Selected Topics in Oil & Gas Process

Design for Graduate Chemical Engineers

Jayanthi Vijay Sarathy

First Edition
Selected Topics in Oil & Gas Process Design for Graduate Chemical Engineers

Foreword
Much has changed in the last three decades with the advent of engineering softwares, causing a
paradigm shift in the way Chemical Process Design calculations are performed. The ‘Art of Hand
Calculations’ is becoming a dying art with computers & engineering software providing
solutions. But this can often turn into a crippling pitfall for Chemical Engineers who lack any
basic understanding of Engineering Fundamentals or Industrial Design Standards when
operating an engineering software. Therefore, it is imperative that chemical engineers recognize
the necessity to focus and gain proficiency in applying process design fundamentals prior to
embarking on any engineering undertaking.

While the academia can sometimes lack focus on meeting Industry standards, engineering
graduates are often left behind from picking up the requisite skills as demanded by Process
Design Industry. With an aim to bridge some of the gaps between academia and the industry, the
following free tutorial book aims to provide a practical approach and guide young chemical
engineers in areas of Process Engineering Design through a series of learning modules with an
emphasis on Design Standards.

The author encourages readers to explore the various topics presented and work them out. No
complex engineering software is required for most of the tutorials, except for MS-Excel or even a
paper, pen and a calculator to work out by hand will do just fine. With each topic progressed
with dedication, engineering practioners can expect to improve their proficiency levels in some
of the areas of Upstream Oil & Gas.

THERE IS A FREE COPY AND STRICTLY ONLY FOR LEARNING PURPOSES

With Best Wishes

Jayanthi Vijay Sarathy, M.E, C.Eng, MIChemE, UK
Chartered Chemical Engineer
Selected Topics in Oil & Gas Process Design for Graduate Chemical Engineers

Contents
Topic Page No.
1. Chemical Process Calculations 1
2. Operating Envelopes for Centrifugal Pumps 6
3. Predicting Performance Curves of Centrifugal Pumps in the Absence of OEM Data 15
4. Affinity Laws for Variable Speed Centrifugal Pumps 31
5. Understanding Centrifugal Compressor Surge and Control 35
6. Variable Speed Drives for Gas Compressor Operations 39
7. Load Sharing for Parallel Operation of Gas Compressors 45
8. Centrifugal Compressor Settle Out Conditions 47
9. Gas Compression Stages – Process Design & Optimization 54
10. Design Considerations for Compressor Antisurge Valve Sizing 58
11. Process Design for Natural Gas Transmission 66
12. Vapour Compression for Propane-Propylene Splitters 80
13. Boil off Gas Analysis of LNG at Receiving Terminals 83
14. Gas Condensate Separation Stages – Design & Optimization 91
15. Process Design for Instrument Air Systems 98
16. Understanding High Integrity Pressure Protection Systems 104
17. Process Safety Valve (PSV) Sizing – API 520/521/526 108
18. Key Process Considerations for Pipeline Design Basis 120
19. Natural Gas Pipeline Transmission Cost and Economics 125
20. Evaluating Pipeline Operational Integrity – Sand Production 136
21. Economic Insulation for Industrial Piping 144
22. Front End Loading for Pipeline Project Management 150
23. Flash Steam and Steam Condensates in Return Lines 154
24. Single Phase Liquid Vessel Sizing for HYSYS Dynamics 159
25. Key Thermo-Physical Properties of Light Crude Oils 160
26. Evaporation Pond Process Design in Oil & Gas Industry 165
27. Exploring LPG Cylinders for Medical Oxygen – A Preliminary Study 169
28. Heating Value Estimation for Natural Gas Applications 172
29. Empirical Approach to Hydrate Formation in Natural Gas Pipelines 176
30. Methodology for Slug Catcher Sizing 179
 Module 1
Chemical Process Calculations
Often engineers are tasked with The process conditions are 40 bara, 500C &
communicating equipment specifications 1,000 kmol/h of natural gas. The process data
with suppliers, where process data needs to requested by the vendor is as follows,
be exchanged for engineering quotations & 1. Natural Gas Molecular Weight & Density
orders. Any dearth of data would need to be
2. Component & Total Mass flow
computed for which process related queries
3. Component & Total Actual Volume flow
are sometimes sent back to the process
engineer’s desk for the requested data. 4. Component & Total Standard Volume flow
5. Component mass %
The following module is a refresher for non-
process engineers such as project engineers, 6. Component Volume %
Piping, Instrumentation, Static & Rotating 7. Component Parts per million (ppm) by
Equipment engineers to conduct basic weight.
process calculations related to estimation of 8. Component Parts per million (ppm) by
mass %, volume %, mass flow, actual & volume.
standard volumetric flow, gas density, parts
Component Molar Flow [M]
per million (ppm) by weight & by volume.
To estimate the component molar flow, the
Problem Statement mixture molecular weight [MW] is evaluated
A vendor requests the project engineer to first by using Kay’s mixing rule as follows,
provide certain natural gas process data for
!" = ∑ %! !"! , Where, i = 1 to n (1)
evaluation. The gas composition is as follows,
Where,
Table 1. Natural Gas Composition & Properties
yi = Mole fraction of each component, -
Component MW Mol%
MWi = Component MW, kg/kmol
- kg/kmol %
The component molar flow rate is calculated
Methane 16.04 76.23 as,
Ethane 30.07 10.00 !! = %! × !, Where, i = 1 to n (2)
Propane 44.01 5.00 Where,
i-Butane 58.12 1.00 Mi = Component Molar Flow, kmol/h
n-Butane 58.12 1.00 M = Total Molar Flow, kmol/h
i-Pentane 72.15 0.30 Component & Total Mass Flow
n-Pentane 72.15 0.10 To estimate the component mass flow [mi] &
total mass flow [m], the relationships are,
n-Hexane 86.18 0.05
'! = !! × !"! , Where, i = 1 to n (3)
H2O 18.02 0.25
' = ∑ '! , Where, i = 1 to n (4)
CO2 44.01 3.00
Where,
H2S 34.08 0.07
mi = Component Mass Flow, kg/h
N2 28.01 3.00
m = Total Mass Flow, kg/h
Page 1
Component & Total Volume Flow Component PPM by Weight, ppm(w)
To estimate the component volume flow & The component PPM by weight, ppm(w) is
total volume flow, the relationship is based on computed as,
the principle that 1 kmol of ideal gas occupies [:; %]! × 10,000 = [>>'(:)]! (12)
22.414 m3 at 00C [273.15 K]. In order to
estimate the volume flow of each component,
Component PPM by Volume, ppm(v)
the volume occupied by a gas at standard The component PPM by volume, ppm(v) is
pressure & temperature [STP], i.e., 1 atm & computed as,
150C, the relationship is corrected to, [ABC %]! × 10,000 = [>>'(A)]! (13)
"".$%$
(! = !! × "&'.%( × [*[℃] + 273.15] (5) Results
The total volume flow rate is, Based on the steps provided, the estimated
results of mass %, volume %, mass flow,
()*+ = ∑ (! , Where, i = 1 to n (6)
actual & standard volumetric flow rates, parts
Where, per million (ppm) by weight & by volume is
Qi = Component Volume Flow [Sm3/h] shown in Appendix B & Appendix C.
Q = Total Volume Flow [Sm3/h] Appendix A: Gas Compressibility
Component Mass % Factor, Z for Natural Gas Estimation
The component mass % is calculated as, To assess the properties of natural gas,
,!
calculations can be begun by estimating the
'! % = ,
, Where, i = 1 to n (7) properties using Kay’s Mixing Rule as follows,
Component Volume % Mixture molecular weight [MW], kg/kmol
The component volume % is calculated as, !" = ∑ %! !"! (14)
- Mixture Pseudo Critical Pressure [Pc], psia
(! % = - ! , Where, i = 1 to n (8)
"#$
D/ = ∑ %! D/,! (15)
Actual Volumetric Flow Rate [Qact]
Mixture Pseudo Critical Temperature [Tc], 0R
The actual volumetric flow is computed as,
,
*/ = ∑ %! */,! (16)
(./* = 0 (9)
%& Gas Specific Gravity [gg], [-]
Where, the density of the natural gas [rNG] is 56
E; = 56 ) ; MWair = 28.96 kg/kmol (17)
computed from the expression that takes into '!*

account the gas compressibility factor, Z as, From the above, Kay’s Mixing Rule does not
512 = 7
3'(# ×56
78/'' (10)
give accurate pseudocritical properties for
'(# ×8×9'(# higher molecular weight mixtures
R = 0.0831447 m3.bar/kmol.K (particularly C7+ mixtures) of hydrocarbon
The gas compressibility factor, Z of natural gases when estimating gas compressibility
gas can be computed based on the DAK factors [Z] and deviations can be as high as
Equation of State [EOS] as described in 15%. Therefore, to account for these
Appendix A. Alternatively the Standard differences, Sutton’s correlations based on
Volumetric Flow rate can be computed as, gas specific gravity can be utilized as follows,
% ×'
!!"# = ∑ !$ = $ (!"# ×)!"#% × $
($#% ×)$#%
% '(* /ℎ (11) D</ = 756.8 − 131.07E; − 3.6E;" (18)
!"# !"# %$#%
*</ = 169.2 − 349.5E; − 74.0E;" (19)
Where, Zstd is taken to be 1.0
Page 2
The above equations are valid for the gas Where,
specific gravities range of 0.57 < gg < 1.68. ".$%&+
!! = '(+
(23)
Using the Sutton correlations, the reduced
properties are calculated as, rr = Pseudo-Reduced Density [-]
"
!! = " (20) Tr = Pseudo-Reduced Temperature [-]
!"
# The constants A1 to A11, are as follows,
#! = (21)
#!" Table 2. DAK EoS A1 to A11 Constants
However the pseudocritical properties are A1 0.3265 A7 –0.7361
not the actual mixture critical temperature
A2 –1.0700 A8 0.1844
and pressure but represent the values that
A3 –0.5339 A9 0.1056
must be used for the purpose of comparing
corresponding states of different gases on A4 0.01569 A10 0.6134

the Z-chart, as shown below in the Standing & A5 –0.05165 A11 0.7210
Katz, 1959 chart for natural gases. A6 0.5475

DAK-EoS has an average absolute error of

0.486% in its equation, with a standard
deviation of 0.00747 over ranges of pseudo-
reduced pressure and temperature of 0.2 <
Ppr < 30; 1.0 < Tpr < 3.0 and for Ppr < 1.0 with
0.7 < Tpr < 1.0. However DAK-EoS gives
unacceptable results near the critical
temperature for Tpr = 1.0 and Ppr >1.0, and
DAK EoS is not recommended in this range.
DAK EoS for NG Mixtures with Acid Gases
Natural Gas is expected to contain acid gas
fractions, such as CO2 and H2S, & applying the
Standing & Katz Z-factor chart & Sutton’s
pseudocritical properties calculation methods
would yield inaccuracies, since they are only
valid for hydrocarbon mixtures. To account
for these inaccuracies, the Wichert & Aziz
Figure 1. Natural Gas deviation factor chart correlations can be applied to mixtures
(Standing & Katz, 1959)
containing CO2 < 54.4 mol% & H2S < 73.8
Due to the graphical method of Standing &
mol% by estimating a deviation parameter
Katz chart, the Z factor can be estimated using
[e], which is used to modify the pseudocritical
Dranchuk and Abou-Kassem Equation of State
pressure & temperatures. The deviation
[DAK-EoS] which is based on the data of
parameter [e] whose units are in 0R, and psia,
Standing & Katz, 1959 and is expressed as,
% % % % K = 120[L=.> − L%.? ] + 15[M=.( − M$ ] (24)
$ = 1 + '($ + ## + #%% + #&& + #'') *! +
$ $ $ $ Where,
%( %) %( %)
'(& + # + ) *!' − (( ' # + ) *!) + A = YCO2 + YH2S in Gas mix [Y = mole fraction]
$ ##
$ $ $ ##
+$# #
+($* (1 + ($$ *!' ) . / 0 ,%** +$
(22) B = YH2S in Gas mixture [Y = mole fraction]
#$%
Page 3
Applying [e], the modified pseudocritical Nevada, USA, 22-26 Sep, SPE-14265-MS
pressure & temperature is, (http://dx.doi.org/10.2118/14265-MS)
/
#-. = #-. − 2 (25) 5. “Calculation of Z Factors For Natural Gases
+
Using Equations of State”, Dranchuk, P.M.
/ "!" #!"
!-. = (26) and Abou-Kassem, H. 1975, J Can Pet
#!" ,1[$,1]4
Technol 14 (3): 34. PETSOC-75-03-
Where, T’pc & P’pc are valid only in 0R and psia. 03 (http://dx.doi.org/10.2118/75-03-03)
Based on the calculated modified 6. “Compressibility Factors for Naturally
pseudocritical pressure [P’pc] and Occurring Petroleum Gases, Piper, L.D.,
temperature [T’pc], the pseudo-reduced McCain Jr., W.D., and Corredor, J.H., SPE
pressure [Pr] & temperature [Tr] is, Annual Technical Conference & Exhibition,
& [),-.] Houston, 3–6 October 1993, SPE-26668-
#)! = &. (27)
,- [),-.] MS (http://dx.doi.org/10.2118/26668-MS)
( [°/ ]
$)! = (. (28) 7. https://petrowiki.org/Real_gases
,- [°/ ]

".$%&,+
!)! = (29)
'(,+
Using the calculated values of Ppr Tpr & rpr,
compressibility factor, Z is determined by
using DAK EoS. Owing to the value of ‘Z’ being
an implicit parameter in calculating rpr as
well as in DAK-EoS, an iterative approach,
whereby Z value is guessed & iteratively
solved to satisfy both modified pseudo-
reduced density [rpr] & DAK EoS.

References & Further Reading
1. “Handbook of Natural Gas Engineering”,
Katz, D.L., 1959, McGraw-Hill Higher
Education, New York

2. “Gases and Vapors At High Temperature

and Pressure - Density of Hydrocarbon”, Kay
W, 1936 Ind. Eng. Chem. 28 (9): 1014-1019
(http://dx.doi.org/10.1021/ie50321a008)
3. “Density of Natural Gases”, Standing, M.B.

and Katz, D.L. 1942 In Transactions of the
American Institute of Mining and
Metallurgical Engineers, No. 142, SPE-
942140-G, 140–149, New York

4. “Compressibility Factors for High-

Molecular-Weight Reservoir Gases”, Sutton,
R.P. 1985. SPE Annual Technical
Conference & Exhibition, Las Vegas,
Page 4
Appendix B: Natural Gas Composition Results

Appendix C: Natural Gas Process Data

Page 5
 Module 2
OPERATING ENVELOPES FOR CENTRIFUGAL PUMPS
In today’s global economy, a necessity exists The following focuses on predicting the
in process facilities to reduce costs and allowable operating range or operating
upgrading component specifications do not envelope for a pump’s range of operation.
necessarily provide fail proof solutions. Of all Problem Statement
the equipment used for both commercial and
A Centrifugal pump is used to transfer water
industrial applications, centrifugal pumps are
from a horizontal vessel to a storage tank as
a common feature. They are characterized by
shown in Figure 2.
their high efficiency with low power
consumption.
To cite a few applications, centrifugal pumps
are used in buildings for water supply, as a
booster and for domestic water supplies,
pumping of sewage and slurries. They are also
used in fire protection systems and for
heating and cooling applications. In addition, Figure 2. Pump System Schematic
they are popular in the beverage, dairy, food The centrifugal pump has the following Head
and oil & gas, petrochemical & chemical (H) vs. Flow (Q) characteristics as shown in
industries. Improper operation of centrifugal Figure 2 and operates with a flow control
pumps, often result in mechanical integrity valve at its discharge line. The flow rate
failures such as, high temperatures, low flow required to be maintained is 189.2 m3/h with
cavitation, low bearing and seal life, reduced the receiving tank pressure at 1.02 bara. The
impeller life, suction and discharge pump suction conditions are as follows,
recirculation.
Table 1. Pump Suction Conditions
To circumvent such operational failures,
Parameter Value Unit
Pump Operating Envelopes are made for a
given pump specification to enable engineers Liquid Temperature [T] 20.0 0C

and operators to make decisions regarding its Liquid Density [r] 997.8 kg/m3
operability. The below picture shows the Dynamic Viscosity [μ] 1.00 cP
various possible failures for the range of 0.001 kg/m.s
operation.
Vapour Pressure [Pv] 0.023 bara

Critical Pressure [Pc] 217.7 bara

Pump Operating Flow [Q] 189.2 m3/h

0.05256 m3/s

Pump NPSHR 0.69 bara

7.05 m

Discharge Head [H] 6.29 bar

64.24 m
Figure 1. Pump Operating Failures [2]

Page 6
The piping details of the pump system piping The pump curves of the centrifugal pump are,
connections are as follows, Table 4. Pump Curves
Table 2. Pump System Piping Details Flow Head Head
Parameter Value Unit [m3/h] [bar] [m]
Suction Pipe Size [ND] 8.625 in 0.0 9.51 97.1
Suction Pipe [WT] 8.18 mm 93 8.76 89.5
Suction Pipe ID [ID] 0.2027 m 100 8.68 88.7
Suction Pipe CS Area [A] 0.03227 m2 120 8.41 85.9
Suction Pipe Roughness [ε] 45.2 μm 140 8.01 81.8
Suction ε/D 0.00022 - 160 7.45 76.1
Suction Pipe Length 5 m 172 7.03 71.8
Discharge Pipe Size [ND] 6.625 in 180 6.70 68.5
Discharge Pipe [WT] 7.11 mm 200 5.73 58.5
Discharge Pipe ID [ID] 0.1541 m 220 4.49 45.8
Discharge Pipe CS Area [A] 0.01864 m2 232 3.61 36.9
Discharge Pipe Roughness [ε] 45.2 μm
The equipment operating pressures and
Discharge ε/D 0.00029 - elevation from pump centerline details are as
Discharge Pipe Length 500 m follows,
The pump system has piping Table 5. Pump Suction Conditions
components/fittings like block valves, check Parameter Value Unit
valves, flow elements, Y-strainers, bends, Suction Vessel Gauge Pressure 0.02 barg
elbows, concentric reducers, etc. which add a
1.04 bara
dynamic pressure loss to the pump flow.
These minor pressure/head losses can be Suction Vessel Elevation 2.00 m

estimated by using the relationship, Suction Vessel Low Low LL 0.25 m

"# ! Discharge Vessel Pressure 1.02 bara
ℎ! = %&
(1)
Discharge Vessel Elevation 1.00 m
Where, K is the Loss Coefficient & can be
Discharge Vessel High High LL 4.00 m
estimated from sources such as Crane’s
Handbook. In this module, the total minor The control valve details to be checked for is,
head loss is assumed to be 0.5 bar. Table 6. Control Valve Details
Additionally, as the liquid exits the discharge
Control Valve Cv@100% Opening FL
piping, exit losses is taken to be 5 m (~0.04
bar) based on Eq. 1 4” Size 236 0.82
Table 3. Minor Losses
6” Size 433 0.84
Parameter Value Unit
Pump System Pressure Losses
SDPMinor Losses 0.5 bar
The pump operating point is the point where
DPExit Losses 0.04 bar the system resistance curve intersects the

Page 7
pump performance curve. This can be The results for suction side pressure drop can
understood as, for a given flow rate, the be summarized as,
resultant pump head arrived at, after Table 7. Suction Side Frictional Losses
accounting for all the static & dynamic losses
Parameter Value Unit
(represented as system resistance) from the
generated Total Dynamic Head (TDH). The Reynolds Number [Re] 329,699 -
factors contributing to system resistance is,
Flow Behaviour Turbulent -
1. Suction Piping & Fittings Frictional Loss
Friction Factor Equation Swamee-Jain -
2. Discharge Piping & Fittings Frictional Loss
3. Control Valve Pressure drop Friction Factor 0.0163 -
4. Total Static Head Suction ε/D 0.00022 -
Suction & Discharge Frictional Losses
Pressure Drop [ΔP] 0.0053 bar
To estimate the total static head, the worst
case scenario of Low Low Liquid Level (LLLL) The calculations are similarly performed for
in the suction vessel & High High Liquid Level discharge side with a discharge side velocity
(HHLL) in the receiving tank is considered. of 2.82 m/s & the results are summarized
This is taken so to ensure that during LLLL, below,
the pump does not suffer from cavitation due Table 8. Discharge Side Frictional Losses
to Net Positive Suction Head (NPSH) Parameter Value Unit
deficiency, i.e., even during LLLL, NPSHA >
NPSHR. To estimate NPSHA, the suction line Reynolds Number [Re] 433,408 -
frictional pressure drop is calculated using Flow Behaviour Turbulent -
the Darcy-Weisbach equation (Appendix B).
The Suction Velocity (VS) is estimated as, Friction Factor Equation Swamee-Jain -

" $.$&'&( Friction Factor 0.0165 -

!! = # = $.$)''* ≈ 1.63 )/+ (2)
!
Suction ε/D 0.00029 -
The Suction Side Reynolds Number is,
$.'$'*+&×+.()×--*.. Pressure Drop [ΔP] 2.123 bar
,- = = 329,372 (3)
$.$$+
Total Static Head
As the calculated suction side Reynolds
number is much higher than 4000, the flow is The suction side static head is computed as,
well into the turbulent region. The friction (/**** */+,+- ).+//+, ×2×3
∆)- = 4!'
(6)
factor can now be calculated using Colebrook
[!.#$*#]×778.9×7.94
equation but owing to its implicit nature, the ∆)- = 4!'
≈ 0.22 -./ (7)
friction factor is calculated using Swamee-Jain
Therefore the pump’s NPSHA is calculated as,
correlation (Appendix B),
!.#$
93:;# = D3/ + 3!012345 − 36 − D37184 (8)
!= "."""$$ '.&( $ = 0.0163 (4)
%&'(!" )
%.&
*
%$)%&$".)
+, 0)12: = 0.22 + 1.04 − 0.023 − 0.0053 (9)
Therefore, the frictional pressure drop is, 93:;# = 1.23 6787 (10)
$.$+()×&×--*..×+.()"
∆3 = = 0.0053 678 (5) The calculated NPSHA is 1.23 bara and is
'×$.'$'*+&×+$#
higher than NPSHR of 0.69 bara.
Page 8
The pump suction pressure is calculated as, !.8#U #XU #
Y4Z
#
CV = D1 + > ? E = 0.9216 (26)
(ℎ0000 +ℎ1213 )415512 ×=×> 97! T$
3! = 365!!59 + − D37 (11)
105 0.1
$.&'($.)*+ &*, & . &
(0.25+2)×997.8×9.81 !"# = # $ -! % + $.'&! ' ≈ 0.75 (27)
')$
3! = 1.04 + 105
− 0.0053 (12)
The inlet pressure in psig for choke flow
3! ≈ 1.25 -./. (12.78 9) (13) equation is calculated as by considering
The Pump Discharge Pressure for the flow pressure drop between Pump Discharge and
rate of 189.2 m3/h & corresponding pump Control valve Inlet is very small as,
head of 6.29 bara is calculated as, !$ = ([6.29 + 1.25] × 14.7) − 14.7 = 96.1 3456 (28)

3< = D3 + 3! = 6.29 + 1.25 (14) )# = 96.1 − (3.59 × 14.7) = 43.4 HIJK (29)
∆)JL[LM3 = 96.1 − 43.4 = 53 HIJ (30)
3< = 7.54 6787 (15)
V !.XT
Control Valve Pressure Drop CI = 0.96 − 0.28LV- = 0.96 − 0.28LX#!! (31)
7
The control valve pressure drop becomes,
CI = 0.957 (32)
∆)FG = )H − D)I*J*KLM*NMO − )PNJJNQ (16) >.?@ @
∆(89:;<= = *>.A@BC+ [(96.1 + 14.7) − 0.957 × 0.34] (33)
D)I*J*KLM*NMO = D)I + D)J + ∆)K + ∆)N (17)
∆)\/]^NH = 87 HIJ (34)
∆)FG = 7.54 − 2.123 − 0.27 − 0.5 − 0.04 − Since DPsizing £ DPChoked, then DPsizing = 53 psi
1.02 ≈ 3.59 -./. (18) +.-.'×=.=$'.(*&
A6 = #%
= 124.3 BC)/DC+E (35)
Total Dynamic Losses +×$.-'+(>&.(()*

The total dynamic losses is computed as, Reinserting the calculated Cv value of 124.3,
;<2 = D)I + D)J + ∆)K + ∆)N + ∆)FG (19) the value of Fp, FLP and new Cv is re-computed
iteratively,
;<2 = 0.005 + 2.12 + 0.27 + 0.5 + 0.04 +
Table 9. FP, FLP & Cv Iterations
3.59 ≈ 6.52 -./. (20)
Iteration FP FLP Cv
Control Valve Cv Required
1 0.92 0.75 124.3
The required Cv of the control valve is
calculated by estimating the valve coefficients 2 0.98 0.82 117.4

first followed by checking if choked flow 3 0.98 0.82 117.1

exists (Appendix A), 4 0.98 0.82 117.1
H T T T
5 0.98 0.82 117.1
=R4 = 1 − >S ? = 1 − @8.794A = 0.937 (21)
!
6 0.98 0.82 117.1
H T T T
=R# = 1 − >S ? 1 − @U.!U$A = 0.811 (22) 7 0.98 0.82 117.1
$

# 8 0.98 0.82 117.1

T$
=4 = 0.5 × @1 − >8.794$ ?A = 0.28 (23)
9 0.98 0.82 117.1
#
T$ 10 0.9789 0.8198 117.1
=# = 1.0 × @1 − >U.!U$$?A = 0.319 (24)

∑ " = 0.28 + 0.319 + 0.937 − 0.811 = 0.726(25) Therefore, for a flow of 189.2 m3/h, the Cv
required is 117.1 gpmÖpsi. For the 4” valve
The selected control valve is 4” valve with Cv selected, the % Cv becomes
of 236 and FL of 0.82 from Table 6.
Page 9
 ++*.+ ∑ = = =4 + =# + =R4 − =R# (38)
% A6 = = 49.6 % (36)
')(
Where,
Pump Operating Envelope
K1 = Upstream fitting resistance coefficient [-]
To generate the pump operating envelope, the
above set of calculations is performed for K2=Downstream fitting resistance coefficient [-]
various % Cv between 20% to 80% to KB1 = Inlet Bernoulli Coefficient [-]
estimate the total dynamic head at various KB2 = Outlet Bernoulli Coefficient [-]
flows. Below is the pump performance curve Where,
that includes the system resistance curves for T
H
various % Cv & flow rates. =R4 = 1 − >S ? (39)
!

H T
=R# = 1 − >S ? (40)
$

Where,
D1 = Pipe Inlet Diameter [in]
D2 = Pipe Outlet Diameter [in]
If the upstream and downstream piping are of
equal size, then, KB1 = KB2, and therefore, are
dropped from the ΣK equation. If the
downstream pipe size is similar to upstream
Figure 3. Pump Operating Envelope
pipe size, i.e., D1 = D2, then KB1 = KB2. The most
Appendix A–Liquid Control Valve Sizing [1] commonly used fitting in control valve
Based on ANSI/ISA S75.01.01 standards, to installations is the short-length concentric
size the liquid control valve, the following set reducer. The equations for these fittings are,
of equations can be used, H$
#
=4 = 0.5 × @1 − >S$?A , for inlet reducer. (41)
Step 1: Calculate Piping Geometry (Fp) & !

Liquid Pressure Recovery Factor (FLP) H$

#
=# = 1.0 × @1 − >S$?A , for outlet reducer (42)
ABC $
∑ " ?2 % %
#< = $1 + ' ( ) (37) If the concentric reducers installed on either
>! @!
side of the control valve are identical, then
Where, Fp = Piping geometric Factor [-] #
H$
N1 = Constant [Value = 1.0] ∑ = = =4 + =# = 1.5 × @1 − > $?A (43)
S

N2 = Constant [Value = 890] If the concentric reducers installed on either

Cv = Valve Coefficient [GPM/Öpsi] side of are identical, then, SK = K1 + K2. The
liquid Pressure Recovery Factor (FLP) is
d = Control Valve Size [inch]
calculated as,
The value of Fp is dependent on the fittings
ABC
such as reducers, elbows or tees that are " D" ? % B %
#!< = $ 3 43 ' 5!( + ) (44)
directly attached to the inlet & outlet >! @ E6!
connections of the control valve. If there are Step 2: Calculate Pressure Drop Required
no fittings, Fp is taken to be 1.0. The term SK for Sizing (DPSizing)
is the algebraic sum of the velocity head loss
coefficients of all the fittings that are attached To estimate the DP required for sizing, DPsizing,
first the liquid critical pressure ratio (FF) is
to the control valve & is estimated as,
calculated. Therefore,
Page 10
 V estimate the piping pressure loss, the Darcy-
CI = 0.96 − 0.28LV- (45)
7 Weisbach correlation is used as follows,
Where, 2×?×@×A "
∆3 = (49)
'B
FF = Liquid Critical Pressure Ratio [-]
Where, DP = Pressure drop [bar]
Pv = Vapour Pressure [psia]
f =Darcy Friction Factor [-]
Pc = Critical Pressure [psia]
L = Pipe Length [m]
Using the value of FF, DPchoked is calculated as,
r = Fluid Density [kg/m3]
I*D #
∆)\/]^NH = @ I A [)4 − CI )P ] (46)
D V = Fluid Velocity [m/s]
If DPValve £ DPChoked, then DP = DPSizing D = Pipe Inner Diameter, ID [m]
Else, Repeat calculations for next size. The Darcy friction factor may be determined
Step 3: Calculate Required Control Valve Cv by either using the appropriate friction factor
The required control valve Cv is calculated as, correlation, or from a Moody Chart which is a
G
function of the Reynolds number (Re).
*F = (47) BA@
∆9:;<;=>
>3 E7 H @3 ,- = (50)
? 1@A B C

Where, µ = Dynamic Viscosity [kg.m/s]

Or, estimating in terms of DP gives us,
% The Darcy Friction Factor [f] depends on the
G M
∆,IJKJL& = $ ) - 3 . (48) Reynolds number follows the criteria,
?5 >3 E7 MA
If Re <= 2100 - Laminar Flow Equation
Where, Q = Flowrate [gpm]
If Re <= 4000 - Churchill Equation
r1/r0 = Specific Gravity of Fluid [-]
DPsizing = pressure drop [psig] If Re > 4000 - Colebrook White Equation
Upon calculating the required Cv, it is required The Laminar Flow equation also referred to as
to check if the calculated Cv is within the Cv the Hagen Poiseuille’s equation is,
limit of the selected control valve. If not, the (=
G = DE (51)
next size of control valve is chosen and the
calculations are repeated. To arrive at The Churchill equation combines both
accurate predictions for Cv of the selected laminar and turbulent flow regime friction
size, the calculations are repeated by re- factor expressions. It is accurate to within the
inserting the calculated Cv & control valve size error of the data used to construct the Moody
(d) value into the Fp equation, i.e., Eq. 1 to diagram. This model also provides an
calculate the new value of Fp & further estimate for the intermediate (transition)
continued to estimate the final value of Cv. If region; however this should be used with
the FL value were to change between caution.
iterations, these values would need to be The Churchill equation shows very good
updated, and Cv re-calculated. agreement with the Darcy equation for
Appendix B – Line Sizing [3] laminar flow, accuracy through the
Pressure loss in piping without any size transitional flow regime is unknown, in the
changes or fittings occurs due to friction turbulent regime a difference of around 0.5-
between the fluid and the pipe walls. To 2% is observed between the Churchill

Page 11
equation and the Colebrook equation. For 3. It is not always possible to operate the
Reynolds number up to ~4000, pump at BEP for the conditions required
+I and hence a flow variation of ±10% of BEP
. +' + +'
is allowed.
G = 8 IJF5K + (#GH)+.# L (52)
4. Minimum stable continuous flow (MSCF) is
+(
+ the minimum flow below which the pump
M = N2.457OP Q ) &.( .
RS (53) is not allowed to operate. Although API 610
J,-K G$.'*/
recommends that the rated region is
+(
T = UJ
)*,&)$
KV (54) located between 80% to 110% of BEP, the
F5 preferred region of flow is between 70% to
The Colebrook equation was developed taking 120% of BEP.
into account experimental results for the flow 5. Clause 6.1.12 of API 610 11th edition states
through both smooth and rough pipe. It is “Setting limits for preferred operating
valid only in the turbulent regime for fluid region and the location of rated flow is not
filled pipes. Due to the implicit nature of this intended to lead to the development of
equation it must be solved iteratively. A result additional sizes of small pumps or preclude
of suitable accuracy for almost all industrial the use of high-specific-speed pumps. Small
applications will be achieved in less than 10 pumps that are known to operate
iterations. For Reynolds number up to ~4000, satisfactorily at flows outside of the
+ N⁄B0 '.&+
specified limits and high specific speed
= −2 log+$ I ).*
+ F5M2L (55) pumps that may have a narrower preferred
M2
operating region than specified should be
Owing to the implicit nature of the Colebrook
offered…” Therefore the Allowable
equation which requires iterations, an Operating Region is set by the
alternate correlation may be used to estimate manufacturer as the allowable region to
the Darcy friction factor, i.e., Swamee-Jain operate with stability whilst conforming to
Correlation which is calculated for Re > predefined API 610 vibration limits.
~4000 as, 6. Pumps that are expected to operate less
N.%O frequently can be chosen such that they
/= C⁄D H.GI !
(56)
PQRS3A * F.G D
JKA.L
+T operate at lower speeds at the cost of
efficiency. Since the pump is selected to
Notes on Operating Curves
operate intermittently, a slightly lower
1. For a given set of hydraulic conditions, a efficiency pump is acceptable compared to
pump is designed to operate for one set of a higher speed pump. This will also ensure
flow and head. Deviation from this a longer operating life cycle.
operating point is allowed only to some
degree. References & Further Reading
2. Pump selection closer to the Best Efficiency 1. Control Valve Handbook, 5th Edition,
Emerson
Point (BEP) yields a more efficient pump
with the least amount of vibration & radial 2. https://www.tapflopumps.co.uk/understa
forces acting on the shaft. Pump system nding-centrifugal-pumps
resistance curve when calculated 3. https://neutrium.net/fluid_flow/pressure-
accurately ensures the pump operates loss-in-pipe/
where the performance curve intersects
the system curve.

Page 12
Appendix C – Pump Performance Curves Estimates

Page 13
Appendix D – Liquid Control Valve Sizing

Page 14
 Module 3
Predicting Performance Curves of Centrifugal Pumps in the
Absence of OEM Data
Chemical and Mechanical Engineers in the oil & gas industry often carry out the task of
conducting technical studies to evaluate piping and pipeline systems during events such as pump
trips and block valve failures that can lead to pipes cracking at the welded joints, pump impellers
rotating in the reverse direction and damaged pipe supports due to excessive vibrations to name
a few. Although much literature is available to mitigate such disturbances, a key set of data to
conduct transient studies are pump performance curves, a plot between pump head and flow.
The present module is aimed at applying engineering research in industrial applications for
practicing engineers. It provides a methodology called from available literature from past
researchers, allowing engineers to predict performance curves for a Volute Casing End Suction
Single Stage Radial Pump. In the current undertaking, the pump in question is not specific to any
one industry but the principles are the same for a Volute Casing End suction radial pump.
1. Introduction studies, such as pipeline/piping studies for
Traditionally performance curves are surge analysis and design pressure.
provided by the pump original equipment The working principle of a centrifugal pump
manufacturers (OEM) based on their involves using centrifugal force of a rotating
customized/proprietary models of pump impeller enclosed in a casing to impart
impellers which are designed using methods energy to a fluid. In doing so, a portion of the
such as computational fluid dynamics (CFD) energy is lost in the form of mechanical
and also field tested to provide guarantee in losses with the remaining being transferred
meeting the requirements of the customer. to the fluid that raises the fluid’s pressure
With wear and tear in pumps systems in when discharging from the pump casing. A
ageing facilities that causes deviation from pump impeller consists of vanes that are
the manufactured OEM pump curves, it positioned on a disc to hold fluid and transfer
becomes difficult to accurately predict if the energy as the impeller rotates. Impeller vane
pump can deliver the required head for the geometry is mainly of three types, namely,
new application. forward positioned, straight positioned and
In Brownfield projects, when a plant backward positioned. Backward positioned
undergoes revamp for new process vanes are popularly used for the reason that
conditions, often existing pumps are re- with increase in volumetric flow, power
evaluated and reused for different consumption decreases.
applications with or without impeller Impellers are also characterized as open
trimming. In case where impeller trimming impellers, semi-open impellers and closed
cannot be applied, but instead a larger impellers. Open impellers consist of vanes
impeller is required, the pump is refurbished mounted on central ring to which a rotating
accordingly based on head required for a shaft is connected. In the case of semi-open
given set of pump constraints. In Greenfield impellers, the vane series is held on a circular
projects, when no pump vendor data is disc only on one side while closed/shrouded
available, a necessity arises to use impellers consist of the vane series
performance curves to conduct hydraulic encapsulated between two discs. The main
Page 15
disadvantage of open/semi-open impellers is Euler’s Equation, the fluid’s velocity
that the vanes are sensitive to wear and tear components are expressed as shown in Fig. 2.
but offer the advantage of maintaining the
clearance of the wear ring. Closed/Shrouded
Impellers on the other hand, are less
susceptible to wear and tear and can also
deal with volatile and explosive fluids. The
disadvantage of closed/shrouded impellers
though efficient initially, suffer efficiency
losses due to an increase in clearance of the
wear ring. A representation of the power
consumption trends between the three vane
geometries is shown in Fig. 1.
Fig 2. Velocity Triangle of Pump Impeller

2.1. Net Theoretical Head Relationship

From the velocity triangle shown in Fig. 1, the
net theoretical head is the head developed
based on a finite number of vanes in the
impeller. The aerodynamic relationship
between the net theoretical head (HNet
Theoretical) developed by the fluid for a given
impeller speed and its respective velocity
components at the impeller inner diameter
(ID) represented by subscript ‘1’ and outer
diameter (OD) represented by subscript ‘2’ is
written as,
1
H Theoretical = [U 2Vq 2 - U 1Vq 1 ] (1)
g
[ 1
]
H Net Theoretical = U 2Vq' 2 - U 1Vq'1
g
(2)
Fig 1. Vane Geometry and Power Consumption
In the current undertaking, a backward vane Volumetric Flow, Q = pD2 b2Vr 2e 2 (3)
positioned, shrouded impeller is chosen
From the above relationships, a contraction
considering lower power consumption at
factor (e2) is applied to estimate the flow that
higher flow rates for an end suction single
stage radial flow pump. takes into account the decrease in inlet area
of the impeller due to vane thickness. The
2. Principle of Performance Curves impeller outlet diameter passage width (b2)
Based on the impeller geometry, is considered to estimate the flow rate (Q)
performance curves are derived from an into the impeller. The chief parameter based
aerodynamic analysis of the pump impeller. on which other impeller parameters such as
The basic equation that governs fluid vane angle, passage width, number of vanes,
behaviour at the pump’s impeller is the etc. are calculated is the impeller inner
Euler’s Turbomachine equation relating diameter (ID), D1 and outer diameter (OD),
pump head and fluid velocity. To apply D2 for a given impeller speed (N).
Page 16
2.2. Pump Specific Speed (Ns) The impeller diameters are calculated by
Pump specific speed is a measure to relating the impeller dimensions to the
determine what kind of pumps can be impeller speed (N). The impeller speed is
selected for a given service. Based on the converted to velocity terms, i.e., angular
pump specific speed value, the choice of velocity (w). The relationship between
pumps can vary from radial, Francis Vane, impeller speed and angular velocity is,
mixed flow or axial flow. The pump specific 2 ´ p ´ N rpm
speed [3] is calculated in metric terms with Angular Velocity, w (m / s ) = (7)
60
the below described equation, 2.4. Impeller Vane Angle (b1, b2)
m3
N#Q rpm. When a fluid is rotated by a surface, a certain
Specific Speed, Ns = $ min
% (4)
H
3#
4 m amount of slippage occurs between the
impeller diameter tip and the fluid making
contact with the impeller tip. This causes the
actual fluid velocity leaving the impeller
diameter to be slightly lower than the
impeller tip speed with slippage expressed as

Fig 3. Pump Specific Speed Chart [4] a slip factor (s). This is incorporated into the
The above graph shows a distribution of the velocity triangle relationship to estimate the
pump specific speeds based on which the tangential velocity terms Vq1 and Vq2, radial
type of pump is selected. It is to be noted that velocity terms Vr1 and Vr2 as,
the values of specific speed can be different V
Vq 1 = U 1s - r1 (8)
based on the units of measurements used and Tanb1
Fig 4 is only for illustrational purposes. Based Vr 2
Vq 2 = U 2s - (9)
on the pump speed the volumetric efficiency Tanb 2
can be calculated [3] using the relationship, The slippage factor (s) is computed by
1 relating to the number of vanes (Z) and inlet
Volumetric Efficiency, h v = (5)
1.124 and outlet diameter vane angle, b1 and b2 as,
1+ 2/3
Ns
Sinb 2 R1
Volumetric efficiency is used to estimate the s = 1- , For, £ e limit (10)
b1 Z 0.7
R2
total flow rate entering into the impeller eye
é æ ö
1
ù
ê ç æç R1 ö÷ - e lim it
3
which in turn is used to calculate the impeller ú
÷
eye diameter. Therefore (Qs’) is computed as, Sinb 2 ê ç çè R2 ÷ø ÷ ú
And, s = 1 - ê1 - ç ÷ ú
Q b1 Z 0.7 ê ç 1 - e lim it
Total Flow rate, Q s' = (6) ÷ ú
hv ê ç ÷ ú
êë è ø úû
Qs’ represents the flow that is required to
enter the impeller to meet the discharge flow R1
For, > e lim it (11)
conditions indicated by ‘Q’ since a portion of R2
æ 8.16 Sinb 2 ö
the incoming fluid is expected to accumulate ç- ÷
And, e limit = e è Z ø
(12)
in the pump. Hence all impeller design and
performance curves calculations are made The number of vanes (Z) required is
with Qs’ to meet conditions of Q. calculated as,

2.3. Speed and Angular Velocity

æ D + D1 ö æ b + b2 ö
Z = 6.5 ´ çç 2 ÷÷ ´ Sinç 1 ÷ (13)
è D2 - D1 ø è 2 ø
Page 17
The vane angle at the inner diameter (ID) is p ´ D1 ´ N
Impeller ID Tip Speed, U 1 = (25)
computed from the velocity triangle 60
relationship by relating it to the radial Qs'
Inlet Passage Width, b1 = (26)
component and impeller tip speed as follows, pD1Vr1e 1
æV ö Qs'
Impeller ID Vane Angle, b1 = Tan -1 çç r1 ÷÷ (14) Outlet Passage Width, b2 = (27)
è U1 ø pD2Vr 2e 2
2.5. Impeller Dimensions Relationship The contraction factor (e) for the inner and
The main parameters required to be outer diameters can be estimated by using
estimated are, End of Main Shaft Diameter the thickness of the impeller passage (t) at
(Ds), Hub Diameter (DH), Hub Length (LH), the inlet and outlet diameters as,
impeller inlet passage width (b1), impeller
outlet passage width (b2) impeller eye Zt
Contraction factor, e = 1 - (28)
diameter (D0), impeller inner diameter (D1), pDSinb
and impeller outlet diameter (D2). The The values of Ku, Km1, Km2 and D2/D1 can be
impeller outer diameter (D2) can be computed from Stepanoff Chart [3],
calculated using Stepanoff Chart [2]. To
calculate the above mentioned parameters,
the following equations can be used.
1
é P(HP ) ´ 321000 ù 3
Shaft Dia, Dsh = ê ú (15)
ë N (rpm ) ´ S s ( psi )û
Hub Diameter, DH = (1.5 to 2.0) ´ Dsh (16)

Hub Length, LH = (1.0 to 2.0) ´ DH (17)

The fluid velocity at Impeller Eye (Veye) is
calculated as,
Veye = [(0.07 to 0.11) + 0.00023 N s ]´ 2 gH (18)
The impeller eye diameter (D0) is taken to be,
4 ´ Qs'
Impeller Eye Dia, D0 = + DH2 (19)
p ´ Veye
The various impeller speeds are as follows,
Impeller OD Tip Speed, U 2 = K u 2 gH (20)

OD Radial Velocity, Vr 2 = K m 2 2 gH (21)

ID Radial Velocity, Vr1 = K m1 2 gH (22)

The impeller diameters is calculated as,
60 ´ U 2
Impeller Outer Diameter, D2 = (23)
p´N
æD ö Fig 4. Stepanoff Chart for Ku, Km1, Km2, D2/D1
Impeller Outer Diameter, D1 = D2 çç 2 ÷÷ (24)
From the equations presented, design
è D1 ø
procedures can be commenced by assuming
Page 18
‘b2’ & iteratively calculating until the actual (U 2Vq 2 - U1Vq 1 ) - (U 2Vq' 2 - U1Vq'1 ) (30)
H circ =
head calculated matches with the required g
pump head. Followed by calculating the net
theoretical head, the actual head is calculated H circ =
( ) (
U 2 Vq 2 - Vq' 2 + U 1 Vq'1 - Vq 1 )
(31)
g
by subtracting the pump losses for a range of
flow rates. (U 2 ´ Vs 2 ) + (U 1 ´ Vs1 )
H circ = (32)
g
3. Pump Losses
In a realistic scenario, centrifugal pumps The slip velocity is normalized by the
experience different forms of mechanical impeller tangential velocity as [6],
losses. The different types of losses expected V
s s = 1- s (33)
during pump operation are (i) Circulation U
losses, (ii) Inlet Incidence losses, (iii) Surface Therefore the slip velocities at the inlet
Friction losses, (iv) Volute Friction losses and diameter (ID) and outlet diameter (OD) are,
(v) Diffusion losses. In addition, parasitic
losses are also considered such as (vi) Disc Vs1 = (1 - s s1 ) ´ U 1 = Vq'1 - Vq 1 (34)
Friction losses and (vii) Recirculation losses. Vs 2 = (1 - s s 2 ) ´ U 2 = Vq 2 - Vq' 2 (35)
When these losses are subtracted from the
theoretical head, the actual head developed With the assumption that the slip factor is
by the pump is arrived at. The below figure nearly equal at both the impeller ID and OD,
shows the difference between the net the whirl velocities are written as,
theoretical head and actual pump head. Vq'1 = U 1 (2 - s s1 ) - Vr1Cot b1 (36)

Vq' 2 = U 2s s 2 - Vr 2 Cot b 2 (37)

3.2. Inlet Incidence Losses

Incidence flow losses are characterized by
losses resulting from a forced change of
velocity when fluid enters the pump impeller.
When fluid enters the impeller eye in a
normal direction, it is followed by a radial
change in the direction of fluid flow.
Additionally due to difference between the
Fig 5. Theoretical Head vs. Actual Head
vane inlet angle and angle at which the fluid
3.1. Circulation Losses enters the vane cascade, a loss of head occurs
Circulation flow losses are characterized by due to forced change in velocity. The
circulatory flow that exists within a closed incidence losses are calculated as [6],
impeller channel when the impeller is
rotating. At this point, there would be a
(U 1 - Vq 1 )2
hin = f in ´ (38)
mismatch of relative velocity (W) between 2g
the inlet side and outlet side of the impeller Where, f in = 0.5 - 0.7
vane. The circulation head is calculated as,
Qs'
H circ = H Theoretical - H Net Theoretical (29) Vq 1 = U 1 - (39)
é Z ´ t ´ b1 ù
Tanb1 ´ êpD1b1 - ú
ë Sin(b 1 ) û

Page 19
3.3. Surface Friction Losses Assuming that, D3 = 1.3 ´ D2 (46)
No pump system has perfectly smooth
Taking Volute Width, b3 = 2 ´ b2 (47)
surfaces but instead has some amount of
roughness. As a result when the fluid enters Volute Throat Area, A3 = p ´ D3 ´ b3 (48)
the impeller eye, friction is caused between
the fluid and the disc surface. Taking into 3.6. Disc Friction Losses
account the losses at the solid boundaries Disc friction losses are the result of a viscous
such as stationary vanes, diffuser and the rest friction between the outside portion of the
of the impeller surfaces, the surface frictional impeller Disc and the surface of the pump
head loss is calculated as, casing. Hence in the case of open impellers,
the Disc friction is lower than the case where
b (D - D1 ) ´ (W1 + W2 )
2

hsf = 2 2 (40) closed impellers are used. The Disc friction

2 ´ Sinb 2 ´ H r ´ 4 g losses can be calculated as [6],
æ pD ö æ D - D1 ö
5
b2 ç 2 ÷ ´ Sinb 2 Pdf = C M ´ r ´ w ´ ç 2
3
÷ (49)
è Z ø è 2 ø
Where, H R = (41)
æ pD2 ö Rearranging with r=Qs’/v,
b2 + ç ÷ ´ Sinb 2
è Z ø 5
æ D - D1 ö
Vr 1 CM ´ w ´ ç 2
3
÷
Where, W1 = (42) Pdf è 2 ø
Sinb1 hdf = = (J / kg ) (50)
m Qs'
Vr 2 5
Where, W2 = (43) æ D - D1 ö
Sinb 2 CM ´ w ´ ç 2
3
÷
è 2 ø æ kJ 1 ö
hdf = çç ´ ÷÷ (51)
3.4. Diffusion Losses Qs' è kg 1000 ø
Diffusion Losses are characterized by a loss 5
æ D - D1 ö
of head when the relative velocity at the inlet CM ´ w ´ ç 2
3
÷
è 2 ø æ 102.04 ö
near the impeller eye exceeds the outer hdf = ç ÷ m (52)
Qs' è 1000 ø
impeller’s relative velocity by a certain factor,
5
causing loss of a portion of the velocity head æ D - D1 ö
0.10204 ´ C M ´ w ´ ç 2
3
÷
difference. The diffusion head loss is [6], è 2 ø
hdf = m (53)
éæ W ö 2 ùW 2 Qs'
hDL = 0.25 ´ êçç 1 ÷÷ - 2ú 2 (44) Where,
êëè W2 ø úû 2 g 0.25 0.1
æ ks ö æ s ö
If, W1 W2 > 1.4 C M = çç ÷÷ ´ çç ÷÷ ´ Re -0.2 (54)
è 0.5 ´ D2 ø è 0.5 ´ D2 ø
3.5. Volute Friction Losses Where, b4 is the volute width
The pump volute receives the fluid pumped
by the impeller. Due to its curved shape and U 2 ´ D2 ´ r
Re = 2 (55)
changing area, pressure head is lost as the µ
fluid moves towards the discharge flange. The value of Disc friction loss coefficient (Cm)
Modifying Ref [1] with respect to volute depends on the Disc surface roughness (ks)
throat area, the volute friction loss is, and also the axial gap width (s).
2 2
é æ D2 öù é Qs' ù 3.7. Recirculation Losses
ê q 2 çç
V ´ ÷÷ú - ê ú
è D øû ë A3 û Recirculation losses are caused due to eddies
hvf = 0.8 ´ ë
3
(45)
2g formed in the pump impeller. The
Page 20
recirculation losses also depend on the size of Ref [9] provides an approximated value of 0.6
the impeller in addition to the flow rates into and this has been incorporated into the
the pump that decide the flow pattern. Hence present undertaking.
with larger diameter impellers the
As per Ref [8], a wearing ring clearance of
recirculation losses increase. Pumps with 0.01 inch for rings up to 6 inch diameter or
high specific speeds also tend to exhibit a less is a good practice. For rings greater than
higher chance of recirculation. The head loss 6 inches and up to 12 inch, the clearance is
due to recirculation is estimated as [5], increased by 0.001 inch for every inch of ring
2.5
w 3 ´ D12 æ Q ö diameter. For over 12 inch, increase by
hRL = 0.005 ´ ç1 - ÷ (56)
rQ çè Q0 ÷ø 0.0005 inches per inch of ring diameter over
12 inches. Therefore the clearance width
Where, Q0 = Design Flow rate
taking into consideration the above criteria,
The value of 0.005 for the loss coefficient is éæ D + D1 ö ù
described as the default value as per Ref [5]. bcl = 0.01in + 0.001´ êç 2 ÷ - 6inú (60)
ëè 2 ø û
Using the default value of 0.005, it is
observed by the Author to be very high and 3.9. Actual Pump Head
yields recirculation losses with negative The Actual Pump Head is calculated by
numbers. The recirculation loss coefficient subtracting all the different head losses
depends on the piping configuration calculated from the theoretical pump head.
upstream of the pump in addition to the Therefore the actual head (HAct) is,
geometrical details of the inlet. The current H Act = H NT - (hcirc + hin + hsf + hdL + hvf + hdf + hRL ) (61)
module does not account for the upstream
piping and the Author iteratively estimates 4. Case Study
that the recirculation losses coefficient is to To understand and validate the described
be taken in the order of 1 ´ 10-3 to 1 ´ 10-2 in methodology, procedures are applied to
order to compensate for the piping losses and estimate the performance curves for a certain
arrive at non-negative recirculation loss model of an industrial water pump with a
coefficients. chosen set of process data. The pump model
used for validation is a Grundfos Model No.
3.8. Pump Leakage Losses NB 200-400/392, 4 Pole, 50 Hz, End Suction
Pump leakage losses cause a loss of head and single stage centrifugal pump, Ref [10]. Table
subsequently efficiency due to leakages 4.1 below gives a summary of the input data
through the Disc and wearing ring. These used to predict the performance curves.
volumetric losses can be modelled as loss of
Table 4.1. Input Process and Mechanical Data
flow through an orifice. From Ref [8] and Ref
Service Industrial Water
[9], the leakage loss can be worked out as,
Flow Rate [Q] 364 m3/h
QL = C L ´ AL ´ 2 ´ g ´ H L (57)
Rotational Speed [N] 1493 rpm
From Ref [5], leakage Area is estimated as, Operating Temperature 250C
AL = p ´ D1 ´ bcl (58) Fluid Density [r] 973.6 kg/m3

And Leakage Head Loss, from Ref [8] as, Suction Pressure [P1] 5.0 bara

HL =
( )
3 é U 22 - U 12 ù Discharge Pressure [P2] 10.0 bara
ê ú (59)
4 ë 2´ g û
Page 21
Required Head [H] 52.4 m From the deviation calculated between the
predicted pump performance data and
OEM Pump Efficiency [hp] 73.1 %
manufacturer’s data, the percentage
Motor Rated Capacity 110 kW
deviation for predicted pump head is largely
OEM Impeller ID [D2] 392 mm within ±5% for most data points. The pump
5. Results hydraulic efficiency calculated however
With the data presented in Table 4.1, shows a deviation of most points in the range
calculations were performed and repeated as of ±10% with respect to manufacturer’s data.
shown in Tables 7.1, 7.2, 7.3 for various The pump curve upon approaching shut-off
range of pump flow rates to arrive at the head droops towards y-axis indicating a fall
pump performance curves as shown in below in head as the pump approaches zero flow.
(H vs. Q, h vs. Q). This is a characteristic of end suction
centrifugal pumps where the volute friction
losses begin to increase at lower flow rates
contributing to a decrease in pump head as
shut-off conditions approach. The key
impeller geometry parameters calculated is
shown in Table 5.1 as follows,
Table 5.1. Calculated Impeller Parameters
Diameter of Main Shaft End 51 mm
Hub Diameter 89 mm
Fig 6. Calculated Pump Performance Curve
Hub Length 133 mm
In deriving the performance curves, the Diameter of Impeller Eye 197 mm
min/max operable region is assigned for a
Impeller Outer Diameter 392 mm
range of 80% to 110% of the best efficiency
point (BEP) while the preferred region of Impeller Inner Diameter 177 mm
operation is 70% to 120% of BEP to Vane Angle at Inlet 26.60
minimize failure due to seal and bearing Vane Angle at Outlet 190
failure. A plot is made between
Number of Impeller Vanes 7
manufacturer’s data and predicted pump
Blade Thickness 3.175 mm
performance curves to assess the deviation
as shown below. Inlet Impeller Passage Width 57 mm
Outlet Impeller Passage Width 26 mm

5.1. Pump Losses

In capturing the pump losses experienced,
which causes a departure of performance
from theoretical head to actual head
developed, it is seen that with an increase of
inlet flow, the losses also increase with the
exception of volute frictional losses. A
decrease in the volute frictional losses at
Fig 7. Deviation between Predicted and OEM
Values higher pump flow can be attributed to the

Page 22
decrease in tangential component of outlet With values of ‘k1’and ‘k2, H vs. Q curves for
velocity (Vq2) when pump inlet flow (!!" ) various speeds of 60%, 70%, 80% and 90%,
increases. can be computed as shown in Fig. 9.
6. Technical Notes
1. For a given set of hydraulic conditions, a
centrifugal pump is designed to operate
for one set of flow and head. Deviation
from this operating point is allowed only
to some degree.
2. Pump selection closer to the BEP will yield
a more efficient pump with the least
amount of vibration and radial forces
Fig 8. Pump Losses
acting on the shaft. Pump system
Between the remaining pump losses the
resistance curve should be calculated
surface friction losses and Incidence losses
accurately because the pump operates
contribute the most at higher flow. Diffusion
where the performance curve intersects
losses also follows a similar trend, though in
the system curve.
magnitude is smaller compared to surface
friction losses and incidence losses when 3. In the case of single volute pumps,
estimating pump head vs. flow curve [Fig. 6]. operating away from the BEP will cause
the shaft to deflect with bearings and seals
5.2. Affinity Laws
rubbing against the casing components.
Pump flow can be treated as incompressible
The fluid flow angle into the impeller will
flow since liquids are largely incompressible
also not align to match impeller speeds
that follow Fan laws. Fan laws can be used to
and vane angles causing suction
derive performance curves for various
recirculation, fluid to stall and cavitation.
speeds based on the following relationships.
QµN (62) 4. It is not always possible to operate the
pump at the BEP for the conditions
H µ N 2 (63)
required and hence a flow variation of
Constants ‘k1’ and ‘k2’ can be estimated for
±10% of BEP is allowed.
the base speed of 1493 rpm by re-writing as,
$ 5. Minimum stable continuous flow (MSCF)
! = ## $ → ## = &%' (64)
#&'( *+, is the minimum flow below which the
.
( = #- $ - → #- = &%!' (65) pump is not allowed to operate. Although
#&'( *+,
API 610 recommends that the rated region
is located between 80% and 110% of BEP
the preferred region of flow is between
70% and 120% of BEP.
6. Clause 6.1.12 of API 610 11th edition
states “Setting limits for preferred
operating region and the location of rated
flow is not intended to lead to the
development of additional sizes of small
Fig 9. Pump Performance Curves – Various Speeds pumps or preclude the use of high-specific-
Page 23
 speed pumps. Small pumps that are known References & Further Reading
to operate satisfactorily at flows outside of 1. ‘Design and Performance Analysis of
the specified limits and high specific speed centrifugal Pump’, Khin Cho Thin, Mya Mya
pumps that may have a narrower preferred Khaing, and Khin Maung Aye, World
operating region than specified should be Academy of Science, Engineering and
offered…” Therefore the Allowable Technology, 22 2008
Operating Region is set by the 2. ‘Design and Construction of Single Stage
manufacturer as the allowable region to Centrifugal Pump (Impeller)’Khin Maung
operate with stability whilst conforming to Htwe, GMSARN International Conference
predefined API 610 vibration limits. on Sustainable Development: Issues and
7. The Net Positive Suction Head Available Prospects for the GMS, 12-14 Nov 2008
(NPSHA) should always be higher than the 3. ‘Performance Prediction of Centrifugal
(NPSHR) required. Pump (Impeller)’, Nwe Ni Win,
8. Pumps that are expected to operate less International Journal for Innovative
frequent can be chosen such that they Research in Multidisciplinary Field, ISSN-
operate at lower speeds at the cost of 2455-0620, Volume 2, Issue 12, Nov 2016
efficiency. Since the pump is selected to 4. https://www.introtopumps.com
operate intermittently, a slightly lower
efficiency pump is acceptable compared to 5. ‘Theoretical Model to Predict the
a higher speed pump. This will ensure a Performance of Centrifugal Pump Equipped
longer operating life cycle. with Splitter Blades’, Berge Djebedjian,
Mansoura Engineering Journal (MEJ), Vol
6.1. Simplification To Estimate Vq1 & Vq2 34, No.2, June 2009
To calculate the tangential component of
6. ‘Development of an Empirical Equation to
inlet and outlet velocity (Vq1, Vq2) based on
predict the Disk Friction Losses of a
Impeller dimensions directly, the following
Centrifugal Pump’, Ali Nemdili, Diete Heinz
simplification is made. Omitting subscripts
Hellman, The 6th International Conference
in Eq. 9, Eq. 27, Eq. 28 and arriving at an
on Hydraulic Machinery and
expression for Vq1 and Vq2, the expression
Hydrodynamics Timisoara, Romania,
for Vq becomes,
October 21-22, 2004
!!" = #$%&# ' (66)
7. API 610, 11th Edition
! %
&$ = () − &'( * (67) 8. Centrifugal Pumps and Blowers, Church
+, A.H., John Wiley & Sons, 1950
' = 1 − -./0( * (68)
9. ‘A One-Dimensional Flow Analysis for the
1"# Prediction of Centrifugal Pump
&$ = () − $%& (69)
&'( *×3-.45
'() +
6 Performance Characteristics’, Mohammed
Ahmed El-Naggar, International Journal of
Using Eq. 69 with subscripts ‘1’ and ‘2’ for
Rotating Machinery Volume 2013
impeller inner and outer diameter
respectively, the tangential velocities are 10. Grundfos Product Center, https://product-
calculated. selection.grundfos.com/front-
page.html?time=1514274719214&qcid=3

06774341
Page 24
Nomenclature P2 Discharge Flange Pressure [bara]
A3 Volute Throat Area [m2] PRequired Power Required [kW]
AL Leakage Area [m2] Qs Flow Rate [m3/s]
b1 Impeller Passage Width at Inlet [m] Qs’ Total Flow Rate [m3/s]
b2 Impeller Passage Width at Outlet [m] QN Maximum Flow Rate [m3/s]
b3 Volute Width [m] R1/R2 Radius Ratio [-]
CL Leakage Loss Coefficient [-] Re Reynolds Number [-]
Cm Disc Coefficient Friction [-] s Axial gap [m]
Cv Volute Flow Coefficient [-] Ss Shaft permissible Shear Stress [psi]
ds Diameter of Main Shaft End [m] t Blade Thickness [m]
D0 Diameter of Impeller Eye [m] t1 Thickness of Impeller Passage at Inlet [m]
D1 Impeller Inner Diameter [m] t2 Thickness of Impeller Passage - Outlet [m]
D2 Impeller Outer Diameter [m] U2 Impeller OD Tip Speed [m]
D2/D1 Stepanoff Coefficient U1 Impeller ID Tip Speed [m]
D3 Volute Mean Diameter [m] Veye Velocity of Impeller Eye [m/s]
DH Hub Diameter [m] Vr1 Radial Velocity of Flow at Inlet [m/s]
f Leakage Loss Coefficient Vr2 Radial Velocity of Flow at Outlet [m/s]
H Pump Head [m] Vq1 Tangential Velocity of Flow at Inlet [m/s]
HCirc Circulation Head Loss [m] Vq2 Tangential Velocity of Flow - Outlet [m/s]
HIn Incidence Head Loss [m] Vq1’ Actual Whirl Velocity Flow at Inlet [m/s]
Hsf Surface Friction Head Loss [m] Vq2’ Actual Whirl Velocity Flow at Outlet [m/s]
Hvf Volute Friction Head Loss [m] Vs1 Slip Velocity at Inlet [m/s]
Hdf Disc Friction head Loss [m] Vs2 Slip velocity at Outlet [m/s]
HdL Diffusion Head Loss [m] W1 Relative Velocity at Inlet [m/s]
HL Leakage Head Loss [m] W2 Relative Velocity at Outlet [m/s]
HRL Recirculation Head Loss [m] Z Number of Impeller Vanes [-]
HActual Actual head Loss [m] b1 Vane Angle at Inlet [degrees]
HR Hydraulic Radius [m] b2 Vane Angle at Outlet [degrees]
ks Disc Surface Roughness [m] elimit1 Limiting Radius Ratio at Inlet [-]
Ku Stepanoff Coefficient elimit2 Limiting Radius Ratio at Outlet [-]
km1 Stepanoff Coefficient e1 Contraction Factor at Inlet [-]
km2 Stepanoff Coefficient
e2 Contraction Factor at Outlet [-]
L Shaft Power [kW]
r Liquid Density [kg/m3]
LAH Available Hydraulic Power [kW]
hp Pump Efficiency [%]
LH Hydraulic Power [kW]
hv Volumetric Efficiency [%]
m Mass flow rate [kg/s]
ss1 Slip Value at Inlet [-]
N Rotational Speed [rpm]
ss2 Slip Value at Outlet [-]
Ns Pump Specific Speed [rpm]
µ Liquid Viscosity [kg/m.s]
P1 Suction Flange Pressure [bara]
w Angular Velocity [m/s]

Page 25
Appendix A
A Set of calculations is presented to demonstrate the methodology for pump data presented.
Table 7.1. Impeller Dimensions Calculations

Shaft Dimensions
-
9.;9;;7 <!×=>?.@AB< - ×C2,/.,1DAE<AB
./.1
7×8 7
Shaft Power [L] ,= = = 70 01
h, 9.>FF

m3
N×2Q rpm. #&'(×√:.;:<
Pump Specific Speed [Ns] Ns = 3 * min
+ = ( = 189
H $4 m =-.& $)

Main Shaft End Diameter #D #D

?(.?)×(-#;;; ( #&<×(-#;;; (
[dsh] 0!> = 1%(*+,)×B (+!C)2 = 1 #&'(×&;;; 2 = 1.99 456ℎ (~51 ;;)
*

# #
hE = +.+!) = +.+!) = 0.967 @A 96.7%
Volumetric Efficiency [hv] #F !
$
#F !
-* ( +./ $(

Hub Dimensions

Hub Diameter [DH] 0. = (1.5 C@ 2.0) × 0!> , GH#45I 1.75, 0. = 1.75 × 51 = 89 ;;

Hub Length [LH] J. = (1.0 C@ 2.0) × 0. , GH#45I 1.5, J. = 1.5 × 89 = 133 ;;

Impeller Dimensions

?
= 0.1045 5 76
1 9.;9;;
Total Flow Rate [Qs’] !!" = h =
3 9.=@>

Velocity of Liquid at Impeller LGHG = [(0.07 C@ 0.11) + 0.00023$! ] × P2I(, GH#45I 0.09
Eye [Veye] LGHG = [0.09 + (0.00023 × 189)] × √2 × 9.81 × 52.4 = 4.3 ;/T

&×$ 0 &×;.#;&= K' -

Diameter of Impeller Eye [D0] 0; = UI×J * + 0.- = U I×&.(
+ &#;;;' = 0.197; @A 197;;
121

Coefficients 'KU' 1.043 (From Stepanoff Charts, Fig. 4)

Coefficients 'Km1' 0.149 (From Stepanoff Charts, Fig. 4)

Coefficients 'Km2' 0.113 (From Stepanoff Charts, Fig. 4)

Coefficients 'D2/D1' 0.452 (From Stepanoff Charts, Fig. 4)

F×-×G!45 F×-×;H=?
Angular Velocity [w] 8= = = 156 5/6
@9 @9

Impeller Outer Diameter [D2] Selecting D2 = 0.392m (392 mm) from Table 4.1

Page 26
 I×./ ;J@×9.?=F
Impeller OD Tip Speed [U2] (F = = = 30.6 5/6
F F

.
Impeller Inner Diameter [D1] $; = $F × =./ > = 0.392 × 0.452 = 0.1775 AB 177 55
2

-×G×.2 -×;H=?×9.;>>
Impeller ID Tip Speed [U1] (; = @9
= @9
= 13.9 5/6

Inlet Flow Radial Velocity [Vr1] &#; = C7; D2EF = 0.149 × √2 × 9.81 × 52.4 = 4.8 5/6

Outlet Flow Radial Velocity

&#F = C7F D2EF = 0.113 × √2 × 9.81 × 52.4 = 3.6 5/6
[Vr2]

Vane Dimensions

Vane Angle at Outlet

Assume b2=26.60 (Note: To be solved iteratively till HActual=HRequired)
[b2]
% H.L
Vane Angle at Inlet [b1] I; = tan5; = K!2 > = tan5; =;?.=> ≈ 199
2

Number of Impeller ' ('

! = 6.5 × ''!)'" ( × sin '
*" (*! ,../+(,.011 +2.2(0/
( = 6.5 × ',../+),.011( × sin '+2.2)0/( ≈ 7
Vanes [Z] ! " +

6.27 89: +/ 6.27 ×89:(/7.7)

'M070,F = N C5 $
D
= N C5 >
D
= 0.594
6.27 89: +2 6.27 ×89:(2?)
R1 C5 D C5 D
Check £ e limit 'M070,; = N $ =N > = 0.685
R2
R1 0.177 Sinb 2
= = 0.452 < e limit1 < e limit 2 , Therefore, s = 1 -
R2 0.392 b1 Z 0.7
N/0(*/ N/0((F@.@)
Slippage Factor [ss1] )!; = 1 − =1− = 0.991
*2 + ,.> ;=×>,.>

N/0(*/ N/0((F@.@)
Slippage Factor [ss2] )!F = 1 − =1− = 0.991
*2 + ,.> ;=×>,.>

Blade Thickness [t] Taking, 0.125 inches (0.0032m)

Thickness of Inlet
Taking, 0.3175 inches (0.0081m)
Impeller Passage [t1]

Thickness of Outlet
Taking, 0.3175 inches (0.0081m)
Impeller Passage [t2]

Inlet Contraction +,2 >×9.99L;

'; = 1 − -×. = 1 − -×9.;>>×/0((;=) = 0.688
Factor [e1] 2 /0(*2

Outlet Contraction +,/ >×9.99L;

'F = 1 − -×. = 1 − -×9.?=F×/0((F@.@) = 0.898
Factor [e2] / /0(*/

Inlet Impeller Passage 1"# 9.;9HJ

%; = -×. = -×9.;>>×H.L×9.@LL ≈ 0.057 5 (5755)
Width [b1] 2 ×%!2 ×Q2

Outlet Impeller 1"# 9.;9HJ

%F = -×. = -×9.?=F×?.@×9.L=L ≈ 0.026 5 (2655)
Passage Width [b1] / ×%!/ ×Q/

Page 27
 éæ D + D1 ö ù
bcl = 0.01in + 0.001´ êç 2 ÷ - 6inú
ëè 2 ø û
Clearance Width (bcl)
éæ 0.392 + 0.177 1000 ö ù
bcl = 0.01 + 0.001´ êç ´ ÷ - 6 = 0.0152 in (~ 0.4 mm )
ëè 2 25.4 ø úû

0.4
Leakage Area (AL) AL = p ´ D1 ´ bcl = p ´ 0.177 ´ = 0.000223 m 2
1000
Leakage Head Loss ( )
3 é U 22 - U 12 ù 3 é 30.6 2 - 13.9 2 ù
HL = ê
( )
ú = ´ê ú = 28.5 m
[HL] 4 ë 2 ´ g û 4 ë 2 ´ 9.81 û

Leakage Head Loss

QL = 0.6 ´ 0.000223 ´ 2 ´ 9.81 ´ 28.5 = 0.00317 m 3 s
[QL]

$*0
Adding Leakage Loss, Qs’=Qs’ + QL, LL# = V# W# − 3×5×6+
Outlet Tangential MNOP+ ×QIR+ S+ TU VW
789:+

Velocity [Vq1] ;.#;&=F;.;;(#<

LL# = (13.9 × 0.991) − <×;.;;(!×;.;=<
≈ 2.7 ;/T
MNO(#'; )×XI×;.#<<×;.;=<TY Z[
789>+/; ?

1"#
&$F = (F )F − $×%×&
Outlet Tangential &'(*/ ×3-./ 4/ 5C '()+ /D6
/
Velocity [Vq2] .3+ = (30.6 × 0.991) −
,.0,45(,.,,.01
≈ 23.4 9/;
$×#.##'!×#.#!(
678(+2.2# )×<=×,../+×,.,+2)> ?@
)*+,!(.(# -

Actual Whirl Velocity 4.8

Vq'1 = U 1 (2 - s s1 ) - Vr1Cotb1 = [13.9 ´ (2 - 0.991)] - = 0.09 m s
at Inlet [Vq1'] Tan(19 )

Actual Whirl Velocity " (.:

LL- = V- W!- − L*- Z@C[- = (30.6 × 0.991) − MNO(-:.:) ≈ 22.9 ;/T
at Inlet [Vq2']

Slip Velocity at Inlet "

L!# = LL# − LL# = 0.09 − 2.7 = −2.6 ;/T
[Vs1]

Slip Velocity at Inlet "

L!- = LL- − LL- = 23.4 − 22.9 = 0.5 ;/T
[Vs2]

Net Theoretical Head "# " ] [((;.:×--.')T(#(.'×;.;')]

(%G\ M>G]*G\C^N_ = ` [V- LL- − V# LL# = = 72 ;
[HNetTheoretical] '.K#

Theoretical Power 1"# ×V×8 (9.;9HJW9.99?;>)×=>?.@×>F

Absorbed [PTheoretical]
QGR, &SRT#R,0U'M = ;9F.9H
= ;9F.9H
= 74.1 01

Theoretical Shut-off U 22 - U 12 30.6 2 - 13.9 2

H Shut -off = = » 75.6 m
Head [Hshut-off] g 9.81
Table 7.2. Pump Losses Calculations

Circulation Losses

(K/ %"/ WK2 %"2 ) (?9.@×9.J)WY;?.=×(5F.@)Z

Circulation Head Loss FX0#U = = ≈ −2.5 5/6
B =.L;

Page 28
 Inlet Incidence Losses

Inlet Incidence Head (U 1 - Vq 1 )2 , Taking, f (13.9 - 2.7 )2

hin = f in ´ in=0.5, hin = 0.5 ´ = 3.2 m
Loss 2g 2 ´ 9.81
Surface Friction Losses

Inlet Relative Velocity Vr1 4.8

W1 = = » 15 m s
[W1] Sinb1 Sin(19 )
Inlet Relative Velocity !/ % ?.@
1F = /0(* = /0((F@.@) ≈ 8 5/6
[W2] /

@A @×,.-?/
Hydraulic Diameter 4/ C $ / D×/0(*/ 9.9F@×C D×/0((F@.@)
>
F[ = @A = @×,.-?/ ≈ 0.025
[DH] 4/ WC $ / D×/0(*/ 9.9F@WC > D×/0((F@.@)

b ´ (D2 - D1 ) ´ (W1 + W2 )
2
,.,+2×(,../+),.011)×(C(05)!
Surface Friction Losses hsf = 2 ; ℎAB = +×DE8(+2.2)×,.,+×4×/.C0 ≈ 4.5 9
2 ´ Sinb 2 ´ H r ´ 4 g

Volute Friction Losses

Volute Throat Diameter

Assuming, D3 = 1.3 ´ D2 = 1.3 ´ 0.392 = 0.51 m
[D3]

Volute Width [b3] b3 = 2 ´ b2 = 2 ´ 0.026 = 0.052 m

Volute Throat Area [A3] A3 = p ´ D3 ´ b3 = p ´ 0.5096 ´ 0.052 = 0.084 m 2

2 2
é æ D2 öù é Qs' ù
êVq 2 ´ çç ÷÷ú - ê ú
è D øû ë A3 û
hvf = 0.8 ´ ë
3
Volute Friction Loss
Head 2g
,.-?/ / ,.2,-1B,.,,-2> /
3F?.H × C ,..2 D6 5C ,.,61
D
ℎ\] = 0.8 × ≈ 13.2 5
F×=.L;

Disc Friction Losses

Taking Disc Surface Roughness [ks] = 5 microns (5 x 10-6)

Disc Friction Loss Head Axial Gap [s] = 12.7 mm (1.27 x 10-2)
Viscosity of Water at 250C = 0.00091 kg/m.s

U 2 ´ D2 30.6 ´ 0.392 ´ 973.6

Reynolds Number Re = 2 = 2 = 6,416,773
µ 0.00091
0.25 0.1
æ ks ö æ s ö
C M = çç ÷÷ ´ çç ÷÷ ´ Re -0.2
Disc Coefficient Friction è 0.5 ´ D2 ø è 0.5 ´ D2 ø
[Cm] 0.25 0.1
æ 5 ´ 10 -6 ö æ 0.0127 ö
C M = çç ÷÷ ´ç -0.2 -3
÷ ´ 6416773 = 2.35 ´ 10
è 0.5 ´ 0.392 ø è 0.5 ´ 0.392 ø

Page 29
 5
æ D - D1 ö
0.10204 ´ C M ´ w ´ ç 2
3
÷
è 2 ø
hdf =
Qs'
Disc Friction Loss Head
5
-3 æ 0.392 - 0.177 ö
0.10204 ´ 2.35 ´ 10 ´ 156 ´ ç 3
÷
è 2 ø » 0.1 m
hdf =
0.1045 + 0.00317

Recirculation Losses
2.5
Recirculation w 3 ´ D12 æ Q ö
hRL = 0.005 ´ ç1 - ÷ , Here 0.005 is replaced with 0.00075
Loss Head rQ çè Q0 ÷ø

Taking Max Flow Rate, Q0= 110% of Rated Flow = 1.1 x 1.1 =0.11 m3/s
2.5
w 3 ´ D12 æ Qs' ö
hRL = 0.00075 ´ ç1 - ÷
rQs' çè Q0 ÷ø
2.5
156 3 ´ 0.177 2 æ 0.1045 + 0.00317 ö
hRL = 0.00075 ´ ´ ç1 - ÷ » 0m
973.6 ´ (0.1045 + 0.00317 ) è 0.11 ø

Diffusion Losses

æ V ö é 4.8 ù J@! (.:

W1 = çç r 1 ÷÷ = ê ú » 15m / s ; \- = &BCOP ' = 1BCO(-:.:)2 ≈ 8 ;/T
è Sinb1 ø ë Sin(19 )û
!

W1 15
Diffusion Loss = = 1.875 > 1.4
Head W2 8
éæ W ö 2 ùW 2 éæ 15 ö 2 ù 82
hDL = 0.25 ´ êçç ÷÷ - 2ú
1 2
= 0.25 ´ êç ÷ - 2ú ´ = 1.24 m
êëè W2 ø úû 2 g ëêè 8 ø úû 2 ´ 9.81

Table 7.3. Total Losses Calculations

Actual H Actual = H Net Theoretical - (hcirc + hin + hsf + hvf + hdf + hRL + hDL )
Head
[HActual] ]cdefgh = ^_ − (−_. ` + a. _ + b. ` + ca. _ + d. c + d + c. _b) = `_. _ e

Required
Power PRe quired =
Qs' ´ r ´ H Act
+ å Losses =
(0.1045 + 0.00317) ´ 973.6 ´ 52.2 + 20.7 = 74.3 kW
[PRequired] 102.04 102.04

Pump 8CDEF(!DG JF.F

Efficiency h^ = 8 = JF.FW;=.L = 72.5 ≈ 73%
CDEF(!DG W8HI""D"
[hP]

Note: The above set of calculations shown is made in MS-EXCEL which performs detailed
calculations. In the above Tables 7.1/7.2/7.3, calculations are shown for rounded-off numbers.

Page 30
 Module 4
Affinity Laws for Variable Speed Centrifugal Pumps
The world is becoming more conscious of pumping applications where the pump
greenhouse gases (GHG) & it’s estimated that duty is not expected to be constant.
nearly 10% of the electricity usage is from 2. Noise & Vibrations are reduced when
operating pumps ranging from domestic running at lower speeds.
pumps, sewerage pumps, air conditioning and 3. Consumers that use only a small portion of
in every other industrial application. The rated flow during varying loads would
mechanical aspect of centrifugal pumps have have the pump running at full load speed
changed fairly little in the last 5 decades, but corresponding to full load power. In such
what has brought a vast change in pump situations, VFD’s help alter the speed to
performance is the control system based on a consume less power during operation.
variable frequency drive (VFD).
4. VFD’s reduce the risk of a motor burnout
A centrifugal pump consists of an impeller in during start-up from excessive in-rush
a casing that raises the fluid’s head for a given current & increases the longevity of the
speed & discharges liquid at a desired equipment.
pressure. For applications that require
attending to a variable flow scenario, a flow When VFD’s Are Not Advantageous
control valve is installed at the pump 1. VFD’s are not to compensate for an
discharge that throttles fluid pressure to improperly selected pump.
achieve the desired flow. But such methods 2. In systems with high resistance like boiler
cause a loss of energy that was initially used feed water (BFW) pumps where the pump
to raise the fluid’s pressure in the pump. has to generate high starting torque to
With the advent of variable speed drives overcome the static head, VFD’s do not
consisting of a pressure sensor & piece of offer much in terms of providing high
circuitry that alters the frequency of the start-up torque.
electric current, the pump’s speed parameter 3. Performance curves that cause the
can be altered to achieve the required flow & operating point to fall off the Q vs. H curves
also avoid throttling using a flow control cause the operating point to operate closer
valve. VFD’s although tend to cause a to the stall region at lower speeds which
temperature rise in the circuitry, sometimes can cause cavitation. In such situations, a
require ventilation systems being new pump with a discharge throttle valve
incorporated for cooling purposes. is required to push the operating point into
Advantages of Variable Frequency Drives the operating envelope which defeats the
purpose of retrofitting with a VSD.
1. VFD’s when newly fitted or retrofitted to
rotating machinery such as pumps are 4. At lower speeds, though noise & vibrations
referred to as variable speed drives (VSD). are reduced, chances exist for structural
For applications where the duty is resonance that can compromise the
expected to be constant without much integrity of the bearing house and support
variation in process conditions, a fixed structures.
speed drive (FSD) should be more cost
effective. But VSD’s are mostly suitable for
Page 31
Selection Process - New Pumps Performance Curves for VSD Retrofit
1. For new pumps its common to oversize the Centrifugal Pumps that run on a fixed speed
pump but this is not recommended as it are characterized by a single Q vs. H curve. In
adds higher initial cost & higher life cycle the event of a retrofit with a VFD, the pump
costs. can operate at various other speeds &
2. When selecting a rotodynamic pump in correspondingly would have their respective
combination with a VSD for a system with Q vs. H curves.
some static head, a pump should be chosen To estimate the Q vs. H curves at other
such that the maximum flow rate is slightly speeds, Fan Laws a.k.a Affinity Laws can be
to the right-hand side of the best efficiency used. Affinity Laws are used under the
point (BEP). The exception is for a constant premise that
flow regulated system, in which case the 1. Liquids are largely incompressible and
recommendation is to select a pump that their density [r] remains fairly constant.
operates to the left hand side of BEP at
2. Frictional losses due to impeller & casing
maximum pressure. This approach construction as well as bearing losses exist
optimizes pump operating efficiency. at lower speeds but are considered to be
3. Some operating profiles may be satisfied lower than the losses experienced at 100%
best by installing multiple pumps, which speed.
could be fixed or variable speed. On/off
As per Affinity Laws, Pump speed [N] is
control can be used to vary flow rate for
related to the Pump Flow [Q], Pump Head [H]
systems in which intermittent flow is
& Pump Hydraulic Power [P] as,
acceptable.
! = #! $ (1)
Selection Process - Retrofit Pumps
1. Often a contingency of 20% - 25% on the % = #" $ " (2)
required system head is added. Therefore & = ## $ # (3)
retrofitting with VSD’s could match pump Where k1, k2, k3 are constants.
systems to actual system requirements
Once the constants k1, k2, k3 are estimated for
more accurately to save considerable
the rated curve [100% speed], the Q vs. H
amounts of energy.
curves can be estimated for other speeds, i.e.,
2. When adding a VSD to an existing electric 90%, 80%, 70%, 60%, 50%, 40%, 30%. It is to
AC motor, the electrical characteristics of be noted that the pump efficiency for a given
the motor & the frequency converter must flow range and for various speeds would be
match. Variable frequency drives work on fairly constant and these are referred to as
the principle of altering the frequency of Constant Efficiency [h] lines.
the incoming current using a frequency
To demonstrate the use of Affinity Laws to
converter that produces a change in the
derive Q vs. H curves for various speeds, a
synchronous speed of the motor for a given
case study is shown. A motor driven
number of poles. Therefore frequency
centrifugal pump operating at 50 Hz, delivers
converters that give smaller levels of
water from 5 bara suction pressure to 10
harmonic current distortion is to be chosen
bara discharge pressure. The process
to avoid over heating the motor windings
parameters & performance curves are as
and avoid the risk of premature failure.
follows,
Page 32
 Table 1. Pump Process Parameters '×-×*
h= +
(7)
Parameter Value
Tabulating the Q vs. H values for the 100%
Service Water
speed case, the values of k1, k2 & k3 are,
Operating Capacity 372 m3/h
Table 3. Fan Law Constants for 100% Speed
Pump Head 52.4 m
Q H P K1,100% K2,100% K3,100%
Rotational Speed 1493 rpm
[m3/s] [m] [kW] [m3/s/rpm] [m/rpm2] [kW/rpm3]
Suction Flange Pressure 5.0 bara
0.024 51.1 34.9 1.60E-05 2.29E-05 1.05E-08
Discharge Flange Pressure 10.0 bara
Liquid Density 973.6 kg/m3 0.039 53.3 40.5 2.61E-05 2.39E-05 1.22E-08

0.049 54.0 45.0 3.29E-05 2.42E-05 1.35E-08

0.067 54.3 53.7 4.50E-05 2.44E-05 1.61E-08

0.085 53.9 62.4 5.68E-05 2.42E-05 1.88E-08

0.090 53.7 64.9 6.02E-05 2.41E-05 1.95E-08

0.095 53.4 67.4 6.36E-05 2.40E-05 2.03E-08

0.100 53.1 69.9 6.70E-05 2.38E-05 2.10E-08

0.105 52.7 72.3 7.04E-05 2.37E-05 2.17E-08

Figure 1. Pump Performance Curves 0.110 52.4 74.6 7.36E-05 2.35E-05 2.24E-08
Pump speeds chosen for retrofitting are, 0.120 51.3 79.5 8.07E-05 2.30E-05 2.39E-08
Table 2. Electric Motor Speeds
0.131 50.1 84.1 8.79E-05 2.25E-05 2.53E-08
Speed [%] Speed [rpm]
0.152 47.1 92.4 1.02E-04 2.11E-05 2.78E-08
100 1493
90 1344 0.202 37.2 104.1 1.35E-04 1.67E-05 3.13E-08

80 1194 Using 100% speed k1, k2, k3 values, Q, H, P, h

70 1045 for other speeds shown in Table 2, are,
60 896 !)%% = #!,!%%% × $)%% (8)
"
%)%% = #",!%%% × $)%% (9)
Design Methodology
#
As described previously, fan constants k1, k2, &)%% = ##,!%%% × $)%% (10)
k3 are estimated for each volumetric flow rate '#$% ×-×[*#$% ⁄!%".%(]
h)%% = (11)
(Q) & corresponding head (H) for 100% +#$%

speed. Using these k values, Q vs. H curves is !2%% = #!,!%%% × $2%% (12)
calculated for various speeds [Table 2]. "
%2%% = #",!%%% × $2%% (13)
'
#!,!%%% = !()# (4) #
&2%% = ##,!%%% × $2%% (14)
* '&$% ×-×[*&$% ⁄!%".%(]
#",!%%% = !()#! (5) h2%% = (15)
+&$%
+
##,!%%% = !()#" (6) !3%% = #!,!%%% × $3%% (16)
"
The hydraulic efficiency is estimated as, %3%% = #",!%%% × $3%% (17)

Page 33
 # '($% ×-×[*($% ⁄!%".%(]
&3%% = ##,!%%% × $3%% (18) h4%% = (23)
+($%
''$% ×-×[*'$% ⁄!%".%(]
h3%% = +'$%
(19)
Note: 1 kJ/kg = 102.04 m.
!4%% = #!,!%%% × $4%% (20) With the above set of calculations made, for
"
%4%% = #",!%%% × $4%% (21) each calculated values of Q, H, P, h for speeds
#
of 90%, 80%, 70%, 60%, the pump
&4%% = ##,!%%% × $4%% (22)
performance curves at constant efficiency is,

Figure 2. Performance Curves at Constant Efficiency for Various Speeds

References & Further Reading Appendix A

1. “Variable Speed Pumping - A Guide to Fluid Head Conversion - ‘metres’ to ‘kJ/kg’
Successful Application”, US Department of 1 * = 1. $. ,
Energy, Energy Efficiency and Renewable
Multiplying and Dividing by 1000 and kgf,
Energy
5 8.9
1000 67 = 1000 67
2. https://southernswater.com.au/variable- ) )

speed-pumps/ Taking 1 kgf = 9.8 N and substituting in above,

3. http://jmpcoblog.com/hvac-blog/modern- ! $.&
1000 "# = 1000 '.( $ → 1000 "# = 102.04 )
!

pump-selection-part-4-why-you-should- ! !

! 65
never-over-head-a-variable-speed-pump = 102.04 ,
67)
4. https://product-selection.grundfos.com


Page 34
 Module 5
Understanding Centrifugal Compressor Surge and Control
Ask a chemical or mechanical engineer, what gear teeth. The power of a surge is also
does a compressor surge do, and he would proportional to the capacity (flow, power,
shudder merely thinking of the pressure ratio) and even in the case of small
consequences. The centrifugal compressor is turbo compressors, the gear teeth wear out
the heart of any oil & gas facility and since when the impeller rotates in the opposite
the last 100 years has been subjected to direction during a surge. The bottom line is -
scrutiny as to what is the perfect control Always Avoid a Surge in Rotating Equipment.
mechanism. Surge in a centrifugal Typical Single Stage Compression System
compressor can be simply defined as a
A typical single stage compressor system
situation where a flow reversal from the
shown in Figure 1, consists of,
discharge side back into the compressor
1. A centrifugal compressor driven by a gas
casing occurs causing mechanical damage.
turbine, steam turbine or electric motor.
The reasons are multitude ranging from 2. A suction scrubber to disengage any
driver failure, power failure, upset process carryover liquids that can potentially wear
conditions, start up, shutdown, failure of anti- out impellers that run at high velocities of
surge mechanisms, check valve failure to the order of 200 m/s to 500 m/s.
operator error to name a few. The
3. A discharge cooler to cool the compressed
consequences of a surge are more mechanical
vapours to the required export
in nature whereby ball bearings, seals, thrust
temperature.
bearing, collar shafts, impellers wear out and
4. Check valves at compressor discharge to
sometimes depending on the how powerful
prevent any backflow of vapours into the
are the surge forces, cause fractures to the
compressor in the event of a surge.
machinery parts due to excessive vibrations.
5. An anti-surge valve (ASV) that recycles
cold gas from the discharge cooler to the
suction to keep the operating point away
from the surge line.
6. A hot gas recycle is included, if the ASV is
inadequate.

Figure 1. Bearings dislodged from containment

Here is an image that shows the bearings

being dislodged from its containment. The
effects of surge are also contagious and due
to excessive shaft vibrations, the gearbox
connected between the compressor and the Figure 2. Typical Centrifugal Compression System
driver is also not spared at the bearings and

Page 35
Compressor Surge Protection Agents Figure 4 shows a hot gas recycle installation
Anti-surge Valve (ASV) that compensates for the ASV’s deficiency,
The chief protecting agent in a centrifugal thereby keeping the operating point away
from the surge line during an ESD.
compressor is the anti-surge system that
consists of a control valve with the associated
piping. The ASV recycles cold gas from the
discharge side cooler back to the compressor
via the suction scrubber to keep the
operating point away from the surge line.
Hot Gas Recycle Valve (HGV)
Although the anti-surge valve is the chief
protector, in brownfield projects, often the
ASV becomes inadequate to deal with a
compressor surge due to addition of new
compressors in parallel or series (e.g., Figure 4. Surge Avoided with Hot Gas Recycle

booster compressors), change of plant piping In recent decades, with tools such as dynamic
or change of vapour composition. In such simulation, the quantity of hot gas to be
situations, a necessity arises to recycle more recycled can be determined without recycling
flow for which an additional ASV with quick immoderate amounts of hot gas that can
opening characteristics is installed in parallel overheat the gas compressor with bearings
to the first ASV. When such solutions still fail and seals failing.
to stop a surge event from occurring, a hot Requirements of an Anti-surge Valve
gas recycle (HGV) is used as a last resort. The (ASV) and Hot Gas Recycle Valve
HGV is always to be used in tandem with the
A hot gas recycle/bypass system consists of
ASV and only during an emergency shutdown
piping with an On-Off Valve that is motor or
(ESD). Excessive hot gas recycle also shortens
pneumatic operated and should have a full
the efficacy of the lube oil that is used for
opening time of < 1 sec (for valves between 4
lubrication purposes. Figure 3 shows an
inch to 16 inch). For larger On-Off Valves
example of ASV inadequacy leading to
(above 16 inch), the time is taken to be < 2
recycling insufficient vapour to the suction
sec. In the case of an electric motor driven
during a sudden trip caused by power failure.
compressor, the power source for the motor
operated HGV must be independent lest
during a power failure the motor operated
HGV becomes futile. The hot gas valve and
ASV should be fail open type and is sized for
twice the flow required to keep the operating
point away from surge. During operation,
fluids velocities must be kept less than 0.3
Mach which otherwise causes damage to the
valve and piping due to erosion. A noise limit
of 110 dB is also placed and operating at
Figure 3. Surge during an Emergency Shutdown around 85 dB is acceptable. For good
pressure throttling, the ASV is equipped with
Page 36
linear opening characteristics or a Nevertheless, compressor operating
combination of equal percentage temperatures must never exceed the
characteristics up to ~40% opening with the stipulated or mutually agreed values in order
remaining travel substituted with linear to protect the compressor's internals.
characteristics. The hot gas piping should be Compressor Control Systems
laid as short as possible between the
discharge and suction to have a fast response. In today's world no piece of machinery can
be said to be protected by modern methods
During an ESD scenario (e.g., power loss), without implementing a control system. A
taking a conservative approach for design surge can occur in a matter of seconds or
purposes, the control output signal from the sometimes even milliseconds giving almost
compressor driver after a trip, takes ~300 no time for operators to intervene. Hence a
msec to reach the Distributed Control System control system becomes a part and parcel of
(DCS) and another ~300 msec from the DCS the compressor package.
to reach the HGV to open. However with
advances in technology, these timings can be Although the good old Proportional-Integral-
considered at ~100 msec. In simple terms, a Derivative (PID) control was enough to avoid
lower response time increases the chances of a surge by minimizing the compressor
responding faster to a compressor surge. recycle flow, it did not aid much in reducing /
optimizing the power requirements. With a
Deviations from Design Criteria steady rise in the oil consumption since the
As a thumb rule, the hot gas system is sized 1970s, the necessity of energy efficiency,
for 50% (max) during FEED stage. However safety and environmental friendliness
this needs to be checked with a dynamic became a priority and demanded better
simulation study since over-sizing the hot gas control systems.
system recycles excessive flow that causes
To respond quickly to any process upsets,
the bearings and seals to overheat.
high computational speeds in controllers also
As per API 617 (7th Edition, 2002), Clause became a necessity. This led to the rise of
2.7.1.3, it states, ‘As a design criteria, bearing specialized control equipment known as
metal temperatures shall not exceed 100°C 'Black Boxes' that was the alternate to panel
(212°F) at specified operating conditions with mounted instruments. Black boxes though
a maximum inlet oil temperature of 50°C addressed response times, suffered from
(120°F). Vendors shall provide bearing frequent hardware and software revisions.
temperature alarm and shutdown limits on Black box technology was proprietary with
the datasheets.’ However clause No. 2.7.1.3.1 its own coding languages and often
of the said document also says, ‘In the event experienced compatibility issues when
that the above design criteria cannot be met, interfacing between different manufacturer's
purchaser and vendor shall mutually agree on models. This also meant having to sometimes
acceptable bearing metal temperatures.’ shutdown the machinery causing monetary
In the Author’s experience, this deviation was implications and increased downtime if not
seen up to ~1350C depending on the made part of plant maintenance.
manufacturer and believes that this is due to Advent of Programmable Logic Controller
a variation of operating conditions between With the limitations of black box technology
string test conditions and actual conditions. being recognized, industry honchos realized
Page 37
the necessity of standardizing and comes down, offering better efficiency,
generalizing control systems and their diagnostics, generic parts and scalability.
respective programming languages. These
standardization efforts led to documenting
the IEC 61131 (International Electrotechnical
Commission Standard for Programmable
Controllers) in 1993 and subsequently
revised in 2003.
Programmable Logic Controllers (PLCs)
provided not only computational power but
also were easily integrate-able to the
compressor controls. PLC's offered the
advantage of scalability where new I/O’s
Figure 6. Compressor System with Integrated Interfaces
could be added during any form of plant
modification/expansion depending on the References & Further Reading
type of PLC used (e.g., modular or stacked). 1. ‘Compressor Control: Moving Beyond “Black
PLCs also offer diagnostics capabilities, for Boxes” to Integrated, Automated Platforms
example, to trace through the logs of in Oil and Gas Production’, Rick McLin,
controller output during a fault analysis. Turbomachinery Controls Business
In earlier systems, as shown in Figure 5, a Manager, Rockwell Automation
primary PLC is supplemented with an 2. ‘Axial and Centrifugal Compressors and
auxiliary PLC that controlled systems like Expander-compressors for Petroleum,
lube oil, seal oil / dry gas seals, start up Chemical and Gas Industry Services’, API
sequencing, interlocks, etc. This also required Standard 617, 7th Edition, July 2002
interfacing them properly to allow operators
to diagnose and do a root cause analysis in
the event of, for example, a compressor trip.

Figure 5. Compressor System with Interfaces

However with integrated systems as shown

in Figure 6, that used a dedicated control PLC
with a backup PLC and the necessary hard
wiring, the cost of implementation also
Page 38
 Module 6
Variable Speed Drives For Gas Compressor Operations
Historically the Oil & Gas industry has been provide information on the 'Absorbed Power’
dominated by mechanical prime movers like which can be defined as the power required
gas turbines or steam turbines to drive to sustain the gas compressor at the required
centrifugal gas compressors for large operating conditions during continuous
industrial applications. With time, these operation. The power required to start a gas
prime mover operations have become compressor will always be higher than the
difficult to maintain, considering the steady state absorbed power since, the
Environmental Protection Agency (EPA) electric motor (EM) has to overcome the
regulations that mandate strict emission inertia of the entire gas compressor system to
compliance. A good alternative to combat bring it to normal operating conditions.
emission concerns is to use electric motors as To understand by how much, excess power
a prime mover in modern times with regards required during start-up varies, requires a
to their high rates of efficiencies of the order transient state set of calculations including
of 80% to 90% in addition to the absence of steady state calculations. With the advent of
hydrocarbon emissions from electric motors. engineering software, such studies can be
Electric Motors can be operated in two made and for this module Aspentech’s HYSYS
modes, namely, Fixed Speed Drive (FSD) 2006.5 is used to investigate the power
Mode where the prime mover’s speed cannot requirements for a compressor start-up
be altered during operation & Variable Speed based on a case study.
Drive (VSD) Mode where the prime mover’s
speed can be altered by altering the Case Study
frequency of the current fed to the electric To understand the effects of VSD and FSD
motor. The following module is written to mode of operation on gas compressor start-
explore the power requirement implications up, a case study is made. Based on a certain
for gas compressor operation when gas compressor performance curves, the
employing these modes of operation. following process conditions are employed.
Table 1. Gas Compressor Process Parameters
Introduction
The standard procedure employed by Parameter Value
engineers to perform gas compressor design Operating Capacity 8,100 m3/h
can be found quoted in literature such as
Suction Flange Pressure 0.2 barg
GPSA, JM Campbell to name a few with
Industry Standards, for example, API 617, to Discharge Flange Temperature 25 0C

conveniently customize centrifugal gas Discharge Flange Pressure 2.7 barg

compressors as per customer’s requirements. Gas MW 18.38
However Chemical Engineers need to work
Compressor Speed [100% Curve] 3,000 rpm
beyond the standard practice of estimating
steady state process parameters to avoid The Performance Curves (Polytropic Head vs.
under-estimating power requirements for an Actual Flow & Polytropic Efficiency vs. Actual
effective compressor start-up & restart. This Flow) used for the study is shown below.
is so, since steady state calculations only
Page 39
 Motor Rating 600 kW
Configuration 4 Pole, 50 Hz
Motor Slip 1%
Rated Speed 1500 rpm [With Gearbox]

The EM is also characterized by a Speed vs.

Torque curve enabling to compute the power
and torque required to be generated.

Figure 1. Polytropic Head vs. Actual Volume Flow

Since no information is available on the actual
location of the surge line, the surge margin is
assumed as 10% on the actual volume flow.
The polytropic efficiency vs. flow rate curves
is shown below as,

Figure 3. EM Speed vs. Torque Characteristics

Design Methodology & Approach
Prior to performing a transient study for
power requirements, the gas compressor’s
remaining piping & equipment system details
needs to be available. The general design
approach consists of performing steady state
calculations, i.e., heat & mass balance as well
as sizing the equipment, valves & lines based
Figure 2. Polytropic Efficiency vs. Actual Volume
on customer’s needs and layout. In the
Flow
current module, since no layout information
To drive the compressor, an asynchronous
is available, an approach is proposed for
electric motor is chosen. The slippage in the
preliminary volumes to avoid surge during
electric motor can vary even up to 5% during
start-up & shutdown. This is followed by
start-up depending on its design. However for
shaping these volumes into detailed piping &
this module, a value of 1% is assumed. Note
equipment estimates. A general schematic of
that, the EM is rated at 600 kW & the start-up
the compressor loop envisaged is shown,
power required would be higher than 600
kW. The power absorbed during operating
conditions is 496 kW. In addition to the VSD
and FSD cases, the case of suction throttling
with an FSD is also investigated.
Table 2. Electric Motor Configuration
Parameter Value
Figure 4. Gas Compressor Proposed Layout
Electric Motor Type Asynchronous Induction

Page 40
Equipment Sizing Discharge Side Volume
Equipment Volumes The discharge side volume is predicted for
The main equipment in addition to piping the worst case of surge & this can happen
that contributes to the compressor loop during an emergency shutdown (ESD) which
volumes are suction scrubber and discharge is dependent on the decay rate of the
cooler. In situations where gas condensation compressor speed as well as the recycle flow
occurs after the discharge side air cooler, an rate through the anti-surge valve (ASV). The
additional discharge scrubber is installed to discharge side volume can be taken as
knock out any liquids. The discharge side approximately, 1/3rd of the suction volume.
1
volumes particularly affect the response time !5%6#78*9+,)*+,%- = !!"#$%&' !#*""/+* ×
3
(3)
of anti-surge system. Excessive discharge side 1
Or, !5%6#78*9+,)*+,%- = 3 × 4.5 = 1.5 33 (4)
volumes result in a delay in recycling
discharge gas. Hence discharge side volumes Therefore, the discharge volume is 1.5 m3.
are to be kept as minimum as possible. Anti-surge Valve (ASV) Size

It must be ensured that the anti-surge take- To estimate the anti-surge valve (ASV) Cv, the
off point before the discharge side check ASV inlet & outlet pressure is required. The
valve and air cooler is chosen, such that the compressor vendor would provide
setup is not too close to the compressor performance curves as a plot of discharge
discharge flange (which can cause the anti- pressure vs. flow rate from which the
surge valve to rattle) inducing noise related discharge line losses is added to the
issues. If the anti-surge line tap off point is compressor discharge pressure to arrive at
too far, then it increases the surge response the ASV’s upstream pressure & the ASV’s
time. The rate at which the compressor coast downstream pressure is the sum of suction
down, i.e., speed decay occurs, also side line losses and compressor suction
determines the size of the anti-surge valve pressure. Considering the maximum possible
that regulates the amount of the recycle flow. flow through the compressor is at the
Based on the above considerations, the stonewall region, the ASV can be sized for this
equipment volumes are calculated as flow. However, to avoid equipment operation
described in the next section. at its limits, a margin of 10% to 15% on the
stonewall flow at 3000 rpm is taken to ensure
Suction Side Volume
that the ASV does not recycle excess fluid
The suction side volume can be calculated back to the suction side. Based on a maximum
initially for twice the rated volume flow. allowable compressor suction flow of 10,440
Therefore, m3/h, the ASV Cv size is taken as 980 as per
!!"#$%&',)*+,%- = #!"#$%&' !#*""/+* × %&'()* (1) ANSI/ISA 75.01-1985 standard estimation.
0122
Or, !!"#$%&',)*+,%- = 3422 × 2 = 4.5 33 (2) Start-up Operations
Therefore on the suction side, a preliminary The compressor loop is checked for stability
volume of 4.5 m3 is taken. This is a in terms of surge and power adequacy for
preliminary estimate that is subjected to different start-up modes. This is applied to
change depending on heat & material balance, 1. Fixed Speed Gearbox coupled Electric
equipment sizing & transient study results for Motor (EM) + Centrifugal Compressor
start-up & shutdown scenarios. (CC) configuration

Page 41
2. Variable Frequency Drive (VFD/VSD) + increasing gas density. This operation
Electric Motor (EM) configuration. reduces the power required to compress the
3. In addition to VFD and FSD incoming gas i.e., compressor start-up power
configurations, the case of using a suction is also decreased. However, a limitation exists
throttle valve at 100% ASV opening is also on the suction throttling operation while
studied. reducing the suction flow rate, because it
Suction throttling methods involve the use of results in the operating point moving closer
a globe valve or a butterfly valve at the to the surge line during start-up. Hence the
suction line of the CC to cool the gas thereby throttling valve operation must be regulated.
Table 3. Gas Compressor Start-up Cases

Electric Motor Configuration Anti-surge Valve (ASV) Position Fixed Speed & Fixed
Start-up ASV opening with /
Case No. without Suction
Fixed Speed Variable Speed Fixed ASV Variable ASV Throttling

1 Ö Ö
2 Ö Ö
3 Ö Ö
4 Ö Ö
5 Ö Ö
6 Ö Ö

Results & Discussions
Case 1: Fixed Speed with Fixed ASV at Start up no surging occurs & the compressor reaches
For Case 1, based on the simulations, the the rated point of 8,100 m3/h and 2.7 barg.
following plot shows the operating point
migration. For the volumes & ASV size used,
no surging occurs & the compressor reaches
the rated point of 8,100 m3/h & 2.7 barg.

Figure 6. Case 2: Operating Point Migration
Case 3: Variable Speed with Fixed ASV Start up
Variable frequency drives (VFD) are
particularly useful when a variation in speed
Figure 5. Case 1: Operating Point Migration is required, to control gas throughput during
Case 2: Fixed Speed with Variable ASV Start up production changes. VFD’s also offer the
For Case 2, based on the simulations, the advantage of lowering compressor speed
following plot shows the operating point during turndown conditions thereby avoiding
migration. For the volumes & ASV size used, gas recycling that causes energy wastage.
Page 42
During start-up, the EM speed ramp-up rate is & ASV size used, no surging occurs & the
achieved by altering the frequency of the compressor reaches the rated point of 8,100
current passing through the EM which m3/h and 2.7 barg discharge pressure.
thereby raises the CC speed gradually. In Case
3, for the volumes & ASV size used, no surging
occurs & the compressor reaches the rated
point of 8,100 m3/h and 2.7 barg discharge
pressure.

Figure 9. Case 5: Operating Point Migration

Case 6: Variable Speed with Suction Throttling

For Case 6, with variable speed drive, 100%
ASV opening and suction throttling, based on
Figure 7. Case 3: Operating Point Migration the simulations, the following plot shows the
Case 4: Variable Speed with Variable ASV Start migration of operating point. For the volumes
& ASV size used, no surging occurs & the
The variable ASV position is achieved by
compressor reaches the rated point of 8,100
using an Anti-surge controller (ASC). In Case
m3/h and 2.7 barg. Between cases 5 & 6, the
4, no surging occurs & the compressor
peak absorbed power is 865 kW and 666 kW
reaches the rated point of 8,100 m3/h and 2.7
respectively which can be attributed to the
barg discharge pressure.
use of variable ASV position during start-up.

Figure 8. Case 4: Operating Point Migration

Figure 10. Case 6: Operating Point Migration

Case 5: Fixed Speed with Suction Throttling

For Case 5, with fixed speed drive, 100% ASV Start-Up Power Results
opening and suction throttling, suction From the 6 cases simulated for start-up
power requirements, a plot of the power
throttling causes the operating point to closer
to the surge line but does not cross. Based on absorbed vs. Time for a start-up time of 180
the simulations, the following plot shows the sec (~3 min) for cases that have VFD
provision and is shown below.
migration of operating point. For the volumes

Page 43
 lower by ~23% w.r.t case 1. Cases 1, 2 & 5 are
fixed speed EM operation cases and show that
the peak power absorbed is higher since the
EM has to reach breakdown torque threshold
of ~150%. In case 2, using a variable ASV
position for a fixed speed operation during
start-up, does not help in energy reduction.

Conclusions
From the study made, it can be inferred that,
1. VFD’s show a significant reduction in start-
Figure 11. Compressor Start-up Power – All Cases up peak power when compared to FSD
The compressor speed variation during start- operation. In oil & gas applications, with
up is also plotted and shown below. varying production rates, VFD’s are better
equipped to alter the compressor speed to
match production demands & give energy
savings & operational savings.
2. Electric motor’s speed vs. Torque curves
are designed for a higher break-down
torque. This means during start-up with an
FSD, there are chances that the EM would
reach the break down torque but can still
fail to bring the compressor online to
match the required process conditions. In
Figure 12. Compressor Speed vs. Time – All Cases such cases, VFD’s are an alternative.
3. With FSD’s, there is no provision for speed
A comparison of the peak absorbed power for
control. Hence during a stat-up if the ramp
all the start-up cases is shown below.
up rate is too fast, a sluggish anti-surge
Table 4. Start-up Power Comparison controller would struggle to recycle
Peak Absorbed % Power Savings
Case
Power [kW] [w.r.t Case 1]
sufficient flow & prevent surge. Use of
VSD’s help slowing down the start-up ramp
1 (Base Case) 869 0
up rate of the electric motor, enough to
2 866 0.3
allow the anti-surge controller to respond
3 711 18.2 & ensure sufficient recycle gas flows to the
4 666 23.4 suction to prevent a compressor surge.

5 865 0.5 References & Further Reading

Aspentech HYSYS 2006.5 Documentation –
6 666 23.4
Dynamic Modelling
From the above table, it is seen that for cases
1, 2 & 5 which use an fixed speed induction
motor coupled CC with Gearbox arrangement,
the start-up power is higher while for cases 3,
4 & 6 which use a VFD, the start-up power is
Page 44
 Module 7
Load Sharing for Parallel Operation of Gas Compressors
The art of load sharing between centrifugal
compressors consists of maintaining equal
throughput through multiple parallel
compressors. These compressors consist of a
common suction and discharge header.
Programmable logic controllers (PLCs) can be
incorporated with load sharing functions or
can be incorporated as standalone controllers
also. Control signals from shared process
parameters such as suction header pressure
or discharge header pressure can be then fed
to individual controllers such as compressor
speed controllers (SC) or anti-surge
controllers (UIC) to ensure the overall load is
distributed efficiently between the
compressors.
The following module covers load sharing Figure 1. Base Load Operation Method

schemes for parallel centrifugal compressor The speed of compressor B is altered based
operation. on the master pressure controller (PIC) set
point (SP) to attend to the swing in flow
Load Sharing Options
throughputs.
The load sharing options covered are as,
During periods of low process demand,
1. Base Load Method
Compressor B (swing machine) can be
2. Suction Header - Speed Control Method recycling & sometimes even close enough to
3. Equal Flow Balance Method the Surge Control Line (SCL) causing the
4. Equidistant to Surge Line Method swing machine to trip. Additionally, due to
differences in piping layouts & pressure loss,
Base Load Method
the compressor operation would not be
In Base Load method of operation, one
symmetrical, causing operators to frequently
compressor is allowed to run on manual
intervene. With these limitations, the base
mode while the other is controlled through
load method is least preferred.
speed manipulation based on the discharge
header pressure. The pressure controller on Suction Header - Speed Control Method
the discharge header is termed as the Master In the suction header - speed control method,
Pressure Controller (MPC) that alters the no base load exists. Instead the master
second compressor’s speed a.k.a “Swings” the pressure controller (PIC) is shifted to the
compressor speed to cater to varying suction header. The advantage offered is, both
throughputs. In Fig 1, the speed of compressors operate independently despite a
compressor A is manually set (HIC) for a common set point provided by PIC to the
maximum throughput, i.e. Base Load. speed controllers (SC) of both compressors.

Page 45
 signal to each speed controller is achieved by
scaling the total load demand (BIAS A & BIAS
B) to the individual flow controller (FC) on
each compressor. Both Compressor
operations are independent of the Anti-surge
valve (ASV) operation.

Figure 2. Suction Header Speed Control Method

It may be noted that both compressors would

not necessarily be running at the same speed
or flow due to differences in the piping layout
as well as during a compressor recycle since
both anti-surge controllers (UIC A/B) also act
independently of each other.
Figure 3. Suction Header Speed Control Method
To ensure no production losses, the
However certain limitations exist with the
configuration consists of standby compressor
flow balancing method. Due to additional
along with working compressors. During the
control elements, CAPEX cost increases.
failure of one of the compressor, say
Furthermore since the flow element &
compressor A, the PIC issues a signal to
transmitter (FT) is installed on the
increase the speed of compressor B, until the
compressor discharge, additional pressure
standby compressor can be brought online to
drop occurs which represents energy losses.
maintain throughput. In case of layouts that
have no standby compressors, a 2 50% For the cascaded control used, PIC FC
configuration, with no recycle during regular SC, the inner loop (FC) must respond faster
operation must be chosen. This enables the than the PIC outer loop. This causes the
remaining working compressor to cater to master pressure control, PIC to be sluggish. A
100% of the throughput/load at higher faster FC loop also means, the compressor
speeds during failure of the one of the speed would increase rapidly than required
compressors. often reaching maximum speed. Hence this
does not offer the best control strategy.
Equal Flow Balance Method
In the equal flow balance method, the Master Equidistant to Surge Line Method
Pressure Controller (PIC) on the common In the equidistant method, the aim is to
discharge header determines the total load ensure, the deviation/distance between the
demand and alters the speeds of Compressors operating point and the surge control line
A & B via SC. The individual flow control (SCL) in both trains is equidistant.
Page 46
 Some Design Considerations
1. The Master pressure controller which
provides shared information across all
compressors can often be subjected to
harsh field conditions. To circumvent these
issues, redundancy with multiple
transmitters can be provided. This ensures
not only maximum availability but also
hardwiring the transmitters prevents any
loss of signals to the Load sharing system.
2. Depending on the reliability of the control
systems, controllers need to be replaced
sometimes with third party OEM vendors,
each with their own proprietary control
systems. Hence load sharing systems must
be able to integrate different vendors.
Figure 4. Equidistant to Surge Line Method
3. Real Time optimization (RTO) techniques
In this configuration, neither the throughputs based on regression models of steady state
through each compressor or the operating data have gained sufficient footing in
compressor speed is the same, but only the recent years. Short Time RTO of the order
deviation between the operating point & SCL. of a few minutes & Long term RTO of the
It may also be noted that the load sharing order of a few days can be employed to
function (LSIC A/B) that alters the determine the best load distributions
compressor speed, is not fed with the signal between compressors.
from the suction flow transmitter (FT), but
instead the anti-surge controller (UIC A/B) References & Further Reading
and the master pressure controller (PIC) 1. “Advanced Load Sharing Controls for
installed on the common discharge header. Compressor Networks”, Alex Benim, Brian
This would mean, both UIC A/B and LSIC A/B Eldridge, Woodward Inc.
have to coordinate in real time.
A significant advantage of the equidistant to
surge line method is the configuration’s
ability to cater to asymmetrical performance
curves, i.e., dissimilar compressors. In
brownfield modifications, any addition of new
compressors can offer synchronicity issues
including variation in throughputs &
pressures due to differences in performance
curves & piping layouts. Therefore the
equidistant method becomes an effective
configuration for varying loads ensuring both
compressors independently adjust their
respective operations and avoid surge.

Page 47
 MODULE 8
CENTRIFUGAL COMPRESSOR SETTLE OUT CONDITIONS
Centrifugal Compressors are a preferred via the vent valve to reduce the pressure
choice in gas transportation industry, mainly and achieve effective re-start.
due to their ability to cater to varying loads. In Estimating Settle Out Conditions
the event of a compressor shutdown as a Although there are many process simulations
planned event, i.e., normal shutdown (NSD), tools that can be used to conduct a transient
the anti-surge valve is opened to recycle gas study to determine settle out conditions, hand
from the discharge back to the suction calculations based on first principles of
(thereby moving the operating point away thermodynamics can also be easily employed.
from the surge line) and the compressor is In order to do so, the gas compressor system
tripped via the driver (electric motor or Gas can be reduced with the assumptions as
turbine / Steam Turbine). follows, with the philosophy of using a
In the case of an unplanned event, i.e., lumped parameter model, in which an energy
emergency shutdown such as power failure, balance is made across the total volume of the
the compressor trips first followed by the compressor loop taking into account, the
anti-surge valve opening. In doing so, the gas compressor deceleration rate.
content in the suction side & discharge side
mix.
Therefore, settle out conditions is explained
as the equilibrium pressure and temperature
reached in the compressor piping and
equipment volume following a compressor
shutdown
Importance of Settle Out Conditions
The necessity to estimate settle out
conditions,
Figure 1. Schematic of Compression System
1. Settle Out Pressure (SOP) & Settle Out
The assumptions made for this module are,
temperature (SOT) determines the design
1. The compressor loop system is a closed
pressure of the suction scrubber & piping.
loop & no gas has escaped the system.
2. The suction scrubber pressure safety
2. The rate of closure of the suction &
valve’s (PSV) set pressure as well as the
discharge block valve in addition to the
dry gas sealing pressures are decided by
check valve on the discharge side is
the settle out pressure.
neglected.
3. When the compressor reaches settle out
3. The air cooler is assumed to be running at
conditions, process gas is locked inside the
constant duty before and after the
piping and equipment and grips the
compressor is shut down. If the cooler
compressor rotor from rotating effectively
failure occurs due a power trip, then heat
when restarted. Hence depressurizing is
rejection (QCooler = 0) is considered to stop
done by routing the locked gas to a flare,
instantaneously.

Page 48
4. The piping is considered to be adiabatic & T SOT
m S c p , S TS m D c p , D TD QCooler mS mD H P t
mS c p,S mD c p,D
no heat escapes from the equipment &
(1)
piping.
5. The suction scrubber, if considered to have Where,
accumulated liquids, then this volume is Hp t AQ
2
B QN CN 2 (2)
subtracted from the equipment volumes.
1
6. The time delay between the fully closed N Nt (3)
1 216000k t t 0
position & fully open position of the N0 2
2 J
Antisurge valve (ASV) and check valve is
Where,
not considered.
7. When the driver coasts down after a trip, HP(t) = Rate of change of polytropic head as
some amount of residual work is done on the compressor coasts down [kJ/kg/s]
the gas. N(t) = Rate of compressor speed decay
8. Compressor shutdown times are also [rpm/s]
influenced by the fluid resistance, dynamic ms = Suction side gas mass [kg]
imbalance, misalignment between shafts,
leakage and improper lubrication, skewed mD = Discharge side gas mass [kg]
bearings, radial or axial rubbing, Ts = Suction temperature before shutdown [K]
temperature effects, transfer of system
TD = Discharge temp before shutdown [K]
stresses, resonance effect to name a few
and therefore in reality, shutdown times Cp,s = Suction Side Heat Capacity [kJ/kg.K]
can be lower than estimated by the above Cp,D = Discharge Side Heat Capacity [kJ/kg.K]
method.
QCooler = Cooler Duty [kJ/s]
Calculation Methodology
k = Fan Power Law Constant
The lumped parameter methodology applied
to the compressor loop can be depicted as, J = Total Inertia of Compressor System [kg.m2]
The Settle Out Pressure (SOP) can be
estimated
m Z avg R SOT
SOP (4)
MW V1 V2

Where,
m = Total gas mass [kg]
Zavg = Average Compressibility Factor [-]
R = Gas Constant [m3.bar/kmol.K]
MW = Gas Molecular weight [kg/kmol]
SOT = Settle Out Temperature [K]
V1 = Suction side volume [m3]
Figure 2. Gas Compressor Loop V2 = Discharge Side Volume [m3]
Based on the assumptions made, the Settle Case Study
Out Temperature (SOT) can be estimated as,

Page 49
A validation case study is made for a Tank EM /GT Speed 1493 rpm
Vapour compressor in a Gas Compression
Operating Speed 9551 rpm
Plant. Suction pressure exists at 1.05 bara,
Gear Ratio (GR) 6.40 -
540C with a discharge pressure of 5.5 bara,
1280C. The coast down period is calculated Fan law constant (k) 8.38E-05 N.m.min2

initially followed by performing settle out Fan law Constants (k)

calculations. An assumption is made, that the % Speed Speed [rpm]
k
[N.m.min2]
air cooler continues to operate after
105 10029 7.57E-05
shutdown. The compressor maps used is
100 9551 7.00E-05
Table 1. Compressor Performance Curves
Hp Q Q/N Hp/N2 95 9073 6.68E-05

[kJ/kg] [Am3/s] [(Am3/h)/rpm] [kJ/(rpm2)] 90 8596 6.42E-05

136.2 3.0778 0.000322 1.493E-06 80 7641 6.03E-05

133.9 3.4278 0.000359 1.468E-06 70 4776 1.66E-04

130.5 3.6806 0.000385 1.431E-06 Avg. Fan Law constant (k) 8.38E-05

126.6 3.8472 0.000403 1.388E-06 It is to be noted, with the Q vs. Hp curve at

123.6 3.9583 0.000414 1.355E-06 9551 rpm, Fan laws were used to derive the
115.8 4.1111 0.000430 1.269E-06
compressor curves for other speeds, from
70% to 105%.
109.6 4.1806 0.000438 1.201E-06
Table 3. Settle Out Conditions Calculations
100.0 4.2500 0.000445 1.096E-06
Suction Piping Data

Piping Volume 74.55 m3

Gas Mass Density 1.66 kg/m3

Mass Specific Heat 1.83 kJ/kg.K

Gas Temperature 54.1 0C

Comp. Factor (Z1) 0.9875 -

Suction KO Drum %Vol. Liq 20.0 %

Gas Mass- Suction Side 98.82 kg

Discharge Piping Data

Piping Volume 7.87 m3

Gas Mass Density 7.53 kg/m3
Figure 3. Compressor Performance Curves
Mass Specific Heat 2.16 kJ/kg.K
Performing calculations as shown in previous
Gas Temperature 128.3 0C
sections in MS-Excel based on Table 2 & 3,
Comp. Factor (Z2) 0.9622 -
Table 2. Compressor Coast down Input Data
Compressor Design Details Discharge KO Drum %Vol. Liq 10.0 %
Gas Mass – Discharge Side 11.22 kg
Compressor Inertia 376 kg.m2
Cooler Data
Gear Box Inertia 38 Kg.m2
EM Inertia 150.6 kg.m2 Cooler Duty 1432 kW

Total Inertia (J) 380.6 kg.m2 Cooler Outlet Specific Heat 2.03 kJ/kg.K

Page 50
Using the estimated coast down time value of ~35% error. The author attributes the error
115 sec for the case studied, the settle out in SOP partly to the suction & discharge valve
pressure (SOP) & Settle Out Temperature closure time in HYSYS when some vapours
(SOT) is calculated as 0.81 bara, 55.70C & a were discharged & the remaining for the
Settle Out Time of 175 sec. The transient plots reasons explained in the next section.
of the SOP & SOT based on HYSYS simulations Effect of Assumptions on Results
of the case study is as follows,
1. Approximation of compressor curves to
Fan Laws – Fan laws are more applicable
to fluids with low compressibility, smaller
pressure ratios & constant density. Use of
these laws would distort the Compressor
manufacturer’s data thereby causing a
difference in calculations. Since the overlap
area is significant, the performance curve
used in the calculations is assumed to be
same through out the period of coast down.
Figure 4. Compressor Coast down Time
Figure 6 shows the shift in the compressor
The calculated Settle out temperature (SOT)
performance curves.
Trend compared with HYSYS 2006.5 is shown
as follows,

Figure 6. Comparison of Performance Curves

between Fan Laws Generated & Vendor Data

2. Equilibrium conditions during Settle

Figure 5. Settle Out Temperature Trend Out – During coast down, equilibrium
A comparison made between HYSYS conditions are not reached in the
Simulations & the methodology presented compressor plant piping since the system
shows, is dynamic with the gas moving & this is
Table 4. HYSYS vs. Calculated Results tracked in HYSYS 2006.5. However the
Parameter HYSYS Calculated %Error calculations methodology considers
SOT [0C] 58.4 55.7 -4.8 complete equilibrium being reached at
SOP [bara] 0.53 0.81 +34.6
every time step. This causes a difference in
the final settle out temperature (SOT) &
Settle Out Time [s] 167 175 +4.8
settle out pressure (SOP).
The SOT & Settle Out Time shows an error
3. Average Mass Specific Heat Capacity –
margin of < 5%. Whereas for SOP, between
The calculations methodology considers a
the HYSYS predicted value of 0.34 bara and
constant averaged mass specific heat in the
calculated value of 0.81 bara, represents
Page 51
 suction & discharge as well as cooler mS VSuction Side %VSuction Scrubber Liquid Volume S

volumes. However, in commercial solvers (1)

such as HYSYS 2006.5, the mass heat Discharge side gas mass
capacity is computed at every time step mD V Disch arg e Side %V Disch arg e Scrubber Liquid Volume D
which affects the final SOP & SOT.
(2)
4. Density & Z Variations – In the Performing heat balance over the closed loop
calculations made, density and system,
compressibility factor (Z) was assumed to
E In EOut (3)
be constant, whereas HYSYS provides
density & ‘Z’ corrections with change in Or, QSuction Q Disch arg e QCC QCooler (4)
temperature & pressure at every time step. Taking that the energy reaching the gas
Design Standards (API 521/NORSOK) through the compressor is acting only on the
mass of gas enclosed & calculating on a per
1. In designing suction side of compressor
second basis,
piping & equipment, providing a design
margin between settle out pressure and m S c p , S T TS m D c p , D T TD mH P QCooler (5)
design pressure prevents unnecessary
Taking m mS m D & rearranging Eq. (5)
flaring. As per API 521, “Pressure relieving
and Depressuring Systems”, 5th Edition, Jan T SOT
m S c p , S TS m D c p , D TD QCooler mS mD H P t (6)
mS c p,S mD c p,D
2007, “Design Pressure shall be a minimum
The mass specific heat for the cooler in Eq. (6)
of 1.05 times the settle out pressure at
is taken to be an average value between the
maximum pressure drop, calculated
upstream & downstream flow. The polytropic
assuming the suction side is operated at
normal operating pressure and compressor head, H P t is treated as a function of time &
discharge pressure is set to the maximum is calculated by fitting the performance curves
achievable”. (Q vs. Hp).
2
Hp Q Q
2. As per NORSOK P-001, “The maximum A B C (7)
2
operating pressure should be determined N N N
as the settle out pressure occurring at Q
A graph is plotted between (along x-
coincident PAHH” (High-High Pressure N
Alarm) “on both suction side and discharge Hp
axis) & (along y-axis) to obtain the
side, adding a 10% margin for determining N2
design pressure or PSV set pressure”. constants A, B & C, followed by rewriting Eq.
Therefore, NORSOK P-001 standard (E.7) as,
2
provides a more conservative estimate of Hp t AQ B QN CN 2 (8)
settle out pressure since it takes into
In Eq. (8), the compressor speed (N) is
account the highest possible suction &
calculated as shown in Eq. (9)
discharge pressures.
1
ANNEXURE A: SETTLE OUT CONDITIONS N Nt (9)
1 216000k t t 0
DERVIATION 2
2 N0J
The settle out conditions is calculated by
The volumetric flow calculated using Fan
considering the suction & discharge volumes
Laws assuming k1 =k2 during coast down is,
as, Suction side gas mass

Page 52
Qt Nt Applying fan power law as an approximation
(10)
Qt 1 Nt 1 in which ‘k’ is relatively unvarying for a given
Nt Qt curve,
1
Or, Q Qt 1 (11) 60 P kN 3
Nt P N3 P kN 3 ; k N m min 2 P
N3 60
It is to be noted that, the value of ‘Q’ flowing (4)
into the compressor is approximated to value Substituting Eq. (4) in Eq. (3),
of ‘m’ in Eq. (5) (which is constant) since the kN 3 dN
density lies between suction & discharge 2 N 2 J (5)
60 dt
density. The settle out pressure is calculated
Rearranging,
using Ideal Gas equation as,
dN kN 2 N dN kN 2 kg m (6)
Z1 Z 2 dt 2
2
J 60 kg m dt 2
2
J 60 kg m sec
2
n R SOT
2 Integrating Eq. (6), and also multiplying by
P SOP
VTotal
(602) to convert sec2 (rev/s) to min2
(12)
m Z avg R SOT (rev/min)
Or, SOP (13)
MW V1 V2 N N t
dN k 60
t t
dt (7)
ANNEXURE B: COMPRESSOR COAST DOWN N N0 N2 2
2
J t0

DERVIATION N 21
N t
k 60 1
N t
60k t t 0
2
t t0 2
The decay rate of driver speed is governed by 2 1 N0 2 J N N0 2 J

the inertia of the system consisting of the (8)

compressor, coupling, gearbox & driver, 1 1 60k t t 0 1 1 60k t t 0
Nt
1
Nt N0 2
2
J Nt N0 2
2
J 1 60k t t 0
which are counteracted by the torque N0 2
2
J

transferred to the fluid. Neglecting the (9)

mechanical losses, Where, N0 is the compressor speed before
dN
T 2 J [N-m] (1) ESD. The 2nd denominator term exists with
dt
Where, units N.m.min/kg.m2 & is converted to min-1
J = System Inertia (Compressor + gearbox + which gives,
driver) [kg-m2], 1
Nt (10)
where, 1 216000k t t 0
2
JM N0 2 J
J JC 2
Gear Ratio
References & Further Reading
N = Compressor Rotor speed [rpm] or [min-1]
1. www.ogj.com, Volume 113, Issue, 3, Feb
The speed decay rate as well as the system
inertia determines the compressor torque. 2015
Therefore, the power transferred to the gas, is
P 2 NT N m (2)
min
Substituting Eq. (1) in Eq. (2), the power
transferred during (ESD),
dN
P 2 N 2 J (3)
dt
Page 53
 Module 9
Gas Compression Stages – Process Design & Optimization
The following demonstrates how to estimate above design criteria cannot be met,
the required number of compression stages purchaser and vendor shall mutually agree
and optimize the individual pressure ratio in on acceptable bearing metal temperatures.
a multistage centrifugal compression system. 4. During gas recycling, (either by cold
A schematic of a 2-Stage compressor unit is, recycling or hot recycling), the compressor
discharge temperature rises above the
temperature pertaining to normal running
conditions. Quantitatively, the rise in
temperature depends on the pressure ratio
of each stage. The maximum discharge
temperature is typically limited to, in the
range of 1500C to 1600C to avoid damage
Fig 1. Two Stage Compressor Unit to the bearings and seals. To ensure these
A schematic of a 3-Stage Compressor Unit is, limits are not crossed, the compressor
discharge temperature at normal running
conditions must be operated at lower
temperatures with a margin of 200C to
250C. This means typical compressor
discharge temperatures (under normal
running conditions) should be limited to
Fig 2. Three Stage Compressor Unit
the range of 1200C to 1350C.
5. Individual compressor pressure ratios
General Notes
must also be optimized to obtain the
1. When vapours are compressed, its
lowest amount of power required to meet
temperature increases & therefore
the final discharge pressure. This also
requires provisions for gas cooling.
enables to reduce the suction scrubber
2. High gas temperatures can affect lube oil
volumes and air cooler duties to save on
characteristics causing them to carbonize
material and operating costs.
and turn in sludge. This results in fouling
causing the bearing pads and seals to wear Case Study
out and performance degradation. A multistage compression system receives 30
3. As per API 617 (7th Edition, 2002), Clause MMScfd of hydrocarbon vapours at 2 bara,
2.7.1.3, it states, As a design criteria, 300C and is required to be raised to 15 bara.
bearing metal temperatures shall not The Polytropic efficiency [ ] for all LP
exceed 100°C (212°F) at specified compressors is assumed to be 82%. An
operating conditions with a maximum inlet optimization study is performed for a 2-Stage
oil temperature of 50°C (120°F). Vendors and 3-Stage centrifugal compression system.
shall provide bearing temperature alarm The vapour composition is as follows,
and shutdown limits on the datasheets.
However clause No. 2.7.1.3.1 of the said
document also says, In the event that the
Page 54
 Table 1. Gas Composition Therefore considering a maximum number of
Components Mole Fraction [-] stages of 3, for a two stage compressor unit,
the first compressor discharge pressure [P1]
Methane [C1] 0.5232
and Pressure ratio [R] is,
Ethane [C2] 0.3001 15
𝑙𝑛[ ]
2
Propane [C3] 0.1096 𝑛= = 1.83 ~ 2 𝑆𝑡𝑎𝑔𝑒𝑠 (6)
𝑙𝑛[3]
iso-Butane [iC4] 0.0106 15
1⁄
2
𝑅=[ ] = 2.7386 (7)
n-Butane [nC4] 0.0346 2

Iso-Pentane [iC5] 0.0076 𝑃1 = 2 × 2.73861 = 5.48 𝑏𝑎𝑟𝑎 (8)

n-Pentane [nC5] 0.0092 For a three stage compressor unit, the LP

compressor discharge pressure [P1] and MP
n-Hexane [C6] 0.0052
compressor discharge pressure [P2] is,
Total 1.0000 1⁄
15 3
MW [kg/kmol] [PR EoS] 26.53 𝑅=[ ] = 1.9574 (9)
2

Density [1 atm, 15.60C] [kg/m3] 1.128 𝑃1 = 2 × 1.95741 = 3.91 𝑏𝑎𝑟𝑎 (10)

𝑃2 = 2 × 1.95742 = 7.66 𝑏𝑎𝑟𝑎 (11)

Methodology
The number of compressors can be chosen by Using these preliminary values, to arrive at
first estimating preliminary discharge optimized discharge pressures, the following
pressures based on equal pressure ratio as, iterative procedure is adopted.
𝑃𝐿𝑎𝑠𝑡 1. Keeping all preliminary estimated
𝑋𝑛 = [ ] (1)
discharge pressures fixed, the LP
𝑃𝐹𝑖𝑟𝑠𝑡
Where, compressor discharge pressure is varied
PFirst = First Compressor Pressure [bara] for a range to obtain total absorbed power
& total cooler duty of all compressors, and
PLast = Last Compressor Pressure [bara]
sizing each suction scrubber. Making a plot
n = Number of stages [-] of the above values, the discharge pressure
X = Maximum number of Stages [-] corresponding to the lowest duty is chosen
Rewriting the expression, [1st Iteration of LP Compressor].
𝑃𝐿𝑎𝑠𝑡 2. The LP compressor initial estimated
𝑛 × 𝑙𝑛𝑋 = 𝑙𝑛 [ ] (2)
𝑃𝐹𝑖𝑟𝑠𝑡 discharge pressure is now replaced with
𝑃
𝑙𝑛[ 𝐿𝑎𝑠𝑡 ] the 1st Iteration’s optimized pressure.
𝑃𝐹𝑖𝑟𝑠𝑡
Or, 𝑛 = (3) 3. Following further, the MP compressor
𝑙𝑛𝑋
The separation ratio is computed as, discharge pressure is also varied for a
1⁄
given range to similarly obtain an
𝑃𝐿𝑎𝑠𝑡 𝑛
𝑅=[ ] (4) optimized discharge pressure
𝑃𝐹𝑖𝑟𝑠𝑡
corresponding to the lowest total
The intermediate pressure is computed as, compressor duty and cooler duty. [1st
𝑃𝑖 = 𝑃𝑓𝑖𝑟𝑠𝑡 × 𝑅𝑖 (5) Iteration of 2nd stage].

Where, 4. The MP compressor initial estimate

pressure is now replaced with the
Pi = Intermediate Pressure at Stage ‘i’
optimized value, [1st Iteration of 2nd stage].
Page 55
5. With the 1st iteration optimized pressures, Based on the optimized compression ratios,
calculations are repeated similar to Step 2 the savings on the total compressor duty and
Step 3 & Step 4, i.e., 2nd Iteration and so total air cooler duty is 1.59% and 1.68% for 2
forth, until a converged solution is reached. stages respectively. For 3 stages, the
respective savings is 1.86% and 2.03%.
Results
Table 3. Savings on Compressor & Air Cooler Duty
With the procedure applied for the calculated
initial estimates, the optimized results of 2- Parameter 2 Stage 3 Stage
Stage & 3-stage system [LP p = 82%] is, Before Optimization
Table 2. Optimized Compressor Stage Pressures
Total Comp. Duty [kW] 3,000 2,930
Discharge Pressure
Stages-Pressure
Pressure Ratio Total Cooler Duty [kW] 2,786 2,717
- [bara] [-] After Optimization
2 Stage LP [2S-LP] 8.12 4.060 Total Comp. Duty [kW] 2,952 2,876
2 Stage HP [2S-HP] 15.00 1.847 Total Cooler Duty[kW] 2,739 2,663
3 Stage LP [3S-LP] 6.15 3.075 % Savings [Compressor] 1.59% 1.86%
3 Stage MP [3S-MP] 8.25 1.341 % Savings [Air Cooler] 1.68% 2.03%
3 Stage HP [3S-HP] 15.00 1.818
Based on the optimized compression ratios,
The plots of total compressor absorbed the suction scrubber sizes for both cases are,
power, total cooler duty for two stage design Table 4. Suction Scrubber Sizes
and three stage design is as follows, Suction Head Design [2:1 Elliptical]
Scrubber D H Vessel
[H/D = 3.0] [mm] [mm] Volume [m3]
2S-LP/3S-LP 2,400 7,200 34.08

Before Optimization

2S-HP 1,900 5,700 17.11

3S-MP 2,100 6,300 22.98

3S-HP 1,800 5,400 14.59

After Optimization
Fig 3. Two Stages –Total Compressor & Cooler Duty
2S-HP 1,800 5,400 14.59

3S-MP 1,900 5,700 17.11

3S-HP 1,800 5,400 14.59

For 2S-HP & 3S-MP cases, the vessel volume

decreases by 14.7% and 25.5% respectively.
References & Further Reading
1. “Example problems for the calculation and
selection of compressors”, Intech GMBH,
Fig 4. Two Stages – Total Compressor & Cooler Duty
(intech-gmbh.com/compr_calc_and_selec_examples/)

2. www.checalc.com
Page 56
Appendix A

Appendix B

Page 57
 Module 10
Design Considerations for Compressor Antisurge Valve Sizing
Centrifugal Compressors experience a
phenomenon called “Surge” which can be
defined as a situation where a flow reversal
from the discharge side back into the
compressor casing causing mechanical
damage.
The reasons are multitude ranging from
driver failure, power failure, upset process
conditions, start up, shutdown, failure of anti-
surge mechanisms, check valve failure to
operator error to name a few. The
consequences of surge are more mechanical
Figure 1. Performance Curves Operating Limits [1]
in nature whereby ball bearings, seals, thrust
3. To ensure process safety & avoid
bearing, collar shafts, impellers wear out and
mechanical damage, the anti-surge valve
sometimes depending on the how powerful
(ASV) must be large enough to recycle flow
are the surge forces, cause fractures to the
sufficiently. An undersized valve would fail
machinery parts due to excessive vibrations.
to provide enough recycle flow to keep the
The following explains how to size an anti- compressor operating point away from SCL
surge valve for a single stage VSD system for and SLL. Whereas over sizing the ASV leads
Concept/Basic Engineering purposes. to excess gas recycling that can drive the
General Notes & Assumptions compressor into the choke flow region.
1. Centrifugal compressors are characterized Oversized valves also create difficulties in
by “Performance curves” which are a plot tuning the controllers due to large
of Actual Inlet Volumetric Flow rate [Q] vs. controller gain values and limited stroke.
Polytropic head [Hp] for various operating
speeds. The operating limits for
performance curves are the surge line and
the choke flow line, beyond which any
compressor operation can cause severe
mechanical damage.
2. Below is an image of performance curves
characteristics which indicates the surge
flow line and choked flow line, both of
which extend from the minimum speed Q
Figure 2. Sizing Criteria for Anti-surge Valve
vs. Hp curve to the maximum speed Q vs. Hp
curve. The surge curve is defined as the 4. To size the anti-surge valve (ASV), the
Surge Limit Line [SLL] and an operating philosophy employed should consider,
margin is provided [e.g., 10% on flow rate] operating the compressor on the right
which is called the surge control line [SCL]. hand side of the SCL while also ensuring
the operating point does not cross the
Page 58
 choke flow line. Towards this, the recycle Anti-Surge Valve Sizing Methodology
flow rates across the ASV can be taken to To size the anti-surge valve, the ANSI/ISA
be 1.8 to 2.2 times the surge flow rate. S75.01 compressible fluid sizing expression is
5. Traditionally ASVs have linear opening chosen for this exercise and the flow rates are
characteristics, though sometimes equal taken for at least 1.8 to 2.2 times the surge
percentage characteristics can be flow rate.
incorporated into the linear trend. Quick Step 1: Piping Geometry Factor (Fp)
opening characteristics are not preferred −1⁄
∑𝐾 𝐶𝑉 2 2
due to poor throttling characteristics while 𝐹𝑃 = [1 + ( 2) ] (1)
890 𝑑
Equal percentage valves suffer from slow
Where,
opening during the early travel period.
Fp = Piping geometric factor [-]
6. The stroking time of the valve should be
Cv = Valve coefficient [-]
ideally less than 2 sec with less than 0.4 sec
time delay and no overshoot. The actuator d = Control valve size [inch]
response time must be less than 100 msec K = Sum of pipe resistance coefficients [-]
and the noise limit is ~85 dBA. The
The value of Fp is dependent on the fittings
maximum noise level allowed is 110 dBA.
such as reducers, elbows or tees that are
7. Anti-surge valves are Fail-open [FO] type directly attached to the inlet & outlet
and should provide stable throttling. Fluid connections of the control valve. If there are
velocities should be less than 0.3 Mach to no fittings, Fp is taken to be 1.0. The term K
avoid piping damage and valve rattling. is the algebraic sum of the velocity head loss
8. The anti-surge valve can be operated coefficients of all the fittings that are attached
pneumatically or by solenoid action. For to the control valve & is estimated as,
valve sizes greater than 16”, a motor ∑ 𝐾 = 𝐾1 + 𝐾2 + 𝐾𝐵1 −𝐾𝐵2 (2)
operated valve can be used to effectuate
Where,
the fast opening requirements.
K1 = Upstream fitting resistance coefficient [-]
9. Although the current module provides a
K2 = Downstream resistance coefficient [-]
methodology to size an ASV which is
suitable during Concept/Basic Engineering KB1 = Inlet Bernoulli Coefficient [-]
stage, a compressor dynamic simulation KB2 = Outlet Bernoulli Coefficient [-]
shall be performed with the actual plant Where,
layout based on detailed design to verify if 𝑑 4
𝐾𝐵1 = 1 − (𝐷 ) (3)
the ASV can cater to preventing a surge 1
𝑑 4
during start-up & shutdown scenarios. 𝐾𝐵2 = 1 − (𝐷 ) (4)
2

10. The final ASV size must be verified and Where,

arrived in concurrence with the D1 = Inlet Pipe Inner Diameter [in]
turbomachinery vendor, valve
D2 = Outlet Pipe Inner Diameter [in]
manufacturer, if the ASV can cater to the
surge control philosophy employed, slope The most commonly used fitting in control
of the performance curves and polytropic valve installations is the short-length
efficiency maps at the choke points. concentric reducer. The expressions are as
follows,
Page 59
 𝑑2
2 X = Pressure Drop Ratio [-]
𝐾1 = 0.5 × [1 − ( 2 )] , for inlet reducer (5)
𝐷1
2
Z = Gas compressibility Factor [-]
𝑑2
𝐾2 = 1.0 × [1 − (𝐷2 )] , for outlet reducer T1 = Inlet Temperature [ K]
2
(6)
Fk = Gas specific heat to air specific heat ratio
Step 2: Calculate Valve Coefficient (Cv)
k1 = Gas specific heat ratio at valve inlet [-]
To calculate the valve Cv, the following
XTP and XT = Pressure drop ratio factor [-]
ANSI/ISA expression is used.
MW = Molecular Weight of gas [kg/kmol]
𝑀
𝐶𝑣 = 𝑋×𝑀𝑊
(7)
To estimate the compressor mass flow rate
𝑁8 𝐹𝑝 𝑃1 𝑌√
𝑇1 ×𝑍
from the suction density [ s] and compressor
𝑋=
∆𝑃
(8) actual inlet flow rate, it can be estimated as,
𝑃1
𝑃×𝑀𝑊
𝑋 𝜌𝑠 = (14)
𝑍×𝑅×𝑇
𝑌 =1− (9)
3×𝐹𝑘 ×𝑋𝑇
𝑀 = 𝑄𝑠 × 𝜌𝑠 (15)
𝑘1
𝐹𝑘 = (10) Where,
1.4

If X > Fk XT, then flow is Critical. R = Gas Constant [0.0831447 m3.bar/kmol.K]

If X < Fk XT, then flow is Subcritical. Qs = Compressor Suction Vol flow rate [m3/h]

For Critical flow, the value of ‘X’ is replaced To arrive at a converged value of Fp, the valve
with Fk XT and the gas expansion Factor [Y] Cv at each iteration, can be computed
and valve coefficient [Cv] is to be computed as, iteratively by replacing the Fp value in each
𝐹𝑘 ×𝑋𝑇
iteration of the Cv equation. Applying the
𝑌 =1− = 0.667 (11) Sizing method, to the four points shown in
3×𝐹𝑘 ×𝑋𝑇
Figure 2, the various sizing scenarios are,
𝑀
𝐶𝑣 = (12)
𝐹 ×𝑋 ×𝑀𝑊
0.667×𝑁8 𝐹𝑝 𝑃1 √ 𝑘 𝑇 a. Minimum Speed - Surge Flow [Q1]
𝑇1 ×𝑍
b. Minimum Speed - Surge Flow [Q1 1.8]
If the control valve inlet and outlet piping is
provided with reducers and expanders, then c. Minimum Speed - Surge Flow [Q1 2.2]
the value of XT is replaced with XTP as follows, d. Maximum Speed - Surge Flow [Q2]
−1 e. Maximum Speed - Surge Flow [Q2 1.8]
𝑋𝑇 𝑋𝑇 (𝐾1 +𝐾𝐵1 ) 𝐶𝑣 2
𝑋𝑇𝑃 = 2 × [1 + (𝐷2 ) ] (13)
𝐹𝑝 1000 1 f. Maximum Speed - Surge Flow [Q2 2.2]
Where, g. Minimum Speed - Choke Flow [Q3]
Cv = Cv value at Valve 100% Open [-] h. Maximum Speed - Choke Flow [Q4]
M = Mass Flow Rate [kg/h] The ASV Cv computed for the surge points
N8 = Constant [Value = 94.8] would be closer to each other in most cases.
Fp = Piping Geometry Factor [-] Similarly, the ASV Cv at the choke points
would also be closer to each other. Therefore,
P = Pressure drop across ASV [bar]
to arrive at conservative results, the higher of
P1 = Inlet Pressure [bara] the Cv values at the surge points & the lower
Y = Gas Expansion Factor [-]
Page 60
of the Cv values at the choke points are to be The surge control line [SCL] chosen for this
considered to determine a suitable ASV size. exercise is taken as 10% on the surge flow
rate at each speed and is as follows,
Case Study
Table 2. Surge Control Line [SCL] Parameters
68.1 MMscfd of hydrocarbon gas at 11.61
bara [suction flange conditions] and 47.470C Speed Surge Flow 10% HP
is to be compressed to 30.13 bara pressure [rpm] [Act_m3/h] [m]
[discharge flange conditions]. The 7,532 2,952 6,721
compressed gas is cooled to 50 C via an air
0
10,043 4,184 12,297
cooler. The centrifugal compressor used is a
12,544 6,363 19,263
variable speed configuration. The gas
composition is as follows, 13,182 7,118 21,077

Table 1. Gas Composition The Gas Properties are as follows for the
Parameter Mol % suction and discharge flange conditions,
Table 3. Gas Properties at Flange Conditions
Methane [CH4] 94.09
Ethane [C2H6] 0.03 Parameter Value Units

Propane [C3H8] 0.02 Gas MW 16.81 kg/kmol

Nitrogen [N2] 3.93 Suction Pressure 11.61 bara

Carbon Dioxide [CO2] 0.96 Suction Temperature 47.5 0C

Water [H2O] 0.97 Discharge Pressure 30.13 bara

Total 100 Discharge Temperature 143.0 C

Inlet Z [Z1] 0.9810 -
The compressor performance curves for
various operating speeds are as follows, Outlet Z [Z2] 0.9848 -
Specific Heat of Gas - Inlet 1.3229 -
Suction Density 7.464 kg/m3
Discharge Density 14.868 kg/m3
Actual Volumetric Flow 7,611 Am3/h
Inlet Mass Flow 56,809 kg/h

The compressor parameters are as follows,

Table 4. Compressor Parameters

Parameter Value Units

Figure 3. Compressor Performance Curves
Adiabatic Head 16,887 m
The upstream and downstream piping for the
anti-surge line is taken as NPS 4”, Ref [2] with Polytropic Head 17,333 m
a thickness of 0.237 inches for this exercise. Adiabatic Efficiency 77.61 %
The anti-surge valve chosen to be checked is a Polytropic Efficiency 79.71 %
NPS 4” valve [OD 4.5”] [Single ported, Cage Power Consumed 3,365 kW
Guided, Globe Style Valve body] with a Cv of
Polytropic Head Factor 1.0009 -
236 and corresponding XT value of 0.69.
Polytropic Exponent 1.3839 -
Isentropic Exponent 1.2881 -

Page 61
ASV Sizing Solution The gas specific heat ratio to air specific heat
Proceeding with the Cv calculation for the ratio is calculated as,
case of Minimum Speed - Surge Flow [Q1], 𝐹𝑘 =
1.3348
= 0.9534 (26)
1.4
4 4
𝐾𝐵1 = 1 − ( ) = 0.0256 (16) The pressure drop ratio factor [XT] is,
4.026
4 5.48
4 𝑋= = 0.321 (27)
𝐾𝐵2 = 1 − ( ) = 0.0256 (17) 17.09
4.026
2 Since the valve construction details are
42
𝐾1 = 0.5 × [1 − ( )] = 0.000083 (18) available, XTP is used instead of XT.
4.0262
−1
42
2 0.69 0.69(0+0.0256) 236 2
𝐾2 = 1.0 × [1 − ( )] = 0.00017 (19) 𝑋𝑇𝑃 = × [1 + ( ) ] (28)
4.0262 12 1000 42

∑ 𝐾 = 0.000083 + 0.00017 + 0.26 − 0.26 (20) 𝑋𝑇𝑃 ≈ 0.69 (29)

∑ 𝐾 = 0.00025 (21) Checking for flow condition,
−1⁄
0.00025 236 2 2 𝐹𝑘 × 𝑋𝑇𝑃 = 0.9534 × 0.69 = 0.6579 (30)
𝐹𝑃 = [1 + ( 2
) ] =1 (22)
890 4 Since X < Fk XTP, flow is Subcritical.
The flow rate for the minimum speed - surge The gas expansion factor is estimated as,
flow is 2,683 Am3/h and gas density at 0.321
compressor inlet is, 𝑌 =1− = 0.8374 (31)
3×0.9534×0.69

𝜌𝑠 =
11.61×16.81
≈ 7.464
𝑘𝑔
(23) Therefore the ASV Cv is computed as,
0.981×0.0831447×320.62 𝑚3
20,028
𝑀 = 2,683 × 7.464 ≈ 20,028
𝑘𝑔
(24)
𝐶𝑣 = 0.321×16.81
(32)
ℎ 94.8×1×17.09×0.8374√
323.15×0.9732
The compressor discharge flange pressure is
𝐶𝑣 = 112.7 (33)
17.44 bara at minimum speed surge flow of
2,683 Am3/h and a discharge air cooler which Re-inserting the value of Cv = 112.72 into the
offers a pressure drop for the flowing gas. Fp expression to iterate, the value of Cv
Taking a max P of 0.35 bar across the becomes,
discharge side, the ASV inlet pressure 0.00025 112.7 2
−1⁄
2
becomes, 17.44 – 0.35 = 17.09 bara. The 𝐹𝑃 = [1 + ( ) ] = 0.9999 (34)
890 42
cooler discharge temperature is 500C, 20,028
𝐶𝑣 = (35)
therefore neglecting heat losses; the ASV inlet 94.8×0.9999×17.09×0.8375√
0.321×16.81
323.15×0.9732
temperature also is at 500C.
𝐶𝑣 = 112.719~113 (36)
Making an approximation that the ASV
discharge side piping and compressor suction Therefore with another iteration the Cv value
side P is negligible; the ASV outlet pressure remains nearly the same at 112.72 ~ 113.
is nearly equal to the compressor inlet The ASV Cv can now be estimated for the case
pressure. Therefore the ASV outlet pressure of Q 1.8 at Cv,min and Q 2.2 at Cv,max.
becomes 11.61 bara. The ASV P is,
𝐶𝑣,𝑚𝑖𝑛 = 1.8 × 113 = 203 (37)
∆𝑃 = 17.09 − 11.61 = 5.48 𝑏𝑎𝑟 (25)
The ASV Inlet Z & k1 value [Cp/Cv] at 17.09 𝐶𝑣,𝑚𝑎𝑥 = 2.2 × 113 = 248 (38)
bara and 500C is 0.9732 and 1.3348. Performing similar calculations for all cases,
Page 62
 Table 5. ASV Sizing Cases – Surge Points
X 0.195 0.537 -
Min Max
Parameter Units
Surge Surge XTP 0.690 0.690 -
Qs 2,683 6,471 Am3/h Flow Condition Subcritical Subcritical -
s 7.46 7.46 kg/m3
Cv, Choke 286 229 -
M 20,028 48,295 kg/h
From the Cv values calculated, the governing
PD 17.44 36 bara
case becomes the Min Speed surge point case.
Discharge P 0.35 0.35 bar
ASV Inlet P1 17.09 35.99 bara
𝐶𝑣,𝑚𝑖𝑛 = 203 ≤ 𝐶𝑣 = 236 ≤ 𝐶𝑣 = 248 (39)
ASV Outlet P2 11.61 11.61 bara Hence the selected 4” control valve with a Cv
ASV P 5.48 24.38 bar of 236 and XT of 0.69 is adequately sized to
provide anti-surge control.
Cooler Outlet T 323.15 323.15 0K

ASV Inlet Z 0.9732 0.9465 - Transient Study to Verify ASV Sizing

With the ASV size selected, a transient study
Cp/Cv-ASV Inlet 1.3348 1.3781 -
is performed to check for ASV adequacy.
XT 0.69 0.69 -
Centrifugal compressors during shutdown
Fk - ASV Outlet 0.9534 0.9843 - experience surging & the ASV must be able to
X 0.321 0.677 - provide sufficient cold recycle flow to keep
XTP 0.690 0.690 - the operating point away from the SLL as the
Flow Condition Subcritical Subcritical - compressor coasts down.

Cv, Min 113 110 - Normal shutdown [NSD] refers to a planned

event where the anti-surge valve is opened
Cv, Min [Q x 1.8] 203 198 -
first by 100%, prior to a compressor trip.
Cv, Max [Q x 2.2] 248 242 -
An emergency shutdown [ESD] is an
Table 6. ASV Sizing Cases – Choke Points
unplanned event, where for example, upon
Min Max
Parameter
Choke Choke
Units loss of driver power, the ASV opens quickly to
Qs 4,805 9,102 Am3/h
recycle flow and prevent the operating point
from crossing the SLL during coast down. For
s 7.46 7.46 kg/m3
this module, the ESD case considered is
M 35,860 67,932 kg/h “Driver trip” where the compressor driver
PD 14.77 25.45 bara experiences a sudden loss of power.
Discharge P 0.35 0.35 bar To simulate the transient case, the air cooler
ASV P1 14.42 25.10 bara and suction scrubber can be sized with
ASV P2 11.61 11.61 bara preliminary estimates to cater to maximum
2.81 13.49 bar
speed choke flow case.
ASV P
Cooler T 323.15 323.15 0K Suction Scrubber Volume

ASV Inlet Z 0.9769 0.9615 -

Using GPSA K-Value method for suction
scrubber sizing, Ref [3], for a flow rate of
Cp/Cv 1.3279 1.3511 -
67,932 kg/h and 11.61 bara operating
XT 0.69 0.69 -
pressure, the suction scrubber size is H D of
Fk 0.9485 0.9651 - 6.9m 2.3m with an ellipsoidal head and

Page 63
inside dish depth of 0.25m. The total scrubber
volume is 30.1 m3.
Air Cooler Volume
Similarly, the air cooler is sized for maximum
speed choke flow case, Ref [4], for a flow rate
of 67,932 kg/h & duty of 4,351 kW. The
overall heat transfer coefficient [U] is
assumed to be 25 W/m2.K. The inlet
temperature is 1420C which is cooled to 500C Figure 4. Compressor Coast down Time
with an air side temperature of 350C. The air From the curve, the compressor is expected
cooler geometry chosen for this exercise is a to reach a standstill in ~124 sec.
single tube pass with 3 tube rows & each tube
ESD and NSD Analysis
is 9.144m in length. The fan & motor
With the equipment volumes, ASV Cv chosen
efficiencies are taken as 75% and 95%
and compressor speed decay rate imposed, an
respectively. With this data, the air cooler has
ESD and NSD analyses is performed to track
a tube OD of 1” [0.0254m] & total number of
operating point during coast down.
tubes of 307 [Tube volume of 1.423 m3].
Compressor Coast down
Coast down time is influenced by a number of
factors including fluid resistance, dynamic
imbalance, misalignment between shafts,
leakage and improper lubrication, skewed
bearings, radial or axial rubbing, temperature
effects, transfer of system stresses, resonance
effect to name a few and therefore in reality,
shutdown times can be lower than estimated Figure 5. ESD/NSD Operating Point Migration
by the method shown below. From the analysis made, it is seen that the
The decay rate of driver speed is governed by selected ASV size of 4” [Cv 236] is sufficient to
the inertia of the system consisting of the prevent a surge during ESD and NSD.
compressor, coupling, gearbox & driver, References & Further Reading
which are counteracted by the torque
1. “Development and Design of Antisurge and
transferred to the fluid. Neglecting the
Performance Control Systems for
mechanical losses, the compressor speed
Centrifugal Compressors”, Mirsky S.,
decay rate can be estimated as, McWhirter J., Jacobson W., Zaghloul M.,
1
𝑁[𝑡] = (42) Tiscornia D., 3rd M.E Turbomachinery
1 216,000×𝑘×[𝑡−𝑡𝑜 ]
+ [2𝜋]2 ×𝐽 Symposium, Feb 2015
𝑁𝑜

Where, ‘N0’ is the compressor speed before 2. Control Valve Handbook, Emerson, 5th Ed.
ESD, ‘J’ is the total system inertia & ‘t0’ is time 3. https://checalc.com/calc/vertsep.html
at which the ESD is initiated. For this exercise 4. https://checalc.com/calc/AirExch.html
the total system inertia is taken as 108 kg.m2. 5. https://www.slideshare.net/VijaySarathy7
The coast down speed calculated is, /centrifugal-compressor-settle-out-
conditions-tutorial
Page 64
Appendix A

Page 65
 Module 11
Process Design for Natural Gas Transmission
Compressor stations form a key part of the During transmission, gas pressure drops for
natural gas pipeline network that moves which a booster station is installed en-route.
natural gas from individual producing well The minimum pressure required at the
sites to end users. As natural gas moves booster station is 10 bara & the ambient
through a pipeline, distance, friction, and pressure and temperature is 1.01325 bara &
elevation differences slow the movement of 300C. The design pressure & design
the gas, and reduce pressure. Compressor temperature of the pipeline-compressor
stations are placed strategically within the transmission system is 40 bara & 2000C.
gathering and transportation pipeline The requirement to be met for pipeline wall
network to help maintain the pressure and stresses is ASME B31.8. As per ASME B31.8,
flow of gas to market. The following is a the Design factor [F], Temperature De-rating
module to perform process design of a [T], Longitudinal Joint Factor [E] for the
natural gas transmission system. chosen pipeline joining methods is as follows,
Problem Statement Table 2. Reference Mechanical Design Parameters
80 MMscfd of Natural Gas at 36 bara & 480C
is Design Factors [F] - Gas Pipeline Location
to be transmitted from a gas plant in a desert
Class Description F
region to a city power station located 50 km
away. The gas composition & critical Class 1, Div 1 Deserted 0.80

properties are as follows, Class 1, Div 2 Deserted 0.72

Table 1. Natural Gas Composition & Properties Class 2 Village 0.60

Component MW Mol% Pc,i Tc,i Class 3 City 0.50

Class 4 Densely Populated 0.40

- kg/kmol % psia 0R

Temperature De-rating [T] for Gas Pipelines

Methane 16.04 76.23 667.8 343.0
T [0F] T [0C] T
Ethane 30.07 10.00 707.8 549.8
250 120 1.00
Propane 44.01 5.00 616.3 665.7
300 150 0.97
i-Butane 58.12 1.00 550.7 765.3
350 175 0.93
n-Butane 58.12 1.00 529.1 734.6
400 200 0.91
i-Pentane 72.15 0.30 490.4 828.7
450 230 0.87
n-Pentane 72.15 0.10 488.6 845.3
Abbreviation Joining Method E
n-Hexane 86.18 0.05 436.9 913.3
SMLS Seamless 1.0
C 7+ 119.00 0.05 453.0 1116.0
ERW Electric Resistance Weld 1.0
H2O 18.02 0.25 3206.2 1164.9 EFW Electric Flash Weld 1.0
CO2 44.01 3.00 1071.0 547.5 SAW Submerged Arc Weld 1.0
H2S 34.08 0.02 1306.0 672.3 BW Furnace Butt Weld 0.6
N2 28.01 3.00 493.0 226.97 EFAW Electric Fusion Arc Weld 0.8

Page 66
The pipeline specification requirement is API The pipeline would include fittings and
5L plain end line pipe specifications ranging elevational differences which offer a pressure
from 6” ND to 80” ND. The product pipeline drop & hence can be expressed as an
specification (PSL) with its respective equivalent length. For this module, the
Specified Minimum Yield Strength (SYMS) to fittings & elevation losses are taken as 2% of
be used as per API 5L are PSL 1 and PSL 2. the total pipeline length. The booster
The pipeline grades are as follows, compressor station is placed at a distance of
Table 3. Product Specification Level (PSL) 20km from the gas plant located in the desert
area & the downstream pipeline travels
SMYS SMYS
Grade Grade another 30 km to reach the power station in
MPa MPa the city. This is shown as follows,
PSL 1 Gr A25 172 PSL 2 Gr B 241 Table 4. Pipeline Lengths
PSL 1 Gr A 207 PSL 2 X42 290 Total Eq.
Pipeline Length Eq. Length
Length
PSL 1 Gr B 241 PSL 2 X46 317 - [m] [%] [m] [m]
PSL 1 X42 290 PSL 2 X52 359 Upstream 20,000 2 400 20,400
PSL 1 X46 317 PSL 2 X56 386 Downstream 30,000 2 600 30,600
PSL 1 X52 359 PSL 2 X60 414
It is assumed that the condensation in the
PSL 1 X56 386 PSL 2 X65 448 pipeline is minimal and hence the pipeline
PSL 1 X60 414 PSL 2 X70 483 efficiency [Ep] is taken as 0.92. To evaluate
PSL 1 X65 448 PSL 2 X80 552
the maximum hydrostatic test pressure, the
difference in elevation of the pipeline
PSL 1 X70 483 - -
between the pipeline high point elevation and
In the current module, the API 5L pipeline elevation at test point for upstream and down
grades chosen for both desert location [Class stream pipelines is taken as 100m and 70m &
1, F = 0.72] & city location [Class 3, F = 0.50] 100m and 90m respectively.
is PSL 1 X65 [SMYS = 448 Mpa]. The pipeline The pipeline booster station consists of a
joining method chosen is Electric Resistance centrifugal compressor operating with
Weld [ERW] with a longitudinal joining factor polytropic efficiency [ p] of 80% at 80
[E] of 1.0. The temperature de-rating factor MMscfd. The minimum pressure required at
[T] of the pipelines before & after the booster the compressor inlet is 12 bara & a minimum
station for a design temperature of 2000C as pressure of 16 bara at the city power station.
per Table 2 is 0.91. The design capacity of the
pipeline is taken as 100 MMscfd. For the Design Methodology
module, the pipelines before & after the To perform a process design of the pipeline &
booster station is laid below ground that has booster compressor station, the design
a constant soil temperature of 300C & soil methodology consists of 3 parts – Process
overall heat transfer coefficient [U] of 35 design of upstream pipeline from the gas
W/m2.K. The pipeline corrosion allowance plant from the desert, process design of the
before & after the booster station is taken as gas compressor at the booster station and
3 mm considering a corrosion rate of 5 downstream pipeline to the city power
mils/year [1 mil = 1/1000th of an inch] over a station. The process design steps for the
25 year operating period. upstream/downstream pipelines are,
Page 67
1. Estimation of Pipeline wall thickness based polytropic efficiency [ p], adiabatic
on design pressure [DP], design efficiency [ a], molar density [[ m], molar
temperature [DT], design flowrate [Qd], volume [Vm], mass flow [m], Molar flow
location class, design factor [F], pipeline [M], Polytropic Head [Hp] & Absorbed
specification [API 5L], specified minimum Power [P].
yield strength [SMYS], derating factor [T]. 3. Check if compressor energy balance
2. Estimation of Mixture Pseudocritical condition, P1V1n-P2V2n = 0 is satisfied.
Properties – Pseudo critical Pressure [Tc], 4. Repeat above steps for compressor
Pseudo critical Temperature [Tc], Reduced discharge side for process parameters.
Pressure [Ppr], Reduced Temperature [Tpr],
Reduced Density [ r], Deviation Parameter Property Estimation Methodology
To assess the properties of natural gas,
[ ], Modified Reduced Pressure [P’pc],
calculations can be begun by estimating the
Modified Reduced Temperature [T’pc],
properties using Kay’s Mixing Rule as follows,
Specific Heat Capacity [Cp], Gas
Compressibility Factor [Z]. Mixture molecular weight [MW], kg/kmol
3. Estimation of gas mixture density [ ], 𝑀𝑊 = ∑ 𝑦𝑖 𝑀𝑊𝑖 (1)

mixture molecular weight [MW], mass flow Mixture Pseudo Critical Pressure [Pc], psia
[m] and actual volumetric flow rate [Q]. 𝑃𝑐 = ∑ 𝑦𝑖 𝑃𝑐,𝑖 (2)
4. Estimation of upstream/downstream Mixture Pseudo Critical Temperature [Tc], 0R
pipeline Fluid Velocity [V], pipeline exit
𝑇𝑐 = ∑ 𝑦𝑖 𝑇𝑐,𝑖 (3)
temperature [Pe] based on soil/ambient
temperature, overall heat transfer co- Gas Specific Gravity [ g], [-]
𝑀𝑊𝑔
efficient [U], pipeline pressure drop [ P], 𝛾𝑔 = 𝑀𝑊 ; MWair = 28.96 kg/kmol (4)
𝑎𝑖𝑟
Pipeline Exit Pressure & pressure drop per
From the above, Kay’s Mixing Rule does not
km [ P/L].
give accurate pseudocritical properties for
5. Estimation of Maximum Allowable higher molecular weight mixtures
Operating Pressure [MAOP], Test Pressure (particularly C7+ mixtures) of hydrocarbon
at 110% of SMYS, Maximum Hydrostatic gases when estimating gas compressibility
Test Pressure & Leak Test Pressure. factors [Z] and deviations can be as high as
For the Booster Compressor Station, the 15%. Therefore, to account for these
design steps are, differences, Sutton’s correlations based on
1. Estimation of Mixture Critical Properties – gas specific gravity can be utilized as follows,
Pseudocritical Pressure [Tc], Pseudocritical 𝑃𝑝𝑐 = 756.8 − 131.07𝛾𝑔 − 3.6𝛾𝑔2 (5)
Temperature [Tc], Reduced Pressure [Ppr],
𝑇𝑝𝑐 = 169.2 + 349.5𝛾𝑔 − 74.0𝛾𝑔2 (6)
Reduced Temperature [Tpr], Reduced
Density [ r], Deviation Parameter [ ], The above equations are valid for the gas
Modified Reduced Pressure [P’pc], Modified specific gravities range of 0.57 < g < 1.68.
Reduced Temperature [T’pc], Specific Heat Using the Sutton correlations, the reduced
Capacity [Cp], Gas Compressibility [Z]. properties are calculated as,
𝑃
2. Estimation of Adiabatic Exponent [k], 𝑃𝑟 = (7)
𝑃𝑝𝑐
compressor inlet gas mixture density [ ], 𝑇
average polytropic exponent [n] based on 𝑇𝑟 = (8)
𝑇𝑝𝑐

Page 68
However the pseudocritical properties are And the constants A1 to A11, are as follows,
not the actual mixture critical temperature Table 5. DAK EoS A1 to A11 Constants
and pressure but represent the values that
A1 0.3265 A7 –0.7361
must be used for the purpose of comparing
A2 –1.0700 A8 0.1844
corresponding states of different gases on
the compressibility factor, Z-chart/Gas A3 –0.5339 A9 0.1056

Deviation Factor, as shown below in the A4 0.01569 A10 0.6134

Standing & Katz, 1959 chart for natural gases. A5 –0.05165 A11 0.7210

A6 0.5475

DAK-EoS has an average absolute error of

0.486% in its equation, with a standard
deviation of 0.00747 over ranges of pseudo-
reduced pressure and temperature of 0.2 <
Ppr < 30; 1.0 < Tpr < 3.0 and for Ppr < 1.0 with
0.7 < Tpr < 1.0. However DAK-EoS gives
unacceptable results near the critical
temperature for Tpr = 1.0 and Ppr >1.0, and
DAK EoS is not recommended in this range.
DAK EoS for NG Mixtures with Acid Gases
Natural Gas is expected to contain acid gas
fractions, such as CO2 and H2S, & applying the
Standing & Katz Z-factor chart & Sutton’s
pseudocritical properties calculation methods
would yield inaccuracies, since they are only
valid for hydrocarbon mixtures. To account
Figure 1. Natural Gas deviation factor chart
(Standing & Katz, 1959) for these inaccuracies, the Wichert & Aziz
Due to the graphical method of Standing & correlations can be applied to mixtures
Katz chart, the Z factor can be estimated using containing CO2 < 54.4 mol% & H2S < 73.8
Dranchuk and Abou-Kassem Equation of State mol% by estimating a deviation parameter
[DAK-EoS] which is based on the data of [ ], which is used to modify the pseudocritical
Standing & Katz, 1959 and is expressed as, pressure & temperatures. The deviation
𝐴2 𝐴3 𝐴4 𝐴5 parameter [ ] whose units are in 0R, is,
𝑍 = 1 + [𝐴1 + + + + ] 𝜌𝑟 +
𝑇𝑟 𝑇𝑟3 𝑇𝑟4 𝑇𝑟5 𝜀 = 120[𝐴0.9 − 𝐴1.6 ] + 15[𝐵 0.5 − 𝐵 4 ] (11)
𝐴7 𝐴8 𝐴7 𝐴8
[𝐴6 + + ] 𝜌𝑟2 − 𝐴9 [ + ] 𝜌𝑟5 + Where,
𝑇𝑟 𝑇𝑟2 𝑇𝑟 𝑇𝑟2
𝜌𝑟2 2 A = YCO2 + YH2S in Gas mix [Y = mole fraction]
+𝐴10 (1 + 𝐴11 𝜌𝑟2 ) ( ) 𝑒 −𝐴11𝜌𝑟 (9)
𝑇𝑟3
B = YH2S in Gas mixture [Y = mole fraction]
Where,
0.27𝑃𝑟
Applying [ ], the modified pseudocritical
𝜌𝑟 = (10) pressure & temperature is,
𝑍𝑇𝑟
′
𝑇𝑝𝑐 = 𝑇𝑝𝑐 − 𝜀 (12)
r = Pseudo-Reduced Density [-]
′
𝑃𝑝𝑐 𝑇𝑝𝑐
′
Tr = Pseudo-Reduced Temperature [-] 𝑃𝑝𝑐 = (13)
𝑇𝑝𝑐 −𝐵[1−𝐵]𝜀

Page 69
Where, T’pc & P’pc are valid only in 0R and psia. wastelands, deserts, rugged mountains,
Based on the calculated modified grazing land, farmland, sparse populations.
pseudocritical pressure [P’pc] and 2. Class 1, division 1 Location – A Class 1
temperature [T’pc], the pseudo-reduced location where the design factor, F, of the
pressure [Pr] & temperature [Tr] is, pipeline is greater than 0.72 but equal to,
𝑃 [𝑝𝑠𝑖𝑎] or less than 0.80 and which has been
𝑃𝑝𝑟 = ′ [𝑝𝑠𝑖𝑎] (14)
hydrostatically tested to 1.25 times the
𝑃𝑝𝑐

𝑇 [°𝑅 ]
maximum operating pressure.
𝑇𝑝𝑟 = ′ [° ] (15)
3. Class 1, division 2 Location - This is a
𝑇𝑝𝑐 𝑅

0.27𝑃𝑝𝑟 Class 1 location where the design factor, F,

𝜌𝑝𝑟 = (16) of the pipeline is equal to or less than 0.72,
𝑍𝑇𝑝𝑟
and which has been tested to 1.1 times the
Using the calculated values of Ppr Tpr & pr,
maximum operating pressure.
compressibility factor, Z is determined by
4. Class 2 Location - This is any 1-mile
using DAK EoS. Owing to the value of ‘Z’ being
section of pipeline that has more than 10
an implicit parameter in calculating pr as
but fewer than 46 buildings intended for
well as in DAK-EoS, an iterative approach,
human occupancy including fringe areas
whereby Z value is guessed & iteratively
around cities and towns, industrial areas,
solved to satisfy both modified pseudo-
and ranch or country estates.
reduced density [ pr] & DAK EoS. Upon
calculating the value of Zin at the pipeline 5. Class 3 Location - This is any 1-mile
inlet, the actual volumetric flow rate [Qin], Gas section of pipeline that has 46 or more
density [ in], gas mass flow [m] is calculated buildings intended for human occupancy
as follows, except when a Class 4 Location prevails,
including suburban housing developments,
Actual Volumetric Flow Rate, Am3/h
shopping centres, residential areas,
𝑃𝑠𝑡𝑑𝑄𝑠𝑡𝑑 𝑍1 𝑇1
𝑄1 = (17) industrial areas & other populated areas
𝑍𝑠𝑡𝑑 𝑇𝑠𝑡𝑑 𝑃1
not meeting Class 4 Location requirements
The value of Zstd is taken to be close to 1.0.
6. Class 4 Location - This is any 1-mile
Gas Density, kg/m3 section of pipeline where multi-storey
𝑃1 𝑀𝑊 buildings are prevalent, traffic is heavy or
𝜌1 = (18)
𝑍1 𝑅𝑇1 dense, and where there may be numerous
Gas Mass Flow, kg/h other utilities underground. Multi-storey
𝑚𝑔 = 𝑄1 × 𝜌1 (19) means four or more floors above ground
including the first, or ground, floor. The
Pipeline Process & Mechanical Design depth of basements or number of
To perform Gas Pipeline design, dimensions basement floors is immaterial.
are chosen based on the following factors,
Line Specification of Gas Pipelines – API 5L
Location of the Gas Pipelines 1. PSL1 pipes are available through size 2/5”
1. Class 1 location - A Class 1 location is any to 80” whereas the smallest diameter pipe
1-mile pipeline section that has 10 or available in PSL2 is 4.5” & the largest
fewer buildings intended for human diameter is 80”. PSL1 pipelines are
occupancy including areas such as, available in different types of ends, such as
Plain end, Threaded end, Bevelled end,
Page 70
 special coupling pipes whereas PSL2 2. No work is performed by the gas.
pipelines are available in only Plain End. 3. Friction factor [f] is a constant function of
2. For PSL2 welded pipes, except continuous pipeline length.
welding and laser welding, all other Based on these assumptions, since natural gas
welding methods are acceptable. For pipelines operate at high Reynolds numbers
electric weld welder frequency for PSL2 that are well in turbulent flow regime &
pipeline is minimum 100 kHz whereas Moody’s friction factor becomes merely a
there is no such limitation on PSL1 function of relative roughness, the Weymouth
pipelines. equation can be applied. The Weymouth
3. Heat treatment of electric welds is equation is expressed as,
required for all Grades of PSL2 pipes 8⁄ 𝑇 𝑃2 −𝑒 𝑠 𝑃22
𝑄 = 433.49 𝐸 [𝐼𝐷 3] [ 𝑏] √ 1 (21)
whereas for PSL1 pipelines, grades above 𝑃𝑏 𝛾𝑔 𝑇𝑓 𝐿𝑒 𝑍
X42 require it. 𝐿[𝑒 𝑠 −1]
𝐿𝑒 = (22)
𝑠
4. All kinds of welding method are acceptable 𝛾𝑔 ×∆𝐻
to manufacture PSL1; however, continuous 𝑠 = 0.0375 (23)
𝑇𝑓 ×𝑍
welding is limited to Grade A25. Where, Tb = Base Temperature [0R] [519.7 0R]
Gas Pipeline Wall Thickness Estimation Pb = Base pressure [psia] [14.7 psia]
The B31.8 code is often used as the standard
P1 = Pipeline Inlet Pressure [psia]
of design for natural gas piping systems in
facilities, such as compressor stations, gas P2 = Pipeline Inlet Pressure [psia]
treatment facilities, measurement & ID = Pipeline Inner Diameter [in]
regulation stations & tank farms. The B31.8 g = Gas Specific Gravity [-]
wall-thickness formula is stated as,
Tf = Gas Flowing Temperature [0R]
𝐷𝑃×𝑂𝐷
𝑡= (20) Le = Pipeline Equivalent Length [ft]
2×𝐹×𝐸×𝑇×𝑆𝑀𝑌𝑆
Where, s = Static head due to elevation change [ft/0R]

t = Minimum design wall thickness [in] H = Elevation Difference [ft]

DP = Pipeline Design Pressure [psi] E = Pipeline Efficiency [-]

OD = Pipeline Outer Diameter [in] Z = Gas Compressibility Factor [-]

SMYS = Specific Minimum Yield Stress [psi] Weymouth Equation is also recommended for
shorter lengths of pipeline segments (<32
F = Design Factor [-]
kms) within production batteries and for
E = Longitudinal Weld Joint Factor [E] branch gathering lines, medium to high
T = Temperature De-rating Factor [-] pressure (+/–100 psig [6.9 barg] to > 1,000
A min. corrosion allowance of 1 mm is taken psig [70 barg]) applications, and a high
for stainless steel & 3 mm is taken for carbon Reynolds number.
steel pipelines respectively. Typically, pipeline efficiency factors [Ep] may
vary between 0.6 & 0.92 depending on the
Gas Pipeline Pressure Drop
pipeline’s liquid content. As the amount of gas
To evaluate the pressure drop across the gas
phase liquid content increases, the pipeline
pipeline, the following assumptions are made,
efficiency factor can no longer account for the
1. Flow is steady state.

Page 71
2-phase flow behaviour and 2-phase flow Considering the static head contribution of
equations must be used. any condensation of liquids in the pipeline, it
is accounted for as an equivalent length of
Pipeline Exit Temperature
For long gas transmission pipelines with pressure drop. Hence H = 0 & L = Le.
moderate pressure drop, the temperature Therefore the pipeline exit pressure [P2 =Pe]
expansion due to pressure drop is considered using Weymouth equation is calculated as,
to be small and hence the pipeline exit 2
temperature [Te] can be calculated as, 𝑄
𝑃𝑒 = √𝑃12 − 𝛾𝑔 𝑇𝑓 𝐿𝑒 𝑍 [ 8 𝑇 ] (27)
−
𝜋×𝑂𝐷×𝑈×𝐿 433.49 𝐸[𝐼𝐷 ⁄3 ][ 𝑏 ]
𝑃𝑏
𝑚𝑔 ×𝐶𝑝
𝑇𝑒 = 𝑇𝑠 + [𝑇1 − 𝑇𝑠/𝑎 ]𝑒 (24)

Where, Te = Pipeline Exit Temperature [K] Since the gas flow through the pipeline is
compressible, the compressibility factor is
Ts/a = Soil/Ambient Temperature [K]
expected to vary along with gas velocity, gas
T1 = Pipeline Inlet Temperature [K]
pressure & gas temperature due to heat
U = Overall HTC [W/m2.K] losses. Though due to temperature expansion,
Cp = Mass Specific Heat [J/kg.K] exit temperature is expected to change; its
mg = Mass Flow rate of Gas [kg/s] contribution is considered to be small. The
compressibility factor [Z] would be used to
OD = Pipeline Outer Diameter [m]
compute the gas exit pressure [Pe] based on
The ideal mass specific heat [Cp], kJ/kg.K, of pipeline flowing gas temperature [Tf,a]. The
natural gas can be computed as, value of Za is an implicit parameter & hence
𝐶𝑝 = [(−10.9602𝛾𝑔 + 25.9033) + (0.21517𝛾𝑔 − has to be solved for iteratively as follows,
0.068687)𝑇 + (−0.00013337𝛾𝑔 + Iteration 1: Guess Initial value of Za,1,
2
0.000086387)𝑇 + (0.000000031474𝛾𝑔 − pipeline exit pressure Pe,1, and calculate
0.000000028396)𝑇 3 )]/𝑀𝑊 (25)

Max Hydro Test & Leak Pressure Test 𝑇1 −𝑇𝑒

𝑇𝑓,𝑎 = 𝑇𝑠/𝑎 + [ 𝑇1 −𝑇𝑠/𝑎
] (30)
The maximum allowable operating pressure 𝑙𝑛(
𝑇𝑒 −𝑇𝑠/𝑎
)
[MAOP] is taken as 90% of the design
𝑇𝑓,𝑎 [°𝑅 ]
pressure & for an 8 hour minimum test 𝑇𝑝𝑟 = ′ [° ] (31)
𝑇𝑝𝑐 𝑅
pressure, the hydro test pressure is based on
the location class and maximum test pressure 𝑃𝑒,1 [𝑝𝑠𝑖𝑎]
𝑃𝑝𝑟,1 = ′ [𝑝𝑠𝑖𝑎] (32)
𝑃𝑝𝑐
becomes the lower value of 8 hour minimum
test pressure & test pressure at low point. 2
The leak test pressure is taken as 80% of the 𝑃𝑒,1 = √𝑃12 − 𝛾𝑔 𝑇𝑓,𝑎 𝐿𝑒 𝑍𝑎,1 [
𝑄
] (33)
8 𝑇
433.49 𝐸[𝐼𝐷 ⁄3 ][ 𝑏 ]
design pressure. 𝑃𝑏

Pipeline Exit Pressure 0.27×𝑃𝑝𝑟,1

𝜌𝑝𝑟,1 = (34)
The pipeline exit pressure can be computed 𝑍𝑎,1 ×𝑇𝑝𝑟
by re-arranging Weymouth’s equation as, Iteration 2: Assigning Pe,1 = Pe,2,
2 𝑃𝑒,2 [𝑝𝑠𝑖𝑎]
𝑃𝑝𝑟,2 = ′ [𝑝𝑠𝑖𝑎] (35)
𝑄 𝑃𝑝𝑐
𝑃12 −𝛾𝑔 𝑇𝑓 𝐿𝑒 𝑍[ 𝑇𝑏 ]
√ 8
433.49 𝐸[𝐼𝐷 ⁄3 ][ ]
𝑃𝑏
𝑃2 = 𝑠 (26)
𝑒

Page 72
𝑍𝑎,2 = 1 + [𝐴1 +
𝐴2
+
𝐴3
+
𝐴4
+
𝐴5
] 𝜌𝑝𝑟,1 + used in the presence of known corrosives
𝑇𝑝𝑟 3
𝑇𝑝𝑟 4
𝑇𝑝𝑟 5
𝑇𝑝𝑟
such as CO2. The minimum gas velocity
𝐴 𝐴 𝐴 𝐴
2
[𝐴6 + 𝑇 7 + 𝑇 28 ] 𝜌𝑝𝑟,1 5
− 𝐴9 [𝑇 7 + 𝑇 28 ] 𝜌𝑝𝑟,1 + should be between 10 and 15 ft/s [3 m/s to
𝑝𝑟 𝑝𝑟 𝑝𝑟 𝑝𝑟
2 4.5 m/s], which minimizes liquid fallout.
2 𝜌𝑟,1 2
−𝐴11 𝜌𝑝𝑟,1
𝐴10 (1 + 𝐴11 𝜌𝑝𝑟,1 ) ( 𝑇3 ) 𝑒 (36)
𝑝𝑟 Pressure Drop & Pressure Drop/km
2
The total pressure drop across the Gas
𝑃𝑒,2 = √𝑃12 − 𝛾𝑔 𝑇𝑓,𝑎 𝐿𝑒 𝑍𝑎,2 [
𝑄
] (37) pipeline is computed as,
8 𝑇
433.49 𝐸[𝐼𝐷 ⁄3 ][ 𝑏 ]
𝑃𝑏
∆𝑃 = 𝑃1 − 𝑃𝑒,𝑛 (44)
0.27×𝑃𝑝𝑟,2 The pressure drop/km is computed as,
𝜌𝑝𝑟,2 = (38)
𝑍𝑎,2 ×𝑇𝑝𝑟
∆𝑃 𝑃1 −𝑃𝑒,𝑛
= (45)
Iteration 3: Assigning Pe,2 = Pe,3, 𝐿𝑒 𝐿𝑒

𝑃𝑒,3 [𝑝𝑠𝑖𝑎] The optimum pressure drop for natural gas

𝑃𝑝𝑟,3 = (39)
′ [𝑝𝑠𝑖𝑎]
𝑃𝑝𝑐 pipelines can be taken to be between 3.5
𝐴 𝐴 𝐴 𝐴 psi/mile & 5.83 psi/mile (0.15 - 0.25 bar/km)
𝑍𝑎,3 = 1 + [𝐴1 + 𝑇 2 + 𝑇 33 + 𝑇 44 + 𝑇 55 ] 𝜌𝑝𝑟,2 +
𝑝𝑟 𝑝𝑟 𝑝𝑟 𝑝𝑟
Booster Compressor Process Design
𝐴7 𝐴8 2 𝐴7 𝐴8 5
[𝐴6 + 𝑇 + 2 ] 𝜌𝑝𝑟,2 − 𝐴9 [𝑇 + 2 ] 𝜌𝑝𝑟,2 + Natural Gas Properties Estimation
𝑝𝑟 𝑇𝑝𝑟 𝑝𝑟 𝑇𝑝𝑟
𝜌2−𝐴11 𝜌𝑝𝑟,2 2 The properties of natural gas at the booster
2
𝐴10 (1 + 𝐴11 𝜌𝑝𝑟,2 ) ( 𝑇𝑝𝑟,2
3 )𝑒 (40)
𝑝𝑟 station inlet can be computed based on the
upstream pipeline exit process conditions
2
√𝑃12 𝑄 which become the compressor inlet input.
𝑃𝑒,3 = − 𝛾𝑔 𝑇𝑓,𝑎 𝐿𝑒 𝑍𝑎,3 [ 8 𝑇 ] (41)
433.49 𝐸[𝐼𝐷 ⁄3 ][ 𝑏 ]
𝑃𝑏 1. As the gas enters into the compressor
0.27×𝑃𝑝𝑟,3 station, a loss of pressure until the
𝜌𝑝𝑟,3 = (42) compressor suction flange is expected. This
𝑍𝑎,3 ×𝑇𝑝𝑟
is assumed to be 1 bar for the module. The
The above set of iterations is to be continued
temperature drop up till the compressor
until the convergence criteria of Pe,n – Pe,n-1 <
suction flange is assumed to be negligible
1e-6 (say) is met.
and hence would be nearly the same as the
Gas Pipeline Inlet/Outlet Velocity upstream pipeline outlet temperature.
The pipeline inlet velocity is computed as, 2. The gas upon being compressed, a
60×𝑄𝑔 ×𝑇×𝑍
𝑉𝑔 = (43) pressure drop is expected to exist between
𝑂𝐷2 ×𝑃
the compressor discharge flange & the
Where, Vg = Gas Velocity [ft/s] downstream pipeline inlet. For this
Qg = Gas Flow rate [MMscfd] module, a drop in pressure of 1 bar is
T = Inlet/Outlet Temperature [0R] taken across the discharge cooler &
downstream piping. To ensure the
P = Inlet/Outlet Pressure [psia]
discharge piping design temperature of
Z = Compressibility Factor [Z] 2000C is not exceeded even during gas
OD = Pipeline Outer Diameter [in] recycle across the anti-surge valve, an air
The velocity in gas lines should be less than cooler is installed to cool the compressed
60 to 80 ft/s [18m/s to 25m/s] to minimize gas to 500C. The temperature drop
noise and allow for corrosion inhibition. A between the air cooler discharge &
lower velocity of 50 ft/s [15 m/s] should be downstream pipeline inlet is taken as 10C.
Page 73
3. With the gas compressor inlet pressure 𝑛𝑛 =
1
(52)
1
temperature, mass flow rate, actual [1−[1− ] ]
𝑘𝑎,𝑛 𝑝

volumetric flow rate, gas density, the With a selected compressor discharge
compressor suction & discharge gas pressure [P2], the modified properties of
properties can be calculated. reduced temperature [Tpr,n], reduced pressure
4. To estimate the discharge conditions, the [Ppr,n] & reduced density [ pr,n] is,
compressor discharge pressure [P2] & 𝑃2
𝑃𝑝𝑟 = (49)
downstream pipeline size is chosen such ′
𝑃𝑝𝑐

that, sufficient discharge pressure exists to 𝑇2,𝑛

𝑇𝑝𝑟,𝑛 = (50)
deliver natural gas through the ′
𝑇𝑝𝑐

downstream pipeline while respecting the 0.27×𝑃𝑝𝑟

𝜌𝑝𝑟,𝑛 = (51)
pipeline velocity, pressure drop criteria 𝑍2,𝑛 ×𝑇𝑝𝑟

and meets the city power station’s The discharge temperature [T2,n] is,
required pressure criteria. 𝑛𝑛 −1
𝑃2 𝑛𝑛 𝑍1
5. With these compressor inlet process 𝑇2,𝑛 = 𝑇1 [ ] [ ] (52)
𝑃1 𝑍2,𝑛
conditions, the DAK EoS with Wichert &
Iteration n=2: Guess Initial value of Z2,n+1 =
Aziz correction is used as shown in the
Z2,n,
previous section to calculate, compressor
𝐶𝑝,2,𝑛+1
inlet’s Modified properties of Reduced 𝑘2,𝑛+1 = (53)
𝐶𝑝,2,𝑛+1 −𝑅
Pressure [Ppr], Reduced Temperature [Tpr],
The average adiabatic exponent becomes
Reduced density [ pr], Specific heat [Cp] &
𝑘1 +𝑘2,𝑛+1
gas compressibility factor [Z]. 𝑘𝑎,𝑛+1 = (54)
2
Booster Compressor Process Conditions
The polytropic exponent [n] for iteration n=2
To evaluate the booster compressor’s process
[nn+1] is calculated as,
conditions, the average adiabatic exponent
1
[ka] & polytropic exponent [n] is required to 𝑛𝑛+1 = 1 (55)
[1−[1− ] ]
𝑘𝑎,𝑛+1 𝑝
be calculated for the compressor discharge.
This can be performed iteratively based on an The reduced temperature [Tpr,n] becomes,
𝑇2,𝑛
initial guess of the compressor discharge 𝑇𝑝𝑟,𝑛+1 = ′ (56)
𝑇𝑝𝑐
temperature [T2], discharge side specific heat
[Cp,2], with which discharge side adiabatic The discharge side gas compressibility factor
exponent [k2] can be calculated as follows, [Z2,n+1] can be calculated as,
Iteration n=1: Guess Initial value of Z2,n, 𝑍2,𝑛+1 = 1 + [𝐴1 +
𝐴2
+
𝐴3
+
𝑇𝑝𝑟,𝑛+1 3
𝑇𝑝𝑟,𝑛+1
compressor discharge temperature, T2,n,
𝐴4 𝐴5 𝐴7
𝑘1 =
𝐶𝑝,1
(46) 4 + 5 ] 𝜌𝑝𝑟,𝑛 + [𝐴6 + +
𝑇𝑝𝑟,𝑛+1 𝑇𝑝𝑟,𝑛+1 𝑇𝑝𝑟,𝑛+1
𝐶𝑝,1 −𝑅
𝐴8 2 𝐴7 𝐴8 5
𝐶𝑝,2,𝑛 ] 𝜌𝑝𝑟,𝑛 − 𝐴9 [ + ] 𝜌𝑝𝑟,𝑛 +
𝑘2,𝑛 = (47) 2
𝑇𝑝𝑟,𝑛+1 𝑇𝑝𝑟,𝑛+1 2
𝑇𝑝𝑟,𝑛+1
𝐶𝑝2,𝑛 −𝑅
2
𝜌𝑝𝑟,𝑛 2
The average adiabatic exponent [ka] becomes
2
𝐴10 (1 + 𝐴11 𝜌𝑝𝑟,𝑛 )( 3 ) 𝑒 −𝐴11𝜌𝑝𝑟,𝑛 (57)
𝑇𝑝𝑟,𝑛+1
𝑘1 +𝑘2,𝑛
𝑘𝑎,𝑛 = (48) 0.27×𝑃𝑝𝑟
2 𝜌𝑝𝑟,𝑛+1 = (58)
𝑍2,𝑛+1 ×𝑇𝑝𝑟,𝑛+1
The polytropic exponent [n] for iteration n=1
The discharge temperature [T2,n+1] is,
[nn] is calculated as,
Page 74
 𝑃2
𝑛𝑛+1 −1
𝑛𝑛+1 𝑍1 P1 = Compressor Suction Pressure [bara]
𝑇2,𝑛+1 = 𝑇1 [ ] [ ] (59)
𝑃1 𝑍2,𝑛+1 P2 = Compressor Discharge Pressure [bara]
The above set of iterations are to be The absorbed power, P [kW] is calculated as,
continued until the convergence criteria of 𝐻𝑝 ×𝑚
T2,n+1 – T2,n < 1e-6 (say) is met. 𝑃= (66)
𝑝

The gas density [kg/m3] is calculated as, Where,

=
𝑃×𝑀𝑊
(60) m = Mass Flow rate [kg/s]
𝑍×𝑅×𝑇
p = Polytropic efficiency [-]
Where,
P = Inlet/Outlet Pressure [bara] Results
With the above procedures executed, the
MW = Molecular Weight [kg/kmol]
pseudocritical properties of natural gas are,
Z = Inlet/Outlet Compressibility Factor [-]
Table 6. Sutton Correlation & Wichert & Aziz
R = 0.0831447 m3.bar/kmol.K Correction

T = Inlet/Outlet Temperature [K]

The molar density [ m] is calculated as,
𝜌
𝑚
= (61)
𝑀𝑊

The molar volume [Vm], m3/kg is,

1
𝑉𝑚 = (62)
𝑚

With calculated molar volume, the following

balance based on polytropic exponent [n] &
molar volume [Vm] is checked as,
𝑛
𝑃1 𝑉𝑚,1 𝑛
− 𝑃2 𝑉𝑚,2 =0 (63) The upstream & downstream pipeline
mechanical design data is estimated as
The adiabatic efficiency [ a] is calculated as, follows,
𝑛−1
𝑛 𝑃 Table 7. Pipeline Mechanical Design Results
[( 2 ) 𝑛 −1]
𝑛−1 𝑃1
1
𝑎
= 𝑘−1 × (64)
𝑘 𝑃2 𝑘 𝑝
[( )
𝑘−1 𝑃1
−1]

The polytropic head Hp [kJ/kg] is,

𝑛−1
𝑍𝑎𝑣𝑔 𝑅𝑇 𝑛 𝑃2 𝑛
𝐻𝑝 = [ ] [( ) − 1] (65)
𝑀𝑊 𝑛−1 𝑃 1

Where, Hp = Polytropic Head [kJ/kg]

Za = [Z1+Z2]/2; Avg. compressibility factor [-]
R = 8.31447 kJ/kg.K
T = Gas Inlet Temperature [K]
MW = Molecular Weight [kg/kmol]
n = Polytropic exponent [-]

Page 75
 Table 9. Booster Compressor Design Details

The upstream & downstream pipeline

process design results are,
Table 8. Pipeline Process Design Details

The Max Hydrostatic & Leak Test Pressure is,

Table 10. Hydrostatic & Leak Test Pressure

The booster compressor design results are,

Page 76
For the iterative steps made for compressor Technol 14 (3): 34. PETSOC-75-03-
discharge temperature [T2], Upstream & 03 (http://dx.doi.org/10.2118/75-03-03)
Downstream Pipeline Exit Pressure [Pe] 6. https://petrowiki.org/Real_gases
calculations, the error percentage in the
7. “Compressibility Factors for Naturally
iterative steps is plotted below.
Occurring Petroleum Gases, Piper, L.D.,
McCain Jr., W.D., and Corredor, J.H., SPE
Annual Technical Conference & Exhibition,
Houston, 3–6 October 1993, SPE-26668-
MS (http://dx.doi.org/10.2118/26668-MS)
8. “Handbook of Natural Gas Transmission and
Processing, Principles and Practices, Saied
Mokhatab, Wilaim A. Poe, John Y Mak, 3rd
Edition

Figure 2. Error Percentage in Iterative Steps

9. “Pipeline Systems, Design, Construction,
Appendix A shows a summary of the MS- Maintenance and Asset Management”,
Excel Steps & Appendix B shows the Nandagopal N.S, Rev. 4
iterations made to calculate the compressor 10. “Standard for Gas Transmission and
discharge temperature [T2], Upstream & Distribution Piping Systems”, ANSI/ASME
Downstream Pipeline Exit Pressures [Pe]. Standard B31.8, 1999

References & Further Reading 11. Recommended Practice for Analysis, Design,
Installation, and Testing of Basic Surface
1. “Handbook of Natural Gas Engineering”,
Safety Systems for Offshore Production
Katz, D.L., 1959, McGraw-Hill Higher
Facilities,” API RP14C, 7th Ed, 2001
Education, New York
12. “Isobaric specific heat capacity of natural
2. “Gases and Vapors At High Temperature gas as a function of specific gravity, pressure
and Pressure - Density of Hydrocarbon”, Kay and temperature”, Lateef A. Kareema,
W, 1936 Ind. Eng. Chem. 28 (9): 1014-1019 Tajudeen M. Iwalewa, James E. Omeke,
(http://dx.doi.org/10.1021/ie50321a008) Journal of Natural Gas Science and
3. “Density of Natural Gases”, Standing, M.B. Engineering 19 (2014) 74-83
and Katz, D.L. 1942 In Transactions of the
American Institute of Mining and
Metallurgical Engineers, No. 142, SPE-
942140-G, 140–149, New York
4. “Compressibility Factors for High-
Molecular-Weight Reservoir Gases”, Sutton,
R.P. 1985. SPE Annual Technical
Conference & Exhibition, Las Vegas,
Nevada, USA, 22-26 Sep, SPE-14265-MS
(http://dx.doi.org/10.2118/14265-MS)
5. “Calculation of Z Factors For Natural Gases
Using Equations of State”, Dranchuk, P.M.
and Abou-Kassem, H. 1975, J Can Pet
Page 77
Appendix A

Page 78
Appendix B

Page 79
 Module 12
Vapour Compression for Propane-Propylene Splitters
Propane-Propylene (C3) splitters are used in between the column top and bottom should
Petrochemical industries to distil propylene not be more than 250F (13.890C).
from propane prior to export. C3 Splitter Additionally, the bottom heat of vaporization
columns are traditionally operated with low should be close to overhead vapour’s heat of
pressure steam (LPS) passed to the column’s condensation with a pressure drop less than
reboiler to heat up the C3 mixture thereby 15 psi (1 bar) across the column internals.
causing the splitting operation. However the The heat of vaporization for propylene and
use of Low Pressure steam is energy- propane are nearly close at 157.6 Btu/lb and
demanding. A C3 splitter is equipped with a 151.7 Btu/lb respectively. The excess energy
required to be supplied by the compressor is
cooler and flash tank at the column top that
around 11% - 12% of reboiler duty [Ref 2]
condenses the column top vapours but in
which represents a high energy savings. Low
doing so, loses precious heat of condensation.
pressure splitters also offer the advantage of
In recent years, the concept of using a heat
fewer trays, shorter column height and lower
pump, such as a compressor has become a
column wall thickness that represents a
standard practice that eliminates the
capital savings with higher relative
necessity of using low pressure steam. volatilities to effectuate product separation.
C3 Splitter with Vapour Compression
Low pressure C3 splitter columns operate
Vapour compression methods for C3 between 90 psig (6.2 barg) to 110 psig (7.6
stripping operations are ideal for compounds
barg) depending on the Technology Licensor.
that have low relative volatilities. Below is a
For the heat pump, the choice of compressor
schematic of a C3 splitter with vapour
used can be centrifugal type with a typical
compression of product (propylene). Low
pressure ratio of 1.8 [Ref 2] and effectively
pressure steam that was used to vaporize the
regulates the column top pressures when the
bottoms propylene product is replaced by
installing a column top compressor and throughput varies. The operation of the
routing a portion of the discharge to vaporize centrifugal compressor can be fixed type or
the bottoms propane via the reboiler. variable speed type with the latter
representing greater control but with higher
installation costs.
C3 Column Top Pressure Control Methods
Suction Throttling Method
The Suction throttling method involves using
a control valve (e.g. butterfly valve) placed at
suction side of the centrifugal compressor.
But these are suitable only for fixed speed
drives like Asynchronous Induction Motors
where the driver speed cannot be
manipulated. Below is a process schematic of
Figure 1. C3 Splitter with Compression
a C3 column top compressor that works on
As per Ref [2], for a C3 splitter to operate
the principle of suction throttling. A cooler on
effectively, the temperature difference
Page 80
the anti-surge line cuts down the inline abatement measure, the suction throttle
pressure loss in the compressor discharge valve would open further to maintain the
and also reduces the compressor discharge required compressor discharge pressure.
side equipment and piping volumes, 3. For both cases of C3 column top pressure
contributing to the fast response of the anti- increasing/decreasing, the anti surge
surge system. The PIC on the compressor valve (ASV) also acts in tandem with the
discharge receives discharge pressures from suction throttle valve via the Anti-surge
discharge side pressure transmitter (PT) to controller. This ensures that the
alter the suction throttle valve opening. compressor operating point does not cross
the surge control line (SCL).
4. From an energy savings perspective, the
pinching operation of the suction throttle
valve would make the compressor operate
closer to the surge line where the power
absorbed is lower.
Variable Speed Motor (VSM) Method
As an alternative, variable speed motor
eliminates the need for a suction throttle
Figure 2. Suction Throttling for C3 Column Top valve and can cater to the tower top
Compressor propylene vapours during column
For C3 column operation, when the operating fluctuations by altering the motor speed
pressure at the column top increases, the based on the discharge pressure of the
suction throttle valve is altered based on the compressor. The discharge side pressure
compressor discharge pressure. The controller PIC cascades its output (OP) to
operational advantage would be as follows, assign a set point to the speed controller (SC)
and controls the motor speed. In doing so,
1. For the case where the column top
both output flow and pressure are regulated.
pressure increases, it results in more
propylene flow into the compressor. This
causes the motor to draw in more
amperage to sustain the required higher
torque but can trip the electric motor and
subsequently the compressor. Therefore,
the suction throttle valve in such an event
closes accordingly to ensure the discharge
pressure required is maintained without
tripping the compressor.
2. In the event, the C3 column pressure falls,
and for a fixed speed of the motor and Figure 3. Variable Speed Motor for C3 Column Top
Compressor
reduced throughput from the column top,
the compressor discharge pressure would In real situations, C3 splitter columns can also
rise causing the column top flash tank experience fluctuations in operating
pressure and the bottom reboiler pressure pressures due to any changes in the
to exceed design limits. Therefore as an upstream side of the C3 splitter. This would
Page 81
also mean the C3 splitter column would take a 3. In vapour compression methods, when the
while to again attain equilibrium across trays use of steam is eliminated to vaporize the
and the vapour compressor at the tower top bottoms product, the tower top vapours
would also need to synchronize itself with from the compressor are used to
the column operating pressures. Comparing effectuate vaporization of the bottoms
suction throttling methods over Variable product in the reboiler. In the event of the
speed methods, the following points are column operating pressure exceeding the
described below. stipulated values, variable speed drives
reduce the compressor speed and hence
1. Suction throttling methods though can be
discharge pressure, thereby supplying a
used but suffer from an operability range
reduced throughput to the reboiler. This
of the throttle valve. This often requires
consequently reduces the reboiler
the anti surge valve (ASV) to open
temperatures, reboiler vaporization rates,
frequently to avoid the compressor
avoids product degradation and aids in
operating point from crossing the surge
bringing the operating pressure of the C3
control line (SCL). A disadvantage of the
splitter columns within the allowable
suction throttling method is, during a
limits.
reduction in the column top pressure,
when the compressor operates closer to References & Further Reading
the surge line though saving power, the 1. ‘Design Guidelines for Propylene Splitters
risk of the operating point crossing the Efficiencies’, Karl Kolmetz, KLM
surge line is high. Any further fluctuation Technology Group
in the column top pressure is detrimental 2. ‘Be Smart About Column Design’, Mark
in creating a Level 1 emergency equipment Pilling, P.E, Daniel R. Summers, P.E, Sulzer
shutdown (ESD-1) scenario and Chemtech, USA, AIChE November 2012
subsequent upsetting of the column
pressure profile. Additionally due to
excessive pressure drop across small
throttle valve openings, the risk of Joule-
Thompson cooling can produce
condensate particles which can erode the
compressor’s impellers.
2. Variable speed compressors also cater
well during turndown conditions, as much
as 30%. However this must be considered
during the process design basis stage
when setting the design and operating
limits of the C3 splitter column. For such
low turndown conditions, the compressor
vendor must also be consulted prior to
equipment selection to understand how
much vapour recycling can occur via the
anti-surge valve during such low
turndown conditions.
Page 82
 Module 13
BOIL OFF GAS ANALYSIS OF LNG AT RECEIVING TERMINALS
Liquefied Natural Gas (LNG) is a cryogenic Typical Design Considerations [3]
mixture of low molecular weight (MW) 1. Single containment tank is either a single
hydrocarbons with its chief component being tank or a tank comprising an inner tank
methane. Its uses cover a gamut of and outer container, designed and
applications from domestic & industrial use, constructed so that only the inner tank is
power generation, to transportation fuel in its required to meet the low temperature
liquid form. LNG is transported in double- ductility requirements for storage of the
hulled ships specifically designed to handle product. A double containment tank is a
low temperatures of the order of -1620C. As of tank designed and constructed so that both
2012, there were 360 ships transporting the inner tank and the outer tank are
more than 220 million metric tons of LNG capable of independently containing the
every year. [1] refrigerated liquid stored.
When LNG is received at most terminals, it is 2. The difference between the double
transferred to insulated storage tanks that containment and full containment is that
are built to specifically hold LNG. These tanks the outer tank of a full containment tank is
can be above or below ground & keep the intended to be capable of both containing
liquid at a low temperature to minimize the refrigerated liquid and of controlled
evaporation & compositional changes due to venting of the vapour resulting from
heat ingress from the ambient. The product leakage after a credible event.
temperature within the tank will remain Among these three types of LNG tanks, the
constant if the pressure is kept constant by full containment type is regarded as the
allowing the boil off gas (BOG) to escape from most advanced type.
the tank. This is known as auto-refrigeration. 3. The inner tank is manufactured with 9%
BOG is collected & used as a fuel source in the nickel steel and the outer tank is composed
facility or on the tanker transporting it. When of reinforced concrete and pre-stressed
natural gas is needed, LNG is warmed enough concrete. The 9% nickel steel is widely
using heat exchangers to vaporize it called re- used as a material for the inner tank since
gasification process, prior to transferring it to it has the strength and toughness enough
the pipeline grid to various users. for cryogenic uses. The inner tank also has
Boil-off gas (BOG) management & assessment a function of preventing LNG from leakage.
of LNG’s thermodynamic properties are key The concrete outer tank is designed to
issues in the technical assessment of LNG resist all the external loads including
storage. Increased vaporization process may seismic load. Insulating materials are
negatively affect the stability and safety of the placed between the inner and outer tank to
stored LNG. For these reasons the rate of preserve the stored LNG.
vaporization (boil off rate) should be 4. The design boil off rate is typically about
precisely determined in storage terminal 0.05 vol%/day & the design vacuum
energy systems. [2]. pressure in the dome is about -0.05 kPag.
5. The roof has a suspended ceiling deck and
a steel lined concrete dome. The steel liner
Page 83
 installed on the inside surface of the outer 4. The LNG tank’s inner heat transfer
concrete tank provides the gas tightness. coefficient is taken as 35 W/m2.K & the
The boil-off rate is determined by the soil’s heat transfer coefficient is taken as 2
insulation system. W/m2.K.
6. In case of an LNG leakage, liquid may 5. Since no information is available from Ref
impact the outer tank. Accordingly, liquid [2] regarding the maximum operating
tightness must be guaranteed by the liquid level in the LNG tank, the max
corner protection system as well as the operating liquid level (H) is taken to be 2 m
polyurethane foam coating installed on the less than the height of the inner
inside surface of the concrete wall. The containment. This would mean a vapour
concrete outer tank protects the inner tank space is available & the heat transfer
in case of emergency from the outside. through this vapour space is neglected.
7. The base of the tank has a bottom heating 6. Certain insulation material related to the
system (BHS) using ethylene glycol as roof top & tank bottom is not presented in
brine fluid. The bottom heating system is Ref [2], hence the insulation material and
installed in order to avoid frost heave. their concerned thickness is assumed as
shown in the next section.
8. The roof liner consisting of a 9% Ni steel
membrane stiffened with rafters in radial 7. Heat transfer due to free convection is
and tangential directions acts as formwork expected. Considering the diameter of the
for the concrete sphere. The steel structure LNG tank to be very large, any section of
is fabricated on the bottom slab and lifted the concrete wall, is assumed to be similar
by air pressure to its final position. Rafters to flow over a vertical flat plate.
and roof liner plates are connected with a 8. The LNG tank outer temperature at the
steel compression ring anchored in the concrete wall is expected to be less than
concrete roof ring-beam by welding. the ambient temperature. In calculating the
free convection, Grashof number is
Case Study & Assumptions
expected to yield a negative value.
To demonstrate the BOG Rate calculations,
However as per Ref [4], Sec. 9.4, Page 568
the following case study is made based on a
single containment LNG Tank dimensions & indicates, that irrespective of Ts > T or Ts <
composition given in Ref [2]. In addition, T , where, T is the ambient temperature &
certain assumptions are made for this Ts is the surface temperature, the foregoing
example case study. results to estimate Nusselt number still
apply. When Ts < T , free convection still
1. The ambient temperature is taken to be
develops, except that the convection
26.850C & wind velocity is taken to be 36
boundary flows downward. Therefore, in
km/h (10 m/s). The soil temperature is
the event of obtaining a negative Grashof
taken as 16.850C.
Number based on Ts < T , the absolute
2. For heat transfer computations, the
average temperature on the LNG Tank value is taken.
surface is taken to be an average of LNG 9. LNG is stored at cryogenic temperatures
temperature inside the tank and the for significant durations and inevitable
ambient air temperature. heat ingress from the surroundings into
3. Radiation plays a role & the emissivity of the storage tank will lead to vaporization.
the outer concrete is taken to be 0.9. The more volatile components (methane
Page 84
 and nitrogen) will vaporize preferentially, 𝑀𝑊𝐿𝑁𝐺 = ∑𝑖=𝑛
𝑖=1 𝑥𝑖 𝑀𝑊𝑖 (1)
resulting in weathering of LNG. If left
𝐿𝑁𝐺 = ∑𝑖=𝑛
𝑖=1 𝑥𝑖 𝑖 (2)
unchecked, weathering can render the
remaining LNG unsellable, because of Where,
regulatory requirements. Furthermore, MW = Molecular Weight [kg/kmol]
weathering increases the overall tank = Latent Heat of Vaporization [kJ/kg]
pressure and in order to avoid over-
Table 2. LNG Latent Heat of Vaporization & MW
pressurization some of the generated
vapour is removed, as boil-off gas (BOG). Component LNG xi.MWi xi. i

10. The LNG industry is specifically concerned - [kJ/kg] [kg/kmol] [kJ/kg]

in minimizing BOG rates and ensuring that Methane 510.8 14.36 457.3
weathering does not greatly impact the Ethane 489.3 2.07 33.7
LNG quality. In particular weathering
Propane 425.6 1.07 10.3
prediction is used in planning operations,
i-Butane 365.1 0.36 2.3
thus ensuring appropriate allocation of
LNG cargoes, its compatibility with stored n-Butane 385.7 0.27 1.8
LNG and avoiding catastrophic events i-Pentane 349.3 0.00 0.0
involving stratification, sudden vapour n-Pentane 367.3 0.00 0.0
release and rollover, Ref [5]. In the current Nitrogen 199.2 0.02 0.2
module, the effects of stratification & roll
MWLNG & LNG 18.1518 505.5
over are neglected.
LNG Tank Design Data To estimate the density of LNG, the specific
molar volumes as per Ref [6] page 145, 146,
For BOG estimation, the following process
(N.B.S. - Technical note 1030 December 1980)
data for heavy LNG is used [2],
need to be computed as,
Table 1. Heavy LNG Composition 𝑀𝐿𝑁𝐺
𝜌𝐿𝑁𝐺 = (3)
Component MW Mol% 𝐿𝑁𝐺
𝑥
𝑁2
- [kg/kmol] [%] 𝐿𝑁𝐺 = ∑ 𝑥𝑖 𝑖 − [𝐾1 + (𝐾1 − 𝐾1 ) 0.0425] 𝑥𝐶𝐻4 (4)

Methane [CH4] 16.04 89.52 Where,

Ethane [C2H6] 30.07 6.89 LNG = Density of LNG [kg/m3]
Propane [C3H8] 44.01 2.42 MLNG = Molar Mass of LNG [kg/kmol]
i-Butane [i-C4H10] 58.12 0.62 LNG = LNG molar volume [L/mol]
n-Butane [n-C4H10] 58.12 0.47 Mi = Molar Mass of Component ‘i’ [kg/kmol]
i-Pentane [i-C5H12] 72.15 0.00 xi =Molar fraction of component ‘i’ [-]
n-Pentane [n-C5H12] 72.15 0.00 = Component molar volumes of ‘i’ at LNG
i

Nitrogen [N2] 28.01 0.08 Temperature [L/mol]

Total 100.0 K1, K2 = Correction Factors [-]

The LNG molecular weight (MW) and latent The values of K1, K2 can be obtained from the
heats of vaporization are estimated using below graphs,
Kay’s Rule of Mixing for the composition
selected and are as follows,

Page 85
 Table 3. LNG Molar Volume

Component i LNG =xi. i

- [L/mol] [lit/mol]
Methane 0.0381 0.034070
Ethane 0.0479 0.003299
Propane 0.0624 0.001511
i-Butane 0.0783 0.000485
n-Butane 0.0768 0.000361
i-Pentane 0.0916 0.000000
n-Pentane 0.0915 0.000000
Figure 1. K1 Correction Factor
Nitrogen 0.0465 0.000037
Total 0.03976

Therefore the LNG density at 112.5K is,

18.1518
𝜌𝐿𝑁𝐺 = 0.03976 = 456.5 𝑘𝑔/𝑚3 (5)

The ambient & soil conditions for the BOG

rate calculations are as follows,
Table 4. Ambient & Soil Conditions

Parameter Value Unit

Ambient Temperature 300 K
Wind Velocity 10 m/s
Figure 2. K2 Correction Factor
Soil Temperature 290 K
The component molar volumes of each
Soil HTC [hsoil] 2 W/m2.K
component ‘i’ at LNG Temperature is,
The air properties between -250C & 500C are
computed using fitted equations as follows,
Air Density [kg/m3] is computed as,
𝜌𝐴𝑖𝑟 = 0.0000158𝑇 2 − 0.0133989𝑇 + 3.7622 (6)

Specific Heat of Air [kJ/kg.K] is computed as,

𝐶𝑝,𝑎𝑖𝑟 = 1.006 (7)

The thermal conductivity [W/m.K] of air is,

𝑘𝐴𝑖𝑟 = −2.69 × 10−8 𝑇 2 + 9.04 × 10−5 𝑇 + 9.56 × 10−4 (8)
The thermal diffusivity [m2/s] of air is,
Figure 3. Component Molar Volume [ i] 𝛼𝐴𝑖𝑟 = 1.99 × 10−10 𝑇 2 + 1.5 × 10−8 𝑇 − 7.96 × 10−7 (9)
Based on the values of K1, K2 & i, the specific The dynamic viscosity [kg/m.s] of air is,
molar volumes of each of the components are 𝜇𝐴𝑖𝑟 = −4.22 × 10−11 𝑇 2 + 7.19 × 10−8 𝑇 + 8 × 10−7 (10)
estimated for an LNG Tank storage The kinematic viscosity [m2/s] of air is,
temperature of -160.650C (112.5K). The LNG
= 1.02 × 10−10 𝑇 2 + 3.1 × 10−8 𝑇 − 2.69 × 10−6 (11)
molar volume based on K1, K2 is computed as, 𝐴𝑖𝑟

Page 86
The Prandtl Number of air is computed as,
𝑃𝑟 = −5.12 × 10−7 𝑇 2 + 3.7 × 10−5 𝑇 + 0.7642 (12)
Based on the above correlations, for an
ambient temperature of 300 K, the air
properties are as follows,
Table 5. Air Properties at Ambient Conditions

Parameter Value Unit

Density [ air] 1.176 kg/m3

Specific Heat [Cp,air] 1.006 kJ/kg.K
Thermal Conductivity [kair] 0.0256 W/m.K
Thermal Diffusivity [ air] 0.000022 m2/s
Figure 5. Containment Representation
Dynamic Viscosity [ air] 0.000019 kg/m.s
The thermal conductivity & wall thickness of
Thermal Exp. Coeff [ air] 0.0033 1/K
each layer of insulation is,
Kinematic Viscosity [ air] 0.00001578 m2/s
Table 7. LNG Tank Insulation
LNG Tank Construction Details Material
Thermal Wall
Cond. [k] Thickness
The construction details (Ref [2]) in addition
- [W/m.K] [m]
to the assumptions made for missing data is,
Table 6. LNG Tank Construction Perlite Insulation [k2] 0.038 0.30

Parameter Value Unit 9% Ni Steel [k5] 90.9 0.005

Tank Height 40 m 9% Ni Steel [k1] 90.9 0.20

Max. Op. Level 38 m Perlite Insulation [k2] 0.038 0.60

Inner Tank ID 74 m Carbon Steel [k3] 42.6 0.15

Inner HTC [Input] 35 W/m2.K Polyurethane [k4] 0.029 0.60

LNG Tank Volume 160,000 m3 9% Ni Steel [k1] 90.9 0.20

Foam Glass [k6] 0.045 0.60
A schematic of the LNG Tank is shown below,
Carbon Steel [k7] 42.6 0.15
Concrete [k8] 1.80 0.60

The various tank Radii/wall thicknesses for

calculating resistances in cylindrical
coordinates & linear coordinates are,
Table 8. LNG Tank Insulation
Tank Radii/Wall
Material
Thickness
- - [m]

Perlite Insulation [k2] r9 0.30

9% Nickel Steel [k5] r10 0.01

Figure 4. Single Containment LNG Tank Schematic 9% Nickel Steel [k1] r1 37.20

The various tank insulations are depicted as, Perlite Insulation [k2] r2 37.80

Page 87
 4⁄
Carbon Steel [k3] r3 37.95 1 1
0.62 𝑅𝑒 ⁄2 𝑃𝑟 ⁄3 𝑅𝑒
5⁄
8
5
𝑁𝑢 = 0.3 + 1⁄ [1 + (282000) ] (16)
Polyurethane Foam [k4] r4 38.55 2
0.4 ⁄3
4
[1+( 𝑃𝑟 ) ]
9% Nickel Steel [k1] r5 0.20
The above correlation is valid for all ranges of
Foam Glass [k6] r6 0.60
Reynolds number (Re) and Pr 0.2, where all
Carbon Steel [k7] r7 0.15
properties are evaluated at film temperature.
Concrete [k8] r8 0.60 It is to be noted that as per [4], Churchill &
Bernstein correlation is reasonable over a
Design Methodology
certain range of conditions but for most
To estimate the heat load, thermal
engineering calculations, the accuracy is not
conductivity and heat transfer coefficients
expected to be much better than 20% because
determine the amount of heat transferred to
these are based on more recent results
the cryogenic LNG through the walls of the
encompassing a wide range of conditions [4].
tank. The modes of heat transfer driven by
4⁄
temperature differences are conduction 0.62 × 2.41 × 107
1⁄
2 0.72926
1⁄
3 2.41 × 107
5⁄
8
5

𝑁𝑢 = 0.3 + [1 + ( ) ]
through the tank wall & its various insulation
1⁄
2⁄ 4 282000
0.4 3
[1 + (0.72926) ]
layers, free convection, forced convection &
radiation from the ambient. (17)

Forced Convection 𝑁𝑢𝑓𝑜𝑟𝑐𝑒 ≈ 23,357 (18)

To calculate the external heat transfer The external forced convection heat transfer
coefficient (hair,overall), taking the assumptions coefficient can be calculated from Nusselt
made, the average surface temperature (Tavg) number as,
of the concrete insulation is taken as an 𝑘𝑎𝑖𝑟 𝑁𝑢𝑓𝑜𝑟𝑐𝑒 0.0256×23357
average of LNG boiling temperature (Tb,LNG) & ℎ𝑓𝑜𝑟𝑐𝑒 = = ≈ 15.8 𝑊/𝑚2 𝐾 (19)
𝐿𝑀𝑎𝑥𝑂𝑝.𝐿𝑒𝑣𝑒𝑙 38

ambient temperature (Ta),

Radiation Heat Transfer
𝑇𝑏,𝐿𝑁𝐺 +𝑇𝑎 112.5+300
𝑇𝑎𝑣𝑔 = = = 206.25 𝐾 (13) To estimate the radiation heat transfer
2 2

Prandtl Number of ambient air is, between the ambient & concrete insulation
𝐶𝑃,𝑎𝑖𝑟 𝜇𝑎𝑖𝑟 1.006×0.000019
on the tank, the expression is written as [4],
𝑃𝑟 = = ≈ 0.72926 (14) 2
𝑘𝑎𝑖𝑟 0.0256 ℎ𝑟 = 𝜀𝜎(𝑇𝑐𝑜𝑛 + 𝑇𝑎 )(𝑇𝑐𝑜𝑛 + 𝑇𝑎2 ) (20)
Based on the assumption that the tank The radiation mode expressed above is
diameter is large; the flow of air over the tank written in a manner similar to convection, i.e.,
is approximated to flow over a flat plate. the radiation rate equation is linearized
Therefore Reynolds number (Re) becomes, making the heat rate proportional to the
𝑅𝑒 =
𝑉𝑎𝑖𝑟 𝐿𝑚𝑎𝑥 𝑜𝑝 𝐿𝑒𝑣𝑒𝑙 10×38
= 0.00001578 = 2.41 × 107 (15) temperature difference rather than to the
𝛾𝑎𝑖𝑟
difference between two temperatures to the
Nusselt number for forced convection over fourth power [2]. Therefore the radiation
circular cylinder with cross flow can be heat transfer coefficient (hr) for emissivity of
estimated using Churchill and Bernstein concrete ( ) taken as 0.9 & Stefan Boltzmann
correlation [4]. This equation is valid for all Constant ( ) of 5.67 10-8 W/m2/K is,
Re.Pr 2 and the correlation is expressed as,
ℎ𝑟 = 0.9 × 5.67 × 10−8 (206.25 + 112.5) × (206.252 + 112.52 )(21)

Or, ℎ𝑟 ≈ 0.9 𝑊/𝑚2 𝐾 (22)

Page 88
Natural/Free Convection Therefore the mode of heat transfer is mixed
To estimate the heat transfer due to natural convection and the combined heat transfer
convection, the correlation by Churchill & Chu coefficient is computed as,
[4] can be used and is of the form, 4 4
1⁄
4
𝑁𝑢𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑 = [𝑁𝑢𝑓𝑟𝑒𝑒 + 𝑁𝑢𝑓𝑜𝑟𝑐𝑒𝑑 ] (34)
2
1⁄
1
0.387 𝑅𝑎 ⁄6 𝑁𝑢𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑 = [233574 + 84974 ] 4 (35)
𝑁𝑢 = {0.825 + 8⁄ } (23)
9 27
0.492 ⁄16
[1+( ) ] 𝑁𝑢𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑 = 23,458 (36)
𝑃𝑟

Therefore the combined external heat

Where, Rayleigh number (Ra) is the product
transfer coefficient, hcomb, is computed as,
of Grashof number (Gr) and Reynolds number
𝑘𝑎𝑖𝑟𝑁𝑢𝑐𝑜𝑚𝑏 0.0256×23458
(Re) computed as, ℎ𝑐𝑜𝑚𝑏 =
𝐿𝑀𝑎𝑥𝑂𝑝.𝐿𝑒𝑣𝑒𝑙
=
38
≈ 15.8 𝑊/𝑚2 𝐾 (37)

𝑅𝑎 = 𝐺𝑟 × 𝑃𝑟 (24) Therefore the external heat transfer

Grashof number is computed as, coefficient, hair, overall, is computed as,
𝑔×𝛽𝑎𝑖𝑟 ×[𝑇𝑠 −𝑇 ]𝐿3 ℎ𝑎𝑖𝑟,𝑜𝑣𝑒𝑟𝑎𝑙𝑙 = ℎ𝑐𝑜𝑚𝑏 + ℎ𝑟 (38)
𝐺𝑟 = 2 (25)
𝛾𝑎𝑖𝑟
ℎ𝑎𝑖𝑟,𝑜𝑣𝑒𝑟𝑎𝑙𝑙 = 15.83 + 0.9 = 16.73 𝑊/𝑚2 𝐾 (39)
Where ‘ ’ is the thermal expansion coefficient LNG Heat Load Estimation
of air approximated as 1/Ta. To estimate the heat load ingress into the
𝐺𝑟 =
9.812×0.00333×|206.25−300|×383
≈ 6.75 × 1014 (26) LNG tank, Fourier’s Law of conduction is
0.000015782
applied along the tank wall, tank roof & tank
The Rayleigh number and Nusselt number is, bottom. For tank wall, the heat load is
𝑅𝑎 = 6.75 × 1014 × 0.7293 = 4.93 × 1014 (27) computed as,
2 −2𝜋𝐻𝑚𝑎𝑥 [𝑇𝐿𝑁𝐺 −𝑇𝑠 ]
1⁄
𝑄𝑊 = (40)
∑ 𝑅1
0.387× (4.93×1014 ) 6
𝑁𝑢 = {0.825 + } = 8497(28)
9
0.492 ⁄16
8⁄
27 Where R1 is the resistances through the tank
[1+(0.7293) ]
wall.
Therefore the external free convection heat 𝐴1 =
1
=
1
= 0.0016 𝑚. 𝐾/𝑊 (41)
𝑟𝑜𝑢𝑡 ℎ𝑎𝑖𝑟,𝑜𝑣𝑒𝑟𝑎𝑙𝑙 37×16.73
transfer coefficient is calculated as,
1 𝑟1 1 37.2
𝑘𝑎𝑖𝑟 𝑁𝑢𝑓𝑟𝑒𝑒 0.0256×8497
𝐴2 = 𝑙𝑛 [ ]= 𝑙𝑛 [ ] = 6 × 10−5 𝑚. 𝐾/𝑊 (42)
2 𝑘9% 𝑁𝑖 𝑟𝑖𝑛 90.9 37
ℎ𝑓𝑟𝑒𝑒 = = ≈ 5.73 𝑊/𝑚 𝐾 (29)
𝐿𝑀𝑎𝑥𝑂𝑝.𝐿𝑒𝑣𝑒𝑙 38
1 𝑟 1 37.8
𝐴3 = 𝑙𝑛 [ 2] = 𝑙𝑛 [ ] = 0.4211 𝑚. 𝐾/𝑊 (43)
To determine whether the heat transfer mode 𝑘𝑝𝑒𝑟𝑙𝑖 𝑟1 0.038 37.2

is dominated by free/natural convection or 𝐴4 =

1 𝑟
𝑙𝑛 [ 3 ] =
1
𝑙𝑛 [
37.95
] = 9 × 10−5 𝑚. 𝐾/𝑊 (44)
𝑘𝐶𝑆 𝑟2 42.6 37.8
forced convection, the condition that is 1 𝑟 1 38.55
𝐴5 = 𝑙𝑛 [ 4] = 𝑙𝑛 [ ] = 0.541 𝑚. 𝐾/𝑊 (45)
required to be satisfied is as follows [4], 𝑘𝑝𝑜𝑙𝑦 𝑟3 0.029 37.95

𝐺𝑟 1 𝑟4 +𝑟8 1 39.15
] = 0.0086 𝑚. 𝐾/𝑊 (46)
≪ 1 ; Free convection is negligible (30) 𝐴6 =
𝑘𝑐𝑜𝑛𝑐
𝑙𝑛 [
𝑟4
]=
1.8
𝑙𝑛 [
38.55
𝑅𝑒 2
1 1
𝐺𝑟 𝐴7 = = = 0.00073 𝑚. 𝐾/𝑊 (47)
≫ 1 ; Forced convection is negligible (31) 𝑟𝑖𝑛ℎ𝐿𝑁𝐺 39.15×35
𝑅𝑒 2
𝐺𝑟 ∑ 𝑅1 = 𝐴1 + 𝐴2 + 𝐴3 + 𝐴4 + 𝐴5 + 𝐴6 + 𝐴7 (48)
≈ 1 ; Mixed Convection Exists (32)
𝑅𝑒 2 ∑ 𝑅1 = 0.973 𝑚. 𝐾/𝑊 (49)
Therefore, checking for mode of convection, The heat load through the tank wall is,
𝐺𝑟 6.75×1014
= (2.41×107)2 = 1.1652 ≈ 1 (33) 𝑄𝑊 =
−2𝜋×38×[112.5−206.25]
= 23,005 𝑊 (50)
𝑅𝑒 2
0.973

Page 89
The heat load through the Tank bottom is 𝑄𝐿𝑁𝐺 = 𝑄𝑊 + 𝑄𝐵 + 𝑄𝑅 (70)
computed in linear coordinates as, 𝑄𝐿𝑁𝐺 = 23 + 60.2 + 54.3 ≈ 137.5 𝑘𝑊 (71)
−𝜋𝑟 2 [𝑇𝐿𝑁𝐺 −𝑇𝑠𝑜𝑖𝑙 ]
𝑄𝐵 = ∑ 𝑅2
(51) With the heat load estimated, the BOG is,
𝑄𝐿𝑁𝐺 ×24×3600 137.5×24×3600
Where R2 is the tank bottom resistance, 𝐵𝑂𝐺 = =
505.5
≈ 23,494𝑘𝑔/𝑑(72)
𝐿𝑁𝐺

1 1
𝐵1 = ℎ = 35 = 0.0286 𝑚2 𝐾/𝑊 (52) For the LNG Tank volume of 160,000 m3,
𝐿𝑁𝐺
𝐵𝑂𝐺×100 23494×100
𝑟5 0.2 2 𝐵𝑂𝐺 𝑅𝑎𝑡𝑒 = = ≈ 0.032% (73)
𝐵2 = 𝑘 = 90.9 = 0.0022 𝑚 𝐾/𝑊 (53) 𝑉𝐿𝑁𝐺 ×𝜌𝐿𝑁𝐺 160000×456.5
9% 𝑁𝑖

𝑟6 0.6 From the estimated BOG Rate of 0.032 vol%,

𝐵3 = 𝑘 = 0.045 = 13.34 𝑚2 𝐾/𝑊 (54)
𝑓𝑜𝑎𝑚 it is lower than the allowable 0.05 vol%/day.
𝑟 0.15
𝐵4 = 𝑘 7 = 42.6 = 0.0035 𝑚2 𝐾/𝑊 (55) Note: The above steps are repeated (under
𝐶𝑆
the assumption Twall = Tb,LNG) until the
𝑟8 0.6
𝐵5 = 𝑘 = 1.8 = 0.3334 𝑚2 𝐾/𝑊 (56) average wall temperature converges, i.e.,
𝑐𝑜𝑛𝑐
Iteration 2, Tb,LNG,2 becomes,
1 1
𝐵6 = ℎ = 2 = 0.5 𝑚2 𝐾/𝑊 (57)
𝑚2 𝐾 𝑄 𝑊
𝑠𝑜𝑖𝑙
𝑇𝑏,𝐿𝑁𝐺,2 [𝐾] = 𝑇𝑏,𝐿𝑁𝐺,1 [𝐾] − 𝑅𝑖𝑛𝑠 [ ] × 𝐴 [𝑚2 ]
𝑊
∑ 𝑅2 = 𝐵1 + 𝐵2 + 𝐵3 + 𝐵4 + 𝐵5 + 𝐵6 (58)
𝑄 𝑊 𝑇𝑤𝑎𝑙𝑙 [𝐾]−𝑇𝑎𝑚𝑏 [𝐾]
2 [ ]= 𝑚2 𝐾
∑ 𝑅2 = 14.2 𝑚 𝐾/𝑊 (59) 𝐴 𝑚2 𝑅𝑡𝑜𝑡𝑎𝑙 [ ]
𝑊

The heat load through the tank Bottom is, References & Further Reading
−𝜋×39.152 [112.5−290]
𝑄𝐵 = = 60,186 𝑊 (60) 1. https://ww2.energy.ca.gov/lng/faq.html
14.2
2. “Analysis of boil-off rate problem in
Similarly, the heat load through the Tank Roof
Liquefied Natural Gas (LNG) Receiving
is computed in linear coordinates as,
Terminals” T Wodek 2019 IOP Conf. Ser.:
−𝜋𝑟 2 [𝑇𝐿𝑁𝐺 −𝑇𝑎 ]
𝑄𝑅 = (61) Earth Environ. Sci. 214 012105
∑ 𝑅3

Where R3 is the tank roof resistance, 3. “Development of the World’s Largest Above
Ground Full Containment LNG Storage
1 1
𝐶1 = ℎ = 35 = 0.0286 𝑚2 𝐾/𝑊 (62) Tank”, Young-myung Yang, Ji-hoon Kim,
𝐿𝑁𝐺

𝑟10 0.0005
Heung-seok Seo, Kangwon Lee, Ihn-soo
𝐶2 = 𝑘 = = 0.00006 𝑚2 𝐾/𝑊 (63)
Yoon, Korea Gas Corporation, 23rd World
9% 𝑁𝑖 90.9

𝑟6 0.3 Gas Conference, Amsterdam 2006.

𝐶3 = 𝑘 = 0.038 = 7.8947 𝑚2 𝐾/𝑊 (64)
𝑝𝑒𝑟𝑙𝑖𝑡𝑒
4. “Fundamentals of Heat and Mass Transfer”,
𝐶4 = 𝑘
𝑟8 0.6
= 1.8 = 0.0033 𝑚2 𝐾/𝑊 (65) Incropera, DeWitt, Bergman, Lavine, 6th
𝐶𝑜𝑛𝑐
Edition.
1 1
𝐶5 = ℎ = 16.73 = 0.0598 𝑚2 𝐾/𝑊 (66) 5. “Weathering of Stored Liquefied Natural
𝑎𝑖𝑟,𝑜𝑣𝑒𝑟𝑎𝑙𝑙
Gas (LNG)”, Calogero Migliore, Amin Salehi,
∑ 𝑅3 = 𝐶1 + 𝐶2 + 𝐶3 + 𝐶4 + 𝐶5 (67)
Velisa Vesovic, Department of Earth
∑ 𝑅3 = 8.32 𝑚2 𝐾/𝑊 (68) Sciences & Engineering, Imperial College
The heat load through the tank roof is, London, Repsol S.A, calle de Mendez
Alvaro, 44, 28045, Madrid, Spain
−𝜋×39.152 [112.5−206.25]
𝑄𝑅 = ≈ 54,281 𝑊 (69)
8.32 6. “LNG Custody Transfer Handbook”, GIIGNL,
The total heat load transferred to the LNG is, 5th Edition, 2017.
Page 90
 Module 14
Gas Condensate Separation Stages – Design & Optimization
The life cycle of an oil & gas venture begins at
the wellhead where subsurface engineers
work their way through surveying, drilling,
laying production tubing and well
completions. Once a well is completed,
gathering lines from each well is laid to
gather hydrocarbons and transported via a
main trunk line to a gas oil separation unit
Figure 2. Three Stage Separation
(GOSP) to be processed further to enhance
their product value for sales. Gas condensate General Notes
wells consist of natural gas which is rich in 1. Stage Separation consists of series of
heavier hydrocarbons that are recovered as separators which flash the incoming well
liquids in separators in field facilities or gas- fluids consisting of gas, oil and water into
oil separation plants (GOSP). their respective constituents. The
oil/condensates exiting the separation
The following is aimed at demonstrating how vessel is eventually routed to a storage
to optimize and provide the required number tank, a.k.a, Stock Tank which operates at
of separation stages to process a gas atmospheric conditions. The liquid in the
condensate mixture and separate them into stock tank are in turn termed as stabilized
their respective vapour phase and liquid crude. To increase vaporization,
phase – termed as “Stage Separation”. Stage sometimes a heater is installed in the
separation consists of laying a series of liquid side of the separation stages.
separators which operate at consecutive
2. The art of flashing well fluids in successive
lower pressures to strip out vapours from the
separation stages to increase liquid yield is
well liquids & resulting in a stabilized liquid.
not linear but inversely proportional to the
Prior to any hydrocarbon processing in a gas
number of stages. Therefore the liquid
processing plant or a refinery, it is imperative
recovery with an extra stage to a single
to maximize the liquid recovery as well as
stage system can be substantial. However
provide a stabilized liquid hydrocarbon. A
adding an extra stage to more than three
schematic of a 2-Stage Separation Unit is as
stages does not produce a significant yield
follows,
in liquid quantities. In theory, as the
number of separation stages increase, so
will the liquid yield increase and the gas
and liquid reach equilibrium. In practice, a
three stage separation process is very
efficient and cost effective to arrive at a
stabilized stock.
3. In terms of terminology, a 2-Stage
Figure 1. Two Stage Separation
separation system means one separator
A schematic of a 3-Stage Separation Unit is as and an atmospheric storage tank (stock
follows, tank) because the storage tank also acts as
Page 91
 a separator. Similarly, a 3-Stage separator tubing pressure (FTP) wells. 3-Stage
consists of a two separators & a stock tank. Separation is for intermediate GOR,
4. The number of stages required to provide a intermediate API gravity oils &
gas and liquid at equilibrium conditions, intermediate FTP wells. 4-Stage Separation
referred to as flashing can be estimated by is for high GOR, high API gravity oils & high
empirical correlations or using Equation of FTP wells.
state (EoS). In this module, the method of 10. Separation vessels also consist of gas
EoS is used. control valve on the gas side which also
5. Separator calculations primarily consist of serve the purpose of controlling the
optimum separator pressure and wellhead back pressure.
temperature, Gas-Oil Ratio (GOR), API 11. For effective operation, minimizing energy
Gravity of the Stock Tank Oil, Oil formation & costs, the primary separator pressure
factor (a.k.a Formation Volume Factor, should be lower than the Wellhead FTP
FVF/Bo) and the respective compositions. and higher than the gas pipeline export
6. The key to selecting the optimum number pressure. Failure to keep the primary
of stages is based on, Minimum GOR, separator pressure higher than the
Maximum API gravity & Minimum Oil FVF. pipeline pressure would require the
installation of gas compressors to boost
7. Minimum GOR implies maximum liquid
gas pressure to export requirements.
yield. GOR is termed as the cumulative gas
flow from all the separators including the Case Study
stock tank divided by the amount of A Gas-Condensate well with an FTP of 14.48
oil/condensate exiting the stock tank. bara [210 psia], 450C and liquid fraction of
Higher API gravity implies, higher is the 1.0, flows to a gas oil separation unit at a
commercial value of the oil/condensate. Oil production rate of 78,295 STBPD [533.3
FVF can be defined as the volume of Am3/h].
reservoir fluid required to produce a barrel Assumptions
of stock tank oil. Therefore lower Oil FVF
1. The stock tank pressure & temperature is
implies more stock tank oil for a given
constant at atmospheric conditions and
volume of reservoir fluids. Oil FVF values
weather remains unchanged.
typically range between 1 bbl/STB to 3
bbl/STB. 2. The composition of the well fluids is taken
to be constant for a given FTP.
8. Since the process conditions of the stock
tank are already fixed to be at atmospheric 3. Both Gas and liquid are assumed to be at
conditions while the primary separator is complete equilibrium upon flashing.
determined by the operating conditions 4. The well fluids composition is chosen to
required to avoid inhibiting well be free from H2S, CO2, N2 and produced
production, the only control had, are the water. Presence of water would require a
operating pressures in the intermediate 3 phase separator. For this module, a 2-
stages. Therefore the intermediate Phase separator is used considering only
pressure is controlled to optimize and gas and condensate.
obtain the highest amount of liquid yield. The gas condensate composition of the well
9. 2-Stage separation is applicable for low fluids is as follows,
GOR, low API gravity oils & low flowing
Page 92
 Table 1. Gas Condensate Composition 𝑃𝑝𝑟𝑖𝑚𝑎𝑟𝑦
𝑃𝑖 = (5)
Mole 𝑅𝑖
Components Units
Fraction Where,
Hydrogen Sulphide [H2S] 0.0000 -
Pi = Intermediate Pressure at Stage ‘i’
Carbon Dioxide [CO2] 0.0000 -
Therefore considering a maximum number of
Nitrogen [N2] 0.0000 - stages of 3, with the well fluids data, for a two
Methane [C1] 0.0385 - stage separation unit, the primary separator
Ethane [C2] 0.0391 - pressure [P1] and Separation ratio [R] is,
14.48
Propane [C3] 0.0516 - 𝑙𝑛[
1.01325
]
𝑛= = 2.42 ~ 2 𝑆𝑡𝑎𝑔𝑒𝑠 (6)
iso-Butane [iC4] 0.0145 - 𝑙𝑛[3]
1⁄
n-Butane [nC4] 0.0575 - 14.48 2
𝑅=[ ] = 3.7801 (7)
1.01325
Iso-Pentane [iC5] 0.0231 -
14.48
n-Pentane [nC5] 0.0346 - 𝑃1 = = 3.831 𝑏𝑎𝑟𝑎 (8)
3.78011
n-Hexane [C6] 0.0491 - The stock tank becomes the second stage and
Heptane Plus [C7+] 0.6920 - the operating pressure is P2 = 1.01325 bara.
Total 1.0000 - For a three stage separation unit, the primary
Heptane Plus [C7+] SG 0.8576 separator pressure [P1] and secondary
separator pressure [P2] is,
Heptane Plus [C7+] MW 227 lb/lbmole
1⁄
14.48 3
𝑅=[ ] = 2.4266 (9)
Methodology 1.01325
The number of separators can be chosen by 14.48
𝑃1 = = 5.968 𝑏𝑎𝑟𝑎 (10)
first arriving at a preliminary pressure 2.42661
14.48
estimate based on equal pressure ratio as, 𝑃2 = = 2.459 𝑏𝑎𝑟𝑎 (11)
2.42662
𝑃𝑝𝑟𝑖𝑚𝑎𝑟𝑦
𝑋𝑛 = [ ] (1) The stock tank becomes the third stage and
𝑃𝑆𝑡𝑜𝑐𝑘
the operating pressure is P3 = 1.01325 bara.
Where,
Using these preliminary stage pressures, a
Pprimary = Primary Separator Pressure [bara]
vapour-liquid equilibrium (VLE) flash
PStock = Stock Tank Pressure [bara] calculation can be performed to estimate
n = Number of stages [-] GOR, API gravity and Oil FVF. For this module,
X = Maximum number of Stages [-] the Peng Robinson EoS is chosen. To arrive at
the final separator pressures, the following
Rewriting the expression,
iterative procedure is adopted.
𝑃𝑝𝑟𝑖𝑚𝑎𝑟𝑦
𝑛 × 𝑙𝑛𝑋 = 𝑙𝑛 [ ] (2)
𝑃𝑆𝑡𝑜𝑐𝑘 1. Keeping all preliminary estimated
𝑃𝑝𝑟𝑖𝑚𝑎𝑟𝑦 pressures fixed, the primary separator
𝑙𝑛[ ]
Or, 𝑛 =
𝑃𝑆𝑡𝑜𝑐𝑘
(3) pressure is varied for a pressure range to
𝑙𝑛𝑋 obtain GOR, Oil FVF & API gravity. Making
The separation ratio is computed as, a plot of the above values, the separator
𝑃𝑝𝑟𝑖𝑚𝑎𝑟𝑦
1⁄
𝑛 pressure corresponding to Min GOR, Max
𝑅=[ ] (4) API gravity & Min FVF is chosen [1st
𝑃𝑆𝑡𝑜𝑐𝑘
Iteration of 1st Stage].
The intermediate pressure is computed as,

Page 93
2. The primary separator initial estimate 38.08. The Oil FVF decreases from 1.0466
pressure is now replaced with the 1st m3/m3 to 1.0435 m3/m3.
Iteration’s optimized pressure. Following Table 3. GOR, Oil FVF and API Gravity
further, the secondary separator pressure Stages GOR Oil FVF API Gravity
is also varied for a given range to similarly
- [scf/STB] [m3/m3] [0API]
obtain an optimized pressure
corresponding to Min GOR, Max API gravity 2 90.77 1.0466 37.93
& Min Oil FVF [1st Iteration of 2nd stage]. 3 86.68 1.0435 38.08
3. The secondary separator initial estimate
The plots of Total GOR, API Gravity & Oil FVF
pressure is now replaced with the
for two stage design and three stage design
optimized value, [1st Iteration of 2nd stage].
are as follows,
4. With the 1st iteration optimized pressures,
flash calculations are repeated similar to
Step 2 and Step 3, i.e., 2nd Iteration and so
forth, until a converged solution is reached.
Results
With the flash procedure applied for the
calculated initial estimates, the results of 2-
Stage Separation and 3-Stage Separation for a
compressor polytropic efficiency of 82% are,
Table 2. Separation Stage Pressures Figure 3. Two Stages – GOR & API Gravity
Operating Compressor
Stages
Pressure Power
- [bara] [kW]

2-Stage Separation [Liquid Rate = 75,235 STBD]

FTP 14.48

Primary Separator 2.60 83.04

Stock Tank 1.014

3-Stage Separation [Liquid Rate = 75,430 STBD]

FTP 14.48 Figure 4. Two Stages – Oil FVF

Primary Separator 3.90 71.47

Secondary Separator 1.70 65.32

Stock Tank 1.014

From the results between the 2-Stage and 3-

Stage separation, the liquid yield increased by
195 STBD, i.e., 0.26%. The GOR, Oil FVF and
API gravity for the separation stages are also
tabulated below. Between both cases, the
GOR decreases from 90.77 scf/STB to 86.68 Figure 5. Three Stages – Primary GOR & API Gravity
scf/STB. API gravity increases from 37.93 to
Page 94
 References & Further Reading
1. “New Method to estimate Surface Separator
Optimum Operating Pressures”, Kegang
Ling, Xingru Wu, Boyun Guo, Jun He, SPE,
Oil & Gas Facilities, June 2013
2. “Example problems for the calculation and
selection of compressors”, Intech GMBH,
(https://intech-
gmbh.com/compr_calc_and_selec_example
Figure 6. Three Stages – Primary Oil FVF s/)
3. “Emulsions and Oil Treating Equipment”,
Maurice Stewart, Ken Arnold, 2009
(https://www.sciencedirect.com/science/
article/pii/B9780750689700000025)
4. “Natural Gas (Second Edition)”, James G
Speight, 2019.
5. Aspentech HYSYS 2006.5 Documentation –
Dynamic Modelling. Aspentech HYSYS is a
Figure 7. Three Stages – Sec. GOR & API Gravity
registered Trademark of Aspentech

Figure 8. Three Stages – Secondary Oil FVF

Conclusions
From the calculations made, a three stage
separator offers the advantage of higher
liquid yield. The Gas-Oil Ratio (GOR) can also
be decreased which allows to recover more
liquids and thereby offering higher
commercial value. Although a 4th stage can be
added to increase liquid yield, since well
pressures are low at ~15 bara, 3 stages would
suffice. It is also seen that with increase in
number of stages, the LP compressor power
requirements also decreases.

Page 95
Appendix A

Page 96
Appendix B

Page 97
 MODULE 15
PROCESS DESIGN FOR INSTRUMENT AIR SYSTEMS
Industrial process facilities consist of a wide removing water vapour using a pressure
variety of pneumatically operated equipment swing adsorption (PSA) process.
which needs to be provided with a motive IA Receiver Design Considerations
force for operation. Towards this, ambient air A key component of the Instrument Air
is one of the commonly used motive fluids to System is the Air Receiving Vessel. The IA
operate. In an Oil & Gas project, the primary
Receiver when designed must consider the
step is to assess the number of elements that
following factors,
need instrument air (IA) and capacities of
each element (e.g., Control valve) to 1. Minimize Pressure Fluctuations – To
determine the required instrument air system meet IA consumers’ demands and during
capacity. The following module focuses on emergency shutdown scenarios, it becomes
sizing an IA air receiver vessel as well as some a necessity to reduce pressure fluctuations.
of the design considerations to made for an IA This also means that sufficient pressure at
system. a steady rate must always be available for
processes that use IA and is measured in
Typical Layout & Operation of IA System
units of time (minutes).
A typical layout of an Instrument Air system is
2. Short Term Air Demand – In process
shown below,
facilities often the demand for instrument
air (IA) can fluctuate sometimes reaching a
peak. This needs to be accounted for in the
air compressor capacity estimates along
with sufficient storage volume in the
associated IA receiver to accommodate the
peak demand IA flow rates.
3. Energy Savings – Instrument Air Systems
run frequently consuming power and
Figure 1. Schematic of Compression System
becomes imperative to achieve power
The main components of an instrument air
savings by operating (loading/unloading)
system consist of an inlet air filter to
the compressor only as and when required.
decontaminate the atmospheric air of dust
When the pressure in the IA Receiver drops
and debris, an air compressor to produce
below a threshold, the IA compressor is
instrument air at the required pressure, a
loaded to achieve the required pressure in
cooler to cool the hot air from the compressor
the IA Receiver. Sizing the IA Receiver for
discharge, a moisture separator to remove
longer cycles enables to cut own on power
any condensates from the compressed air, an
consumption while providing a steady flow
air receiver that stores the compressed air, a
of IA to the end users.
set of molecular sieve air dryers that act as a
desiccant to dry the instrument air to the Instrument Air Quality Standards
required dew point. Air dryers are operated A commonly used industry standard to set
in cycles whereby when one dryer is under instrument air quality standards is the
operation, the other dryer is regenerated by ANSI/ISA –S7.0.01-1996. As per the standard,
Page 98
1. Pressure Dew Point – Pressure dew point charged/discharged through a 10 sec cycle &
is defined as the temperature at which free the operating pressure band between lower
moisture condenses out from the and upper pressure of the IA receiver is 10
instrument air into liquid water for a psi. From Steam Tables, moisture content in
specific pressure. Pressure dew point at the free air is as follows,
air dryer outlet should be at least 100C Table 1. Mass of H2O in Air (kg.H2O/m3 Free Sat. Air)
(180F) below the minimum temperature to Temperature Pressure
which any part of the IA system is exposed
[oC] 0 barg 8 barg
and also shall not exceed 40C (390F) at the
line pressure. 0 0.0045 0.00051

2. Particle Size – The IA supplied to 20 0.018 0.0019

consumers is expected to contain 40 0.059 0.0062
particulate matter and for most 60 0.18 0.017
pneumatically operated devices, a
80 0.65 0.041
particulate size of 40 m is acceptable. In
cases where, particulate size of < 40 m is Water Condensation in IA Receiver
required, additional air filtration modules To estimate the amount of water condensing
can be installed to achieve < 40 m sized in the wet air IA Receiver, the mass of water
particulate matter. in air at 100% RH is taken from Table 1.
3. Lubricant Content – In cases where the 1. Water Content in saturated air [100% RH]
installed Air compressor in the IA system is entering Comp. [0 barg] = 0.018 kg H2O/m3
lubricated with lube oil, there is always a 2. Water Content in IA compressor suction at
risk of oil carry over along with the 60% RH = RH x Water Content at 100% RH
compressed air. This poses a threat to the = 0.6 x 0.018 = 0.0108 kg H2O/m3 Air
pneumatic devices that receive the
3. Water Content at IA Comp. Discharge
instrument air and affects their operation.
[100% RH] at 8 barg, 300C = 0.00341 kg
Hence the lubricant content should be close
H2O/m3 Air [Table 1].
to 0 ppm but cannot exceed 1 ppm w/w.
4. Water extracted from compressed air
4. IA System Location – The location of the
discharge & drained via IA receiver liquid
Instrument Air system is also important to
outlet, 0.0108–0.00341=0.00739 kg
prevent contaminants, hazardous and
H2O/m3
flammable gases from being drawn into the
Inlet air filters. 5. Water drain rate in IA Wet Receiver =
0.00739 (600 x 0.0283168) 60 =
Instrument Air System Design
7.5 kg.H2O/h (Note: 0.0283168 is
To evaluate the process data of an instrument conversion factor for ACFM to Am3/s, i.e.,
air unit, an example case study is used to [600 0.0283168] / 60 = 0. 283168 Am3/s.
explain. An Instrument air package is to be
6. Relative Humidity [RH] of Air Leaving the
designed to deliver 600 ACFM of dry air at 8
IA Receiver to Air Dryer = [0.00341/0.018]
barg to pneumatic device users. The ambient
= 19%
location is 200C, with a relative humidity (RH)
of 60%. The barometric pressure is 1.01325 Pressure Dew Point in IA Inlet & Receiver
bara. The IA delivered to the IA receiver is The pressure dew point of the instrument air
required to be ~300C. The IA receiver is processed at the IA compressor inlet and IA
Page 99
receiver exit can be calculated using Arden- f = Charge/discharge per IA receiver Cycle [s]
Buck equation as follows,
PU–PL = Pressure band of IA Receiver [psia]
T T
RH (b− )( )
γm [T, RH] = ln [ e d c+T ] (1) Pa = Barometric Pressure at Location [psia]
100
c × γm [T,RH] The volume of the IA receiver is computed as,
TDew Point = [ ] (2)
b−γm [T,RH]
600×10×14.7
𝑉𝐼𝐴 𝑅𝑒𝑐𝑒𝑖𝑣𝑒𝑟 = [ ] = 147 𝑓𝑡 3
Where, 60×10

Constant ‘b’ = 18.678; Constant ‘c’ = 257.14 (10)

0C; Or, 𝑉𝐼𝐴 𝑅𝑒𝑐𝑒𝑖𝑣𝑒𝑟 ≈ 4.2 𝑚3 (11)
Constant ‘d’ = 234.50 0C; IA Dew Point at Atmospheric Pressure
RH = Relative Humidity The performance guarantee parameter for
Therefore, the pressure dew point at IA most industrial IA systems is based on a
compressor Inlet is, typical dew point requirement of -400C at
60 20 20 atmospheric pressure at the outlet of the air
] (3)
(18.678− )( )
𝛾𝑚 [200 𝐶, 60%] = 𝑙𝑛 [ 𝑒 234.5 257.14+20
dryer. In this module, no calculations are
100

shown for air dryer unit, however taking the

𝛾𝑚 [200 𝐶, 60%] = 0.831 (4)
IA receiver process conditions, it can be
257.14×0.831
TDew Point = [18.678−0.831] ≈ 120 C (5) estimated, what should be the air dryer’s
The pressure dew point at the air leaving the pressure dew point (i.e., at 8 barg) to achieve
IA Receiver is computed as, a performance guarantee dew point of -400C
30 30
at 1 atm at the air dryer outlet.
19 (18.678− )( )
𝛾𝑚 [300 𝐶, 19%] = 𝑙𝑛 [100 𝑒 234.5 257.14+30 ]
To estimate the dew point at atmospheric
(6) pressure, the following dew point graph
𝛾𝑚 [300 𝐶, 19%] = 0.274 (7) between atmospheric pressure & indicated
257.14×0.274
pressure is used.
𝑇𝐷𝑒𝑤 𝑃𝑜𝑖𝑛𝑡 = [18.678−0.274] ≈ 3.80 𝐶 (8)

Instrument Air Receiver Size

The Instrument Air Receiver which collects
the compressed gas is sized based on the
principle of excess pressure in the IA receiver
volume in which the quantity of stored
compressed air is above the facility’s
requirements. Using the actual volume flow
rate flowing into the IA Receiver, the storage
volume, taking into account the time cycle for
charging/discharging, pressure band &
barometric pressure, can be computed as,
Q ×f×Pa
VIA Receiver = [ (Pc ] (9)
U −PL ) Figure 2. Air Dew Point Conversion Chart
Where, From Fig. 2, for a pressure dew point of 3.80C
Qc = Instrument Air Capacity [ACFM] at 8 barg in IA Receiver, the dew point at 1atm
is -290C. Therefore, to achieve an atmospheric
Page 100
dew point of -400C at the air dryer outlet, the 4. IA distribution lines must be sized with line
pressure dew point temperature should be - P < 1 bar between the air dryer outlet and
140C. To estimate the relative humidity [RH] the farthest user of IA. Typically, a user can
at air dryer outlet for an atmospheric dew be taken to use 0.015 m3 (0.5 Scf) of
point of -400C, Eq. (1) and Eq. (2) can be re- air/min.
arranged, 5. Air dryer regeneration methods are of two
types – Air purge regeneration & Heater
𝑏×𝑇
𝛾𝑚 [𝑇, 𝑅𝐻] = [𝑐+𝑇 𝐷𝑃 ] (12) regeneration. Air purge regeneration is a
𝐷𝑃

eγm [T,RH]
commonly used method where the packed
RH = [ T T ] × 100 (13) column of molecular sieves is dried by
(b− )( )
e d c+T
diverting a fraction of the dry air from the
Assuming the temperature rise in the air
active air dryer vessel enabling adsorption
dryer is 400C, the RH of the IA exiting the air of the moisture and expelling via a purge
dryer is, line. Whereas in heater regeneration
18.678×(−14) methods, a heater-blower setup is installed
𝛾𝑚 [𝑇, 𝑅𝐻] = [257.14+(−14)] = −1.075 (14)
in the regeneration line that heats ambient
𝑅𝐻 = [
𝑒 −1.075
] × 100 = 2.8% (15) air & routes the heated air through the
40 40
𝑒
(18.678− )(
234.5 257.14+40
)
regenerating dryer. The hot air heats the
Table 2. IA System Results Summary regenerated dryer till the moisture reaches
Parameter Value Units boiling point and is subsequently expelled
through the purge line.
H2O Extracted in IA Receiver 0.0074 kg.H2O/m3
6. The regeneration time of each air dryer
H2O Condensate in IA Receiver 7.5 kg.H2O/h
shall not be more than 6 hours as per IPS–
Air RH leaving IA Receiver 19 %
G-IN-200(2). The recommended cycle time
TPressure Dew Point - IA Comp. Inlet 12.0 0C
between regeneration cycles for normal
TPressure Dew Point - IA Receiver 3.8 0C operation is 6 hours for regeneration and 2
TPressure Dew Point – Air Dryer Exit -14.0 0C hours of standby. The maximum allowable
cycle time between regeneration cycles is 6
Instrument Air Receiver Size 4.16 m3
hours for regeneration and 4 hours
IA System Design Considerations standby. Hence the air dryers must be
1. In process facilities, it is prudent to install designed to be capable of drying for at least
IA systems with a 1W + 1S configuration. 10 hours without increase in dew point.
The standby can be diesel driven or steam 7. The air dryer adsorption operation is
driven, subjected to the utility available. exothermic & causes the dried air to reach
2. For most utility applications, nominal as high as 600C. If its temperature is not
instrument air line pressure for the utility expected to cool to ~400C, additional after-
industry should be ~690 kPa (100 psi). coolers would be required at the air dryer
3. Since most industrial facilities operate with outlet.
IA air at about 7 barg to 8 barg, the set 8. After coolers can be air-cooled type or
pressure of the relief valve (RV) must be water-cooled type. Water-cooled
set higher accordingly but must not exceed aftercoolers are usually sized to cool outlet
the vessel design pressure. air to within ~5°C to 8°C of the inlet
cooling water temperature. Whereas Air-
Page 101
 cooled aftercoolers are usually sized to cool & ball valves. Globe valves provide the
outlet air to within 14°C to 17°C of the advantage of regulating system flow rates
ambient air temperature & provide tight shut-off. On the down side,
9. Compression increases the partial pressure they cause reduced flow rates, increased
of the water vapour present. If the water pressure loss, and allow places for
vapour partial pressure is increased to the particulates to collect causing valve
saturation water vapour pressure, leakages. Gate valves & Globe valves are
condensation occurs. If the saturation used for on/Off isolation & provide full,
water vapour pressure is reduced to the line-size port for air flow with minimal
partial pressure of the water vapour pressure drop and are conducive to
present, water or ice will result. Therefore, internal cleaning. The disadvantages of
moisture removal is a major consideration gate valves are that they allow particulates
of instrument air treatment systems. Water to collect in disc guides, and valve discs can
droplets entrained in the air can initiate the separate from their stems. The
formation of rust or other corrosion disadvantage of ball valves is that they are
products which block internal passageways more expensive than comparably-sized
of electric to pneumatic converters globe or gate valves, and their sealing
resulting in sticking and/or binding of surfaces are susceptible to leakage from
moving parts. Water droplets can also particulate scoring.
obstruct the discharge ports on solenoid References & Further Reading
air pilot valves thus reducing their ability
to function properly. Therefore, an 1. ANSI/ISA-7.0.01-1996, “Quality Standard
for Instrument Air”, 12th Nov 1996
automatic drain (e.g., timer drain, float
drain or an electronic drain) with a manual 2. IPS-G-IN-200(2), “General Standard for
bypass should be located near the bottom Instrument Air Systems”, 2nd Ed, Feb 2016
of the air receiver to dispose of the 3. Engineering ToolBox, (2008), Compressed
condensate. Air and Water Content
10. In cold climates, water extracted from the 4. https://www.omnicalculator.com/physics/
atmospheric air accumulates at the low air-density
points in the IA system. Hence, in such cold 5. http://www.nationalcompressedair.com/d
climates, insulation and steam tracing at/files/600.pdf
should be provided to both piping as well
ANNEXURE A: Saturation Vapour Pressure
as up to a sufficient height from the bottom
at Dew Point & Actual Vapour Pressure
portion of the air receiver.
The saturation vapour pressure of air at its
11. For the design and construction of the
dew point can be calculated as,
vessels, ASME Boiler and Pressure Vessel
7.5×𝑇𝐷𝑒𝑤 𝑃𝑜𝑖𝑛𝑡 ×14.7
[ ]
Code, Section VIII, Div. 1 or any other 𝑃1 = 6.1078 × 10 𝑇𝐷𝑒𝑤 𝑃𝑜𝑖𝑛𝑡 +237.3
(16)
approved standard of equivalent authority
Actual vapour pressure is, 𝑃𝑣 = 𝑃1 × 𝑅𝐻 (17)
is acceptable. For design, fabrication,
erection and testing of piping ASME B16.5 For air leaving IA receiver at TDP 3.80C,
and ASME B31.3 are acceptable. 19%RH
7.5×3.8×14.7
12. In IA distribution systems, commonly used 𝑃1 = 6.1078 × 10[ 3.8+237.3
]
= 8 ℎ𝑃𝑎 (18)
types of valves are globe valves, gate valves
𝑃𝑣 = 8 × 0.19 = 1.59 ℎ𝑃𝑎 (19)
Page 102
ANNEXURE B: MS Excel Calculation Sheet
Plant Instrument Air Receiver - Process Design Tool
Standard Pressure [PS] 1.01325 [bara] Mass of Water in Air (kg.H2O/m3 free saturated air)
Standard Temperature [TS] 273.15 [K] Temperature Pressure (barg)
o
Ambient Pressure [P1] 0 [barg] ( C) 0 2 4 6 8
0
IA Compressor Suction Temperature [T 1] 20 [ C] 0 0.0045 0.0015 0.00091 0.00065 0.00051
Ambient Air Relative Humidity [RH] 60 [%] 20 0.018 0.0058 0.0035 0.0025 0.0019
IA Compressor Discharge Pressure [Saturated] [P2] 8 [barg] 40 0.059 0.019 0.011 0.0079 0.0062
0
IA Compressor Cooler Temperature [Saturated] [T 2] 30 [ C] 60 0.18 0.053 0.031 0.022 0.017
4973 [SCFM] 80 0.65 0.14 0.078 0.054 0.041
Compressor Flow Capacity [Q C] 600 [ACFM] 100 - 0.38 0.19 0.13 0.094
3
0.2832 [Am /s] 120 - - 0.49 0.29 0.21
3
Water Content at 100% RH [Atmospheric Conditions] 0.0180 [kg.H2O/m free air]
3
Water Content in IA Compressor Suction [60% RH] 0.0108 [kg.H2O/m free air]
Water Content in IA Compressor Discharge [100% RH] 0.0034 [kg.H2O/m3 free air]
3
0.0074 [kg.H2O/m free air]
Water Extracted inIA Receiver & Sent to Drain
0.690 [g.H2O/kg free air]
Water in Air Leaving the IA Receiver 0.318 [g.H2O/kg free air]
Water Drain Rate in Wet Air IA Receiver 7.5 [kg.H2O/h]
Air Relative Humidity [RH] - IA Receiver Exit to Air Dryer 19 [%]
Dew Point Calculation [Arden-Buck Method]
Constant 'a' 6.1121 [mbar]
Constant 'b' 18.678 [-]
Constant 'c' 257.14 [0C]
Constant 'd' 234.5 [0C]
g(T, RH) at IA Compressor Inlet 0.8309 [-]
Pressure Dew Point [TPressure Dew point ] at IA Compressor Inlet 12.0 [0C]
g(T, RH) at IA Receiver Exit 0.2741 [-]
Pressure Dew Point [TPressure Dew point ] at IA Receiver Exit 3.8 [0C]
Instrument Air Receiver Size
IA Compressor Capacity [Q C] 600 [ACFM]
Charge/Discharge per IA Receiver Cycle [f] 10 [sec]
Pressure Band of IA Receiver [PU-PL] 10 [psi]
147 [ft3]
Instrument Air (IA) Receiver Size [VReceiver]
4.16 [m3]
Air Dryer Process Design
0
Required Dew Point at 1 atm Pressure at Air Dryer Outlet -40.0 C
Temperature Rise in Air Dryer 40.0 [0C] Pressure Dew Point at 8 barg - Air Dryer Outlet = -14 0 C

Pressure DP required at Air Dryer Outlet [8 barg] for -40C Atm DP [DP Conv. Graph] -14.0 [0C]
g(T, RH) at IA Air Dryer Outlet -1.075 [-]
Relative Humidity [RH] Required at Air Dryer Outlet for -40C Atm DP 2.82 [%] Dew Point at Atm Pressure - IA Receiver = -29 0 C
SUMMARY
Water Extracted & Sent to IA Receiver Drain 0.00739 [kg H2O/m3]
Water Drain Rate in Wet Air IA Receiver 7.5 [kg H2O/h] Density of Air [Based on RH]
TPressure Dew point at IA Comp. Inlet 12.0 [0C] Saturation Vapour Pressure at Dew point [p 1] 8.0 [hPa]
TPressure Dew point at IA Receiver Exit 3.8 [0C] Actual Vapour Pressure [p v ] 1.52 [hPa]
0
TPressure Dew point Required at Air Dryer Outlet -14.0 [ C] Pressure of Dry Air [p d] at IA Receiver Exit 9012 [hPa]
Instrument Air (IA) Receiver Size [VReceiver] 4.16 [m3] Density of Air [r Air] at IA Receiver Exit 10.71 [kg/m3]

Page 103
 Module 16
Understanding High Integrity Pressure Protection Systems
No chemical process facility is immune to the Temperature (HPHT) environments are also
risk of overpressure to avoid dictating the increasingly becoming common. Standard
necessity for overpressure protection. For design methods involve designing the entire
every situation that demands safe well head to export systems to fully rated
containment of process gas, it becomes an conditions (1500#, 900#, etc.) depending on
obligation for engineers to equally provide the operating pressures and temperatures.
pressure relieving and flaring provisions However such methods would unnecessarily
wherever necessary. The levels of protection increase project costs and affect installation
are hierarchical, starting with designing an foot print depending on how flammable or
inherently safe process to avoid overpressure toxic is the process fluid, sometimes to the
followed by providing alarms for operators to point of not giving any viable cost benefits.
intervene and Emergency Shutdown To attend to such unviable scenarios, the
provisions through ESD and SIL rated concept of de-rating Non-HPHT equipment in
instrumentation. Beyond these design and downstream operations with overpressure
instrument based protection measures, the protection can be employed. For these
philosophy of containment and abatement purposes, HIPPS is treated as a Safety
steps such as pressure relieving devices, Instrumented System (SIS) that is based on a
flares, physical dikes and Emergency Safety Integrity Level (SIL). From an SIL
Response Services is employed. perspective, HIPPS follows a minimum of SIL
High Integrity Pressure Protection Systems 3 rating where the Average Probability of
(HIPPS) are related to the third layer of Failure on demand is of the order between
protection whereby process shutdown can be ≥10-4 to <10-3. It must be noted that HIPPS is
initiated by shutdown valves that receive an SIS that aids more as risk reduction for
instructions from a logic solver which in turn prevention measure rather than a risk
are fed by pressure transmitters. mitigation measure. The typical architecture
of HIPPS is shown in the Figure below.
In the oil and gas industry, process facilities
are often subjected to erratic fluctuations in
wellhead pressure and flow trends. Such
process systems in recent years are tended to
for overpressure protection with the
installation of HIPPS. HIPPS aid in shutting
down instead of having to flare sour gas
through pressure relieving devices that are
subsequently routed to a flare system.
Figure 1. Example of HIPPS Architecture
The following covers key guidelines and
requirements for HIPPS from industry HIPPS Operating Philosophy
experience and, standards. A typical HIPPS architecture consists of three
With world gas demand increasing steadily (3) pressure transmitters (PT) that constantly
over the years, High Pressure High record the line pressure which are fed to a
logic solver. In the event of an overpressure,
Page 104
the logic solver initiates a shutdown operation (PST), Tight Shut-off (TSO) (e.g., Class V or
of two (2) consecutive Fail-close (FC) valves Class VI of ANSI FCI 70-2), Fast acting, Fire
which are installed on the same line thereby Safety tested to for example, API 607.
shutting down fluid flow. A pressure alarm Environmental constraints must also be met
(PA) serves the purpose of informing the for fugitive emissions such as ISO 15848-1
operation personnel. The purpose of installing standards.
the said number of transmitters and valves
HIPPS Engineering Standards
are as follows,
HIPPS can cater to many applications such as
1. To avoid compromising the HIPPS
offshore/onshore well heads, flare headers
functionality due to failure of any one
and chemical process industries. ASME
shutdown valve (SDV), a second valve is
Section VIII, UG-140 (Overpressure Protection
added to provide higher redundancy. Both
Systems) provides a range of applications for
valves are operated on a 1oo2 voting
which HIPPS can be used, such as,
philosophy that decides which Fail-Close
(FC) valve closes. 1. High Propagation Chemical reactions
resulting in loss of containment prior to the
2. To avoid receiving a premature or false
relief device opening or processes that
signal from the pressure transmitter, a
yield impractical large vent areas
2oo3 voting philosophy is employed as
against a 1oo3 voting philosophy. This 2. Runaway Polymerization, Exothermic or
means that unless 2 pressure transmitters Reactive reactions that produce large
concur that there is an overpressure vapour rates rendering relief devices
scenario, HIPPS is not activated. insufficient to cater to over-pressurization
scenarios.
HIPPS Valve Selection To keep the module brief, the focus is made on
HIPPS Valves can be operated hydraulically or Oil and Gas applications. HIPPS for the Oil and
by solenoid methods. The two (2) types of Gas industry are based on two aspects –
valves used are either ball type or butterfly prescriptive and performance based.
type. Ball valves provide the best shutoff Standards such as API, ASME, ANSI to suggest
conditions and can range from 2 inch to 56 a few are for design, manufacture and
inch depending on the manufacturer. Whereas implementation and examples are API 14C
butterfly valves can be provided from 2 inch (Recommended Practice for Analysis, Design,
to 100 inch, again depending on the Installation, and Testing of Basic Surface
manufacturer. For HPHT applications, the Safety Systems for Offshore Production
piping class can vary from as high as 2500# Platforms), API 6A (Specification for Wellhead
which can be provided in the ball class range and Christmas Tree Equipment) for offshore
with material ranging from carbon steel, applications, API 520/521, API 17O (Subsea
stainless steel, duplex as well as special alloys. High Integrity Pressure Protection Systems –
The typical stroke time for HIPPS valves HIPPS) to name a few. The other aspect is the
should be of the order of <2 sec. Valve IEC standards, chiefly IEC 61508
selection must also consider that HPHT supplemented by IEC 61511 which are more
applications can witness temperatures as high of performance-based standards that describe
as 5000C. HIPPS valves must also be able to how to arrive at a solution rather than
cater to Partial Stroke Testing capability prescribing a solution. This would leave room

Page 105
for elucidation between different operators, mitigated risk is lower than the risk tolerance
contractors and suppliers thereby resulting in criteria, to allow for the removal of associated
lack of commonly accepted industry relief devices from flare load calculations.
specifications. The IEC 61508, for example Traditionally, pressure vessels are equipped
focuses much on the functioning of the logic with pressure relieving devices that are
solver and touching minimally on the final routed to an industrial flare. However when
control element. These gaps left in IEC 61508 the flare load capacity is insufficient to deal
regarding final control elements such as with excess capacities, HIPPS offers the
valves and solenoids are covered in IEC advantage of risk reduction, for example,
61511. shutting down well heads in oil and gas
IEC provides SIL ratings with Probability of applications. But this also obliges the engineer
Failure on Demand (PFD) and respective to ask, if when a HIPPS system is installed,
architecture not for individual components, does it always necessitate pressure relieving
but for the system as a whole which must devices to be installed as well.
include the actuators, initiators, final control Pressure vessels that operate above
elements and logic solvers. When different atmospheric conditions of 15 psig are
manufacturers assume certain architecture designed as per Code ASME Section VIII
for HIPPS components provided, the Division 1 and to cover matters of
individual components Probability of Failure overpressure protection, UG-125 to UG 140 of
on Demand (PFD) would not necessarily the said code provides basic requirements.
represent the overall system’s PFD which is The ASME UG-140 requirements and
used to define the SIL rating. Therefore the procedures are commonly known as Code
PFD for a SIL assessment needs to always be Case 2211.
investigated on a case to case basis prior to The requirements of a relief device covered by
understanding the limitation on the SIL rating UG-125 to UG-138 are to be designed as per
arrived at. API 521. For cases where the requirement of a
relief device can be overcome is based on UG-
Pressure Relieving Devices vs. HIPPS
140(a) and UG-140 (b) of the said code which
A point of contention arises when one asks, if pertains to Inherently Safe Design and HIPPS
when a piece of equipment is equipped with based design under specific cases
multiple relieving devices to deal with
respectively. Industrial use of HIPPS certainly
overpressure scenarios, wherefrom arises the
provides the option of installing a smaller
necessity to install a HIPPS. To suggest so,
sized relieving device but cannot eliminate the
means a justification is required to install
necessity of relieving devices, although in
HIPPS. For any successful implementation of
certain specific cases, the need for PRV’s can
HIPPS, an examination of applicable
be eliminated.
regulations, standards, local codes and
insurer’s requirements that may mandate the As per API 521 and Code Case 2211 of ASME
need for relieving devices is required. This is Section VIII, Division 1 and 2, HIPPS is
to be followed up by a Hazard Analysis allowed in lieu of a Pressure relieving device
(HAZAN) by a multi-disciplinary team. The provided HIPPS meets or exceeds the
process risk needs to be evaluated based on protection that would have been provided by
frequency and consequence such that the the PRV. However as per, ASME Section VIII,
HIPPS proposed can demonstrate that the Division 1, para UG-125(a) Section VIII,

Page 106
Division 2, para, AR-100, it is required to The disadvantage of employing multiple
install a pressure relieving device on all suppliers causes increased lead time as well
pressure vessels. as procurement costs. An alternative would be
Therefore, the question of whether a PRV is to source HIPPS from a single manufacturer
necessary in tandem with HIPPS depends on who can provide all individual components
identifying credible overpressure scenarios in and have it certified by an independent SIL
the operating system prior to installing certifying body. This reduces the lead time
relieving devices. HIPPS typically can be found required for procurement as well as costs
in applications where hazardous gases are associated.
part of critical operations. Any addition of a
References & Further Reading
relieving device acts more like insurance to
1. ‘Introduction to Process Unit Flares’,
the safety of the process.
Kolmetz. K, KLM Technology Group
HIPPS Procurement Life Cycle 2. ‘Application of UG-140 for Overpressure
HIPPS system which consists of various Protection’, Sushant G Labhasetwar,
components such as logic solvers, actuators, Chemical Engineering World, September
valves, pressure transmitters can be supplied 2013
by various manufacturers. A Request for 3. ‘Consideration in Designing HIPPS’, Willem-
Quotation (RFQ) is placed with different Jan Nuis, Rens Wolters, www.safan.com
manufacturers by the procurement division of 4. ‘High Integrity Protection System (HIPS) for
the EPC contractor which in turn is provided
Flare Load Mitigation’, Angela E. Summers,
to the engineering teams such as process, SIS-TECH Solutions, LP
piping, Instrumentation and Safety
departments.

Figure 2. HIPPS Procurement and Certification Life Cycle

The EPC procurement team has to perform an

additional task of project management to
ensure all associated items in the bill of
quantities (BOQ) are received and handed
over to the engineering team. The integrated
HIPPS components would then require a SIL
certification by an independent certifying
body for SIL 3 requirements before being
implemented at the End User’s facility.

Page 107
 Module 17
Pressure Safety Valve (PSV) Sizing - API 520/521/526
No chemical process facility is immune to the 2. Balanced Bellow Safety Relief Valve -
risk of overpressure to avoid dictating the A balanced safety relief valve provides a set
necessity for overpressure protection. For of bellows to reduce the effect of back
every situation that demands safe pressure on the operational characteristics.
containment of process gas, it becomes an 3. Pilot Operated Safety Relief Valve -
obligation for engineers to equally provide In a pilot operated safety relief valve the
pressure relieving and flaring provisions major relieving device has a self-actuated
wherever necessary. The levels of protection auxiliary pressure relief valve to control
are hierarchical, starting with designing an the relieving conditions.
inherently safe process to avoid overpressure
4. Power Actuated Safety Relief Valve -
followed by providing alarms for operators to
In a power actuated safety relief valve, the
intervene and Emergency Shutdown
major relieving device is controlled with an
provisions through ESD and SIL rated
external source of energy.
instrumentation. Beyond this design and
instrument based protection measures, the 5. Temperature Actuated Safety Relief
philosophy of containment and abatement Valve-
steps such as pressure relieving devices, In a temperature-actuated safety relief
flares, physical dikes and Emergency valve, the actuation takes place by external
Response Services is employed or internal temperature or by inlet side
pressure.
A pressure safety valve (PSV) is a safety
device used to protect equipment from over 6. Pressure Vacuum Safety Relief Valve -
pressure conditions. Over pressure refers to A vacuum relief valve is designed to allow
any condition which would cause a system to fluid to prevent excessive internal vacuum
increase beyond the specified design pressure & close to prevent further fluid flow after
or maximum allowable working pressure normal conditions have been restored.
(MAWP). PSVs must open at a predetermined Pressure Safety Valves Terminology
set pressure, flow a rated capacity at a The terminology associated with pressure
specified overpressure, and close when the safety relief Valve can be inferred from API
system pressure has returned to a safe level. Recommended Practice 520 as,
Types of Pressure Safety Valves 1. Set Pressure - The inlet gauge pressure at
which the pressure relief device is set to
Pressure safety valves can be chiefly classified
open under service conditions.
into the following types,
2. Back Pressure - The pressure that exists at
1. Conventional Safety Relief Valve – In a
the pressure relief device outlet as a result
conventional safety relief valve it has a
of the pressure in the discharge system. It
spring housing that vents fluids to the
is the sum of the superimposed and built-
discharge side of the PSV. The operational
up back pressures.
characteristics (opening pressure, closing
3. Built up Back Pressure – This is the
pressure, & relieving capacity) are directly
increase in pressure at the outlet of a
affected by changes in PSV back pressure.
pressure relief device that develops as a
Page 108
 result of flow after the pressure relief relief valve per the preliminary sizing
device opens. equations of API 520.
4. Superimposed Back Pressure – The static 12. Rated Coefficient of Discharge – The rated
pressure that exists at the outlet of a coefficient of discharge is determined in
pressure relief device at the time the device accordance with the applicable code or
is required to operate. It is the result of regulation and is used with the actual
pressure in the discharge system coming discharge area to calculate the rated flow
from other sources and may be constant or capacity of a pressure relief valve.
variable.
PERMITTED SIZES OF PSV –API 526
5. Opening Pressure – This is the value of
As per API 526, the size of pressure safety
increasing inlet static pressure at which
valves are designated by their alphabetical
there is a measurable lift of the disc or at
designation. API 526 gives a list of
which discharge of the fluid becomes
permissible sizes of PSVs which are described
continuous, as determined by seeing,
below,
feeling or hearing. Table 1. Standard Orifices Sizes [API 526]
6. Closing Pressure – The value of decreasing Orifice
Sr. No. in²
inlet static pressure at which the valve disc Designation
re-establishes contact with the seat or at 1 D 0.110
which lift becomes zero as determined by
2 E 0.196
seeing, feeling or hearing.
3 F 0.307
7. Actual Discharge Area – The minimum net
area determining the flow through a valve 4 G 0.503

8. Effective Discharge Area – This is the 5 H 0.785

nominal or computed area used with an 6 J 1.287
effective discharge coefficient to calculate
7 K 1.838
the minimum required relieving capacity
8 L 2.853
for a pressure relief valve per the
preliminary sizing equations contained in 9 M 3.600
API 520. API 526 provides effective 10 N 4.340
discharge areas for a range of sizes in
11 P 6.380
terms of letter designations, D through T.
12 Q 11.05
9. Inlet Size – The nominal pipe size (NPS) of
the valve at inlet connection, unless 13 R 16.00

otherwise designated. 14 T 26.00

10. Outlet Size – The nominal pipe size (NPS) OVER PRESSURE SCENARIOS
of the valve at discharge connection, unless
Sizing a pressure relief valve begins with
otherwise designated.
identifying the applicable credible scenario
11. Effective Coefficient of Discharge – The which determines the relieving capacity.
effective coefficient of discharge is a Based on the relieving capacity, API provides
nominal value used with an effective procedures to estimate the required relieving
discharge area to calculate the minimum area followed by choosing standardized sizes
required relieving capacity of a pressure of relief devices from API 526. The following
Page 109
module covers the below described scenarios 5. Fire Case [Liquid Filled Vessel] – All
which are not comprehensive but represent a equipment in a process facility is prone to
set of commonly encountered scenarios. exposure to fire due to equipment failure
1. Blocked Liquid Discharge Case - This or man-made errors. This can result in the
refers to closure of a valve on the outlet of contents of the equipment fluid, expanding
equipment. With continuing liquid flow and vaporizing to create an over pressure
into the equipment and no provision to scenario. Fire cases are of two types – Gas
drain the liquid, fluid accumulates to Filled Vessel and Liquid Filled Vessel. In the
building pressure to as high as the design case of liquid filled vessels, the vessel is
pressure of the upstream equipment. In expected to contain a certain amount of
addition to this, static head of the liquid in liquid that wets the lower part surface of
the upstream equipment also contribute to the vessel through which the liquid
the build-up of pressure. Therefore, the transfers latent heat causing liquid
minimum relieving rate to be considered is expansion and vaporization.
the normal operating inlet flow. CASE STUDIES
2. Blocked Gas Outlet [Non-Fire Case] – To demonstrate the sizing of pressure safety
Similar to the above case of liquid filled, gas valves for the described scenarios as per API
accumulation in the vessel also contributes recommended procedures, the following
to the rise in pressure when the gas side examples are shown.
valve fails to function by staying closed. Blocked Liquid Discharge Case
With pressure continuing to rise, a relief
device is required to relieve the equipment Consider a vessel relieving hydrocarbon at
of the excess pressure. 300,000 kg/h which has a relief valve
3. Gas Control Valve Fail Open – This case pressure set at 18 barg. Considering a non-
refers to a scenario where when a control fire case, the over pressure is taken to be
valve placed between equipment fails open, 10%. For preliminary sizing, the back
[whereby the upstream equipment has a pressure at the relief valve discharge is
higher design pressure and the considered to vary between 0 barg to 4 barg.
downstream equipment is at a lower A rupture disc exists and the back pressure is
design pressure] causes over considered to be a variable for which a
pressurization. balanced bellow type of relief valve is
4. Thermal Expansion – This case refers to recommended followed by the pressure relief
scenarios where liquid locked inside liquid valve requiring capacity certification as per
lines. With exposure from sunlight, heat ASME Sec VIII, Division I. The hydrocarbon
ingress occurs through the piping causing a
fluid properties are as follows,
temperature rise to vaporize the liquid
Table 2. Fluid Properties – Blocked Liquid Outlet
resulting in over pressure. The quantity of
fluid required to be relieved in Parameter Value Units
temperature safety valves may be very Liquid Specific Gravity 0.85 [-]
small & therefore for thermal expansion Liquid Viscosity 450 cP
cases the safety valve size of NPS ¾" x NPS 1"
(DN 20 x DN25) should be sufficient as per API As per Ref [1], Sec 3.8.1.2, the initial orifice
521, [Ref 4]. size of the PSV is sized as,

Page 110
 11.78×𝑄 𝑆𝐺 The orifice area based on corrected viscosity
𝐴𝑅 = ( ) × √(𝑃 −𝑃 ) (1) is then calculated as,
𝐾𝑑 ×𝐾𝑤 ×𝐾𝑐 ×𝐾𝑣 1 2
𝐴
Where, Q = flow rate, lit/min 𝐴𝑐𝑜𝑟𝑟 = ( 𝑅) (4)
𝐾 𝑉
AR = Required effective discharge area, mm2
Applying the above to estimate the pressure
Kd = rated coefficient of discharge that should relief device, the relief flow rate is,
be obtained from the valve manufacturer. For 𝑚 1000
a preliminary sizing, an effective discharge 𝑄 = ( )×( ) 𝑙𝑖𝑡/𝑚𝑖𝑛 (5)
𝜌 60
coefficient can be used as follows: 0.65 when 300,000
𝑄= ( ) = 5,882 𝑙𝑖𝑡/𝑚𝑖𝑛 (6)
PSV is installed with or without a rupture disk in 850×60
combination & 0.62 when PSV is not installed and 𝑃1 = 𝑃𝑠𝑒𝑡 × 1.1 = 19.8 𝑏𝑎𝑟𝑔 = 1980 𝑘𝑃𝑎𝑔 (7)
sizing is for a rupture disk in accordance with
𝑃2 = 4 𝑏𝑎𝑟𝑔 = 400 𝑘𝑃𝑎𝑔 (8)
3.11.1.2 Ref [1].
The percent of gauge back pressure is,
Kw = correction factor due to back pressure. If
𝑃
the back pressure is atmospheric, use a value %𝑃𝐵𝑎𝑐𝑘 = ( 𝐵 ) × 100 (9)
𝑃𝑠
for Kw of 1.0. Balanced bellows valves in back
4×14.7
pressure service will require the correction %𝑃𝐵𝑎𝑐𝑘 = ( ) × 100 = 22.2% (10)
18×14.7
factor determined from Figure 31 of Ref [1]. Since a rupture disc exists and back pressure
Conventional and pilot operated valves is considered to be a variable, a balanced
require no special correction. bellow type of relief valve is recommended.
Kc = combination correction factor for Hence the coefficient of discharge [Kd] is 0.65.
installations with a rupture disk upstream of The combination correction factor for use of
the pressure relief valve. Kc value = 1.0 when rupture disc [Kc] in combination with a relief
a rupture disk is not installed & 0.9 when a valve and in the absence of any published
rupture disk is installed in combination with a value is 0.9. The correction factor due to back
pressure relief valve and the combination pressure [Kw] for balanced bellows is
does not have a published value. determined from Fig 31 of Ref [1] as 0.955.
Kv = correction factor due to viscosity. For
conventional or pilot operated relief valve Kv
can be taken as 1.0. For balanced bellows as
determined from Figure 36 of Ref [1] or from
the following equation:
2.878 342.75 −1
𝐾𝑉 = (0.9935 + + ) (2)
𝑅0.5 𝑅1.5

SG = Liquid specific gravity at flowing

temperature referred to water at standard
conditions. R = Reynolds Number as,
𝑄×18800×𝑆𝐺
𝑅= (3)
𝜂×√𝐴𝑅

P1 = upstream relieving pressure, kPag (Set

pressure plus allowable overpressure) Figure 1. Capacity Correction Factor, Kw, due to
P2 = back pressure, kPag Back Pressure on Balanced-Bellows Pressure Relief
Valves in Liquid Service

Page 111
It is to be noted that, as per Ref [1], Sec Table 3. Fluid Properties – Blocked Gas Outlet
3.3.4.4, the curve above represents values Parameter Value Units
recommended by various manufacturers. This
Specific Heat Ratio [k] 1.2 [-]
curve may be used when the manufacturer is
Compressibility Factor [Z] 0.9428 [-]
not known. Otherwise, the manufacturer
should be consulted for the applicable Gas Molecular Weight [MW] 33 [lbm/lbmol]
correction factor. Therefore, the initial orifice As per Ref [1], Sec 3.6.2.1 and Sec 3.6.3.1,
area sizing with no viscosity correction [Kv], required effective discharge area of a
i.e., Kv = 1.0 is, conventional PSV for critical flow is,
11.78×5882 0.85 𝑊 𝑇𝑍
𝐴𝑅 = ( ) × √1980−400 (11) 𝐴𝑐𝑎𝑙𝑐 = √ (17)
0.65×0.955×0.9×1 𝐶×𝐾𝑑 ×𝑃1 ×𝐾𝑏 ×𝐾𝑐 𝑀𝑊

Where,
𝐴𝑅 = 2877 𝑚𝑚2 ≈ 4.46 𝑖𝑛2 (12)
AR = Required effective discharge area, mm2
The Reynolds number of the relieving fluid is,
Q = flow rate, lit/min
5882×18800×0.85
𝑅= = 389 (13) C = Coefficient determined from an
450×√2877
expression of the ratio of the specific heats (k
The correction factor due to viscosity [Kv] is,
= Cp /Cv) of the gas or vapour at inlet relieving
−1
2.878 342.75
𝐾𝑉 = (0.9935 + (38950.5 ) + (38951.5 )) (14) conditions. Where k cannot be determined, it
is suggested that a value of C equal to 315 be
𝐾𝑉 = 0.961 (15) used. The units for C are
The orifice area based on corrected viscosity, √𝑙𝑏𝑚 ×𝑙𝑏𝑚𝑜𝑙 ×𝑅𝑎𝑛𝑘𝑖𝑛𝑒
𝐶= (18)
4.46 𝑙𝑏𝑓 × ℎ𝑟
2
𝐴𝑐𝑜𝑟𝑟 = ( ) = 4.64 𝑖𝑛 (16)
0.961 Kd = Effective coefficient of discharge. For
From Table 1, the PSV chosen for a calculated preliminary sizing, the following values can
orifice area of 4.64 in2 is a ‘P’ designated be used as follows: 0.975 when PSV is installed
orifice which has an orifice area of 6.38 in2. with or without a rupture disk in combination &
Blocked Gas Outlet – Non-Fire Case 0.62 when PSV is not installed and sizing is for a
rupture disk in accordance with 3.11.1.2 Ref [1].
Consider a vessel relieving hydrocarbon
vapours at 10,000 kg/h which has a relief P1 = upstream relieving pressure, kPaa (Set
valve pressure set at 8 barg relieving at a pressure plus allowable overpressure plus
temperature of 420C. Considering a non-fire atmospheric pressure)
case, the over pressure is taken to be 10%. Kb = Capacity correction factor due to back
For preliminary sizing, the back pressure at pressure. This can be obtained from the
the relief valve discharge is considered to be manufacturer’s literature or estimated for
fairly constant at 4 barg. A rupture disc is preliminary sizing from Figure 30 of Ref [1].
installed upstream of the relief valve and The back pressure correction factor applies to
conventional type of relief valve is balanced bellows valves only. For
recommended. conventional and pilot operated valves, use a
The pressure relief valve requires capacity value for Kb equal to 1.0.
certification as per ASME Sec VIII, Division I. Kc = combination correction factor for
The fluid properties of the hydrocarbon is installations with a rupture disk upstream of
shown below, the pressure relief valve. Kc value = 1.0 when
Page 112
a rupture disk is not installed & 0.9 when a The coefficient of subcritical flow [F2] is,
rupture disk is installed in combination with a
𝑘−1
pressure relief valve and the combination 𝑘 2 1−𝑟 𝑘
𝐹2 = √( )×𝑟 ×[ 𝑘 ] (24)
does not have a published value. 𝑘−1 1−𝑟

T = Relieving temperature of the inlet gas or Where, r is the ratio of back pressure to
vapour, R (°F + 460) [K (°C + 273)] upstream relieving pressure, P2/P1
Z = Compressibility factor for the deviation of 𝐾𝑑
𝑟= (25)
the actual gas from a perfect gas, a ratio 𝑃1
evaluated at inlet relieving conditions. The condition to check whether flow is critical
MW = Molecular weight of the gas or vapour or subcritical is, if the back pressure at relief
at inlet relieving conditions valve discharge is less than or equal to critical
V = Required flow through device, scfm at flow nozzle pressure, then flow is critical, else
14.7 psia, 60°F [Nm3/min at 0°C, 101.325 subcritical.
kPaa] Applying the above to estimate the pressure
G = Gas Specific gravity at standard conditions relief device, the critical flow pressure ratio is,
referred to air at standard conditions [normal 1.2
2 1.2−1
conditions]. In other words, G = 1.00 for air at 𝑃𝑐𝑓𝑟 = [ ] = 0.5645 (26)
1.2+1
14.7 psia and 60°F [101.325 kPaa and 0°C] The upstream relieving pressure is,
The critical flow pressure ratio is, 10
𝑘 𝑃1 = ([8 × 14.5] × [1 + (100)]) + 14.7 (27)
2 𝑘−1
𝑃𝑐𝑓𝑟 = [ ] (19) 𝑃1 = 142.3 𝑝𝑠𝑖𝑎 (28)
𝑘+1
Where, The critical flow nozzle pressure [Pcf] is,
K = Ratio of Specific heat [Cp/Cv] 𝑃𝑐𝑓 = 142.3 × 0.5645 (29)
The critical flow nozzle pressure is, 𝑃𝑐𝑓 = 80.3 𝑝𝑠𝑖𝑎 ≈ 4.5𝑏𝑎𝑟𝑔 (30)
𝑃𝑐𝑓 = 𝑃𝑐𝑓𝑟 × 𝑃1 (20) From the above calculation it is seen that the
𝑃𝑜𝑝 back pressure at relief valve discharge is 4
𝑃1 = ([𝑃𝑠𝑒𝑡 × 14.5] × [1 + (100)]) + 14.7 barg which is less than the calculated critical
(21) flow nozzle pressure value of 4.5 barg. Hence
Where, the flow through the relief valve is critical.
Pset = Set Pressure [psia] Therefore, applying the relevant formulae for
Pop = Over Pressure [%] critical flow through the relief device,
When the value of the ratio of specific heat is 𝑊 = 10,000 × 2.20462 = 22,046 𝑙𝑏/ℎ
known, Coefficient, C is, (31)
𝑘+1 1.2+1
2 𝑘−1
𝐶 = 520√𝑘 × [𝑘+1] (22) 𝐶 = 520√1.2 × [1.2+1]
2 1.2−1
≈ 337 (32)

For subcritical flow, required effective 𝑇 = [(1.8 × 42) + 32] + 459.67 = 567°𝑅 (33)
discharge area of a conventional PSV for
The pressure relief device is installed with a
critical flow is,
rupture disc and Kd is 0.975. The correction
𝑊 𝑍×𝑇 factor due to back pressure Kb is 1.0 for
𝐴𝑐𝑎𝑙𝑐 = √𝑀𝑊×𝑃 ×(𝑃 −𝑃 (23)
735×𝐹2 ×𝐾𝑑 ×𝐾𝑐 1 1 2) conventional relief valve. The combination
Page 113
correction factor for use of rupture disc [Kc] Sizing Coefficients for Single-Ported, Globe-
when rupture disc is installed is 0.9. Style Valve Bodies” (Table 5.10.1) [Ref 2].
Therefore, for the critical flow behaviour the With the above described, taking a single
required effective orifice area is estimated as, ported, globe style valve, cage guided valve
22046 567×0.9428 plug, equal percentage flow characteristic
𝐴𝑐𝑎𝑙𝑐 = √ (34)
337×0.975×142.3×1×0.9 33 with a flow coefficient of Cv of 224 with an
𝐴𝑐𝑎𝑙𝑐 = 2.11 𝑖𝑛2 (35) inlet conditions of 6.5 barg and 350C and
outlet conditions of 5.8 barg and 32.70C
From Table 1, the PSV chosen for a calculated
(5510R) at fail open position.
orifice area of 2.11 in2 is an ‘L’ designated
orifice which has an orifice area of 2.85 in2. Considering a non-fire case, the RV set
pressure is 4.5 barg & over
Control Valve Fail Open Case
pressure/maximum accumulated pressure is
For the case of a control valve fail open case, taken to be 10%. For preliminary sizing, the
the PSV relief rate is determined by the back pressure at the RV discharge is
control valve’s flow capacity. For example, in considered to be fairly constant at 1 barg.
a gas oil separator, the control valve on the
A rupture disc is not installed upstream of the
gas side that feeds to the downstream
relief valve considering no corrosive fluid
equipment, when it fail opens, the
exists and a conventional type of relief valve
downstream PSV’s relieving rate determines
is recommended. The pressure relief valve
the PSV size. In the current module,
requires capacity certification as per ASME
procedures provided by Fisher’s Control
Sec VIII, Division I. The fluid properties of the
Valve Handbook, 5th Edition is used to
hydrocarbon are shown below,
demonstrate an example case. It is to be noted
Table 4. Fluid Properties – Control Valve Fail Open
that Fisher’s Handbook, for the case of sizing
control valves for compressible fluids Parameter Value Units
employs ISA’s standardized procedure, Specific Heat Ratio [k] 1.25 [-]
namely the ANSI/ISA S75.01 for calculating Compressibility Factor [Z] 0.9791 [-]
the required valve flow coefficient, Cv. Flow is
Gas Molecular Weight [MW] 19.37 [kg/kmol]
a dependent variable. Based on the ISA
procedure, control valves can be sized The first step in calculating the PSV size for a
depending on the fluid properties available control valve fail open case begins with
i.e., estimating the fluid flow rate through the
1. Mass Flow Rate & Fluid Density control valve when the valve fails open (say,
2. Mass Flow Rate & Gas Compressibility due to failure of instrument air). This can be
Factor calculated using the expression,
3. Standard Volumetric Flow Rate & Gas 𝑚 𝑇1 𝑍1
Compressibility Factor 𝐶𝑣 = √ (36)
94.8×𝐹𝑃 ×𝑃1 × 𝑌 𝑥×𝑀𝑊
The sizing equations for a control valve are Rearranging to calculate mass flow rate ‘m’,
also influenced by the piping geometry &
𝑥×𝑀𝑊
attached fittings. In this module, it is assumed 𝑚 = 𝐶𝑣 × 94.8 × 𝐹𝑃 × 𝑃1 × 𝑌√ (37)
𝑇1 𝑍1
that the piping geometry is similar to the line
Where,
size valve and no fittings are present. The
𝑥
data related to the control are taken from 𝑌 =1− (38)
3 𝐹𝐾 𝑋𝑇
Fisher’s Handbook based on “Representative
Page 114
 𝑘 ∆𝑃
𝐹𝐾 = (39) 𝐼𝑓 < 𝐹𝑘 𝑥𝑇 𝑡ℎ𝑒𝑛 𝑆𝑢𝑏 𝑐𝑟𝑖𝑐𝑡𝑖𝑐𝑎𝑙 𝐹𝑙𝑜𝑤 (49)
1.4 𝑃1
∆𝑃 10
𝑥 = ∆𝑃⁄𝑃1 (40) = = 0.0932 (50)
𝑃1 109
m = Mass flowrate [kg/h]
𝐹𝑘 𝑥𝑇 = 0.891 × 0.72 ≈ 0.6415 (51)
CV = Flow rate coefficient at rated capacity [-] Therefore, based on the condition for
P1 = Valve upstream absolute pressure [bara] calculating the pressure drop ratio, x is
MW = Gas molecular weight [kg/kmol] 0.0917. The gas expansion Factor [Y] is
T1 = Control valve inlet temperature [K] calculated as,
k = specific heats factor [Cp/Cv] 𝑌 = 1 − 3𝐹
𝑥 0.0932
= 1 − 3×0.891×0.72 = 0.952(52)
𝐾 𝑋𝑇
Z = gas compressibility factor [-]
Therefore, the mass flow through the control
∆P = Pressure drop at rated flow [bar]
valve is estimated as,
Fp = Piping geometry factor (Fp =1) for line
𝑚 = 224 × 94.8 × 1 × 7.5 × 0.952 × √(35+273.15)×0.9791 (53)
0.0932×19.37

sized valve and no attached fittings [-]

FK = Ratio of specific heats factor [-] 𝑚 = 11,742 𝑘𝑔/ℎ (54)
x = Pressure drop ratio [-] Based on the calculated mass flow rate, the
XT = Choked flow pressure drop factor (refer PSV sizing commences & is similar to blocked
to Vendor´s catalogue) [-] gas outlet case. Hence proceeding on similar
lines,
Y = Gas expansion factor [-] 1.25
2
Solving for the problem at hand, 𝑃𝑐𝑓𝑟 = [1.25+1]
1.25−1
= 0.5555 (55)
∆𝑃 = 𝑃1 − 𝑃2 (41) 10
𝑃1 = ([4.5 × 14.5] × [1 + (100)]) + 14.7 (56)
𝑃1 = 6.5 𝑏𝑎𝑟𝑔 ≈ 109 𝑝𝑠𝑖𝑎 (42) 𝑃1 = 86.5 𝑝𝑠𝑖𝑎 (57)
𝑃2 = 5.8 𝑏𝑎𝑟𝑔 ≈ 99 𝑝𝑠𝑖𝑎 (43) 𝑃𝑐𝑓 = 86.5 × 0.5555 (58)

∆𝑃 = 109 − 99 = 10 𝑝𝑠𝑖 (44) 𝑃𝑐𝑓 = 48 𝑝𝑠𝑖𝑎 ≈ 2.3 𝑏𝑎𝑟𝑔 (59)

The ratio of specific heat factor is estimated Therefore, since the PSV back pressure of 1
as, barg is lower than the critical flow nozzle
𝑘 1.247 pressure [Pcf] of 48 psia (2.3 barg), the flow
𝐹𝐾 = 1.4 = = 0.891 (45) behaviour across the PSV is critical.
1.4

The rated pressure drop factor [xT] from 𝑊 = 11,742 × 2.20462 = 25,886 𝑙𝑏/ℎ (60)
Table 5.10.1 for the chosen valve is 0.72 and
1.25+1
the piping geometry factor [Fp] is taken to be 2
𝐶 = 520√1.25 × [
1.25−1
] ≈ 342
1.0 for this example. A condition that has to be 1.25+1

satisfied to calculate the pressure drop ratio (61)

and whether the flow behaviour is critical or 10
𝑃1 = ([4.5 × 14.5] × [1 + (100)]) + 14.7 (62)
subcritical is,
∆𝑃 𝑃1 = 86.5 𝑝𝑠𝑖𝑎 (63)
𝐼𝑓 > 𝐹𝑘 𝑥𝑇 𝑡ℎ𝑒𝑛 𝑥 = 𝐹𝑘 𝑥𝑇 (46)
𝑃1 𝑇 = [(1.8 × 32.7) + 32] + 459.67 = 551°𝑅 (64)
∆𝑃 ∆𝑃
𝐼𝑓 < 𝐹𝑘 𝑥𝑇 𝑡ℎ𝑒𝑛 𝑥 = (47) The coefficient of discharge for the pressure
𝑃1 𝑃1
relief device when installed with or without a
∆𝑃
𝐼𝑓 > 𝐹𝑘 𝑥𝑇 𝑡ℎ𝑒𝑛 𝐶𝑟𝑖𝑐𝑡𝑖𝑐𝑎𝑙 𝐹𝑙𝑜𝑤 (48) rupture disc in combination, Kd is 0.975. The
𝑃1
correction factor due to back pressure Kb is
Page 115
1.0 for conventional relief valve. The pressure, then the pressure relieving device
combination correction factor for use of should be capable of handling the vapour
rupture disc [Kc] when no rupture disc is generation rate. If discovery and correction
installed is 1.0. Therefore, for the critical flow before liquid boiling is expected, then it is not
existing the required effective orifice area is necessary to account for vaporization in
estimated as, sizing the PRD [Ref 4]. Two general
25886 551×0.9791
applications for which thermal-relieving
𝐴𝑐𝑎𝑙𝑐 = √ (65) devices larger than above described are above
342×0.975×86.5×1×1 19.37
ground long uninsulated pipelines of large
𝐴𝑐𝑎𝑙𝑐 = 4.74 𝑖𝑛2 (66)
diameter and large vessels or exchangers
From Table 1, the PSV chosen for a calculated operating liquid full [Ref 4].
orifice area of 4.74 in2 is a ‘P’ designated
Relief Valve Sizing – External Fire Case
orifice which has an orifice area of 6.38 in2.
Liquid Filled Vessel
Relief Valve Sizing – Thermal Expansion For the case of a relief valve sizing – External
For the case of a relief valve sizing – thermal Fire Case Liquid Filled Vessels that are
expansion, consider a vessel with a liquid exposed to fire begins with estimating the
density of 850 kg/m3 (SG 0.85 at 15.60C). The vapour flow rate in the event of a fire.
total heat transfer rate to the vessel is 1000 When vessels are exposed to heat, hydraulic
kJ/s (3,412,140 BTU/h). The specific heat of expansion or thermal expansion is expected
the trapped fluid is 3.9 kJ/kg.K (0.9315 to occur wherein there is an increase in liquid
Btu/lb.0F). The cubic expansion of the volumes due to increase in temperature. In
hydrocarbon is 0.0005 1/0F. The relief valve the case of external fire, API 521 distinguishes
flow rate is calculated as, between wetted vessel & un-wetted vessels. A
𝑞=
𝛼𝑣 ×∅
(67) wetted vessel contains a liquid in equilibrium
500×𝑑×𝑐
with its vapours. During an external fire,
Where, q = Volumetric flow rate at relieving
partial evaporation of liquid occurs, such that
conditions, [USGPM]
the portion of the vessel in contact with the
= cubic expansion coefficient for the liquid liquid within a distance of 25 feet (7.62 m as
at relieving conditions [1/0F] in ISO 23251) receives heat transfer from the
d = Relative Density referred to water [-] exposed fire and must be considered for
c = Trapped fluid’s specific heat capacity sizing. If there is thermal cracking of the
[Btu/lb.0F] vessel occurs leading to vapour generation,
Therefore, the relieving rate is, other alternate sizing methods need to be
0.0005×3412140 considered.
𝑞= (68)
500×0.85×0.9315 An un-wetted vessel is either thermally
𝑞 = 4.3 𝑈𝑆𝐺𝑃𝑀 ≈ 1 𝑚3 /ℎ (69) insulated on the thermal walls or filled with
gases, vapours or super critical fluids. In
Since the flowrates are very small for thermal
comparison to wetted vessels, the thermal
expansion cases the safety valve size of NPS
flow from the walls to the interior are low in
¾" NPS 1" (DN 20 x DN25) should be
un-wetted vessels due to the large thermal
sufficient as per Ref 4. This calculation
resistance. If the vessel is subjected to
method provides only short-term protection
exposure to the fire for prolonged periods, the
in some cases. If the blocked-in liquid has a
vessel temperature would be high enough to
vapour pressure higher than the relief design
cause thermal rupture of the vessel.
Page 116
The total heat of absorption, Q for the wetted The possible mounting position of the
surface can be estimated with adequate partially filled wetted vessels with liquids can
drainage and prompt firefighting facilities as, be horizontal or vertical with hemispherical
𝑄 = 21000𝐹𝐴0.82
𝑤𝑠 (70) or elliptical heads. The figure below gives a
description of the described vessel positions,
Without adequate drainage and prompt
firefighting facilities,
𝑄 = 34500𝐹𝐴0.82
𝑤𝑠 (71)
Q = Heat load/total heat of absorption [Btu/h]
Fenv = Environmental Factor [-]
Aws = Wetted Surface Area [m2]
The value of ‘Fenv’ for a bare vessel, water
application facilities on bare vessels,
depressurizing and empty facilities is taken to
be 1.0. For insulated vessels, the value of ‘Fenv’ Figure 3. Possible mounting positions of the
Partially Filled Wetted vessels with liquids
is taken as per Ref [5], Table T7-6
Table 5. Environmental Factor – Insulated Vessel From the mounting positions described,
Insulation Thermal considering a vessel without a boot, the
[Fenv]
Conductivity effective total height of liquid surface [K1] is,
[W/m.K] [-] 𝐾1 = 𝐻 + 𝐹 (73)
22.71 0.3
If a boot exists at the bottom of the vessel, the
11.36 0.15 vessel elevation, H is computed by subtracting
5.68 0.075 the boot dimensions from the vessel elevation
3.8 0.05 to compute K1 as,
2.84 0.0376 For Hemispherical head,
2.27 0.03 𝑑𝑏𝑜𝑜𝑡
𝐻𝑤𝑖𝑡ℎ 𝑏𝑜𝑜𝑡 = 𝐻 − [ℎ𝑏𝑜𝑜𝑡 + ] (74)
2
1.87 0.026
For Elliptical Head,
Plotting [Fenv] for the insulated vessel vs.
𝑑𝑏𝑜𝑜𝑡
Insulation Thermal Conductivity [K], a linear 𝐻𝑤𝑖𝑡ℎ 𝑏𝑜𝑜𝑡 = 𝐻 − [ℎ𝑏𝑜𝑜𝑡 + ] (75)
4
relationship can be arrived at as, K1 = Effective total height of liquid surface [m]
H = Vessel Elevation without boot [m]
Hwith boot = Vessel Elevation with boot [m]
F = Liquid Level in Vessel [HLL/NLL] [m]
hboot = Height of Boot [m]
dboot = Diameter of Boot [m]
Considering that the portion of vessel
exposed to fire, only that portion in contact
with the liquid within a distance of 25 feet
Figure 2. Environmental Factor Vs. Insulation (7.62 m) above the fire source must be
Thermal Conductivity considered for sizing. Therefore, the effective
i.e, 𝐹𝑒𝑛𝑣 = (0.01319 × 𝐾) + 0.000302 (72) total height of the liquid surface, K1,eff is,

Page 117
𝐼𝑓 𝐾1 < 7.62 𝑚 𝑡ℎ𝑒𝑛 𝐾1,𝑒𝑓𝑓 = 𝐾1 (76) The vapour flow rate generated due to the
𝐼𝑓 𝐾1 > 7.62 𝑚 𝑡ℎ𝑒𝑛 𝐾1,𝑒𝑓𝑓 = 7.62 𝑚 (77) external fire is now calculated as,

Where, K1,eff = Effective total height of liquid 𝑊 = 𝑄⁄ (86)

surface[m] Where, W = Vapour Flow Rate [lb/h]
The initial liquid level, E1 for vessel without a Q = Heat load [Btu/h]
boot is calculated as, = latent Heat of Vapourization [Btu/lb]
𝐸1 = 𝐾1,𝑒𝑓𝑓 − 𝐻 (78) The vapour rate estimated now becomes the
relieving rate for the PSV based on which the
With boot, the initial liquid level, E1 becomes,
PSV orifice size is estimated. For this module,
𝐸1 = 𝐾1,𝑒𝑓𝑓 − 𝐻𝑤𝑖𝑡ℎ 𝑏𝑜𝑜𝑡 (79)
consider a horizontally mounted vessel with
Therefore the effective liquid level, E1,eff that boot with hemispherical ends for both boot
would be used in the calculations are, and vessel. The vessel is insulated and
𝐼𝑓 𝐸1 ≤ 0 𝑡ℎ𝑒𝑛 𝐸1,𝑒𝑓𝑓 = 0 (80) adequate drainage & firefighting measures
𝐼𝑓 𝐸1 > 0 𝑡ℎ𝑒𝑛 𝐸1,𝑒𝑓𝑓 = 𝐸1 (81) are also available. The design parameters are,
Table 6. Liquid Filled Fire Case – Vessel Data
The effective liquid level angle, becomes,
2𝐸1,𝑒𝑓𝑓
Design Parameters Value Units
𝛽 = 𝐶𝑜𝑠 −1 [1 − ] (82)
𝐷 Tank Diameter [D] 4.50 m
= Effective Liquid Level Angle [degrees] Tank Length [T/T] [L] 21.0 m
D = Diameter of the Vessel [m] Boot Diameter [dboot] 1.20 m
It is also considered that a portion of the Boot Height [hboot] 2.00 m
associated piping is also subjected to external
Tank Elevation w/o Boot [H] 5.50 m
heat & is assumed that it is completely filled
with liquid. Additionally, the wetted area Operating Liquid Level [F] 4.00 m

computed with hemispherical ends is Insulation Thermal Conductivity [K] 3.00 W/m.K
considered approximately equal to wetted % Piping Exposed to Fire 20.0 %
area with elliptical heads since a margin is
The fluid properties are as follows,
added to account for piping. The difference is
Table 7. Liquid Filled Fire Case – Fluid Properties
expected to be accommodated in the margin.
Design Parameters Value Units
Taking that the boot in the horizontal vessel is
always liquid filled; the boot wetted area is Ratio of specific heats [k = Cp/Cv] 1.39 -
added to the vessel wetted area. Therefore, Gas Compressibility Factor [Z] 0.53 -
the wetted surface area, Aws for a horizontal Gas Molecular Weight [MW] 79.3 lb/lbmol
vessel without boot is calculated as,
(𝐿−𝐷)𝛽 %𝑝𝑖𝑝𝑖𝑛𝑔
Relieving Temperature [T] 277 0C

𝐴𝑤𝑠 = (𝜋𝐷 × [𝐸1,𝑒𝑓𝑓 + ] ) × [1 + ] (83)

180 100 Latent Heat of Vaporization [ ] 150 kJ/kg
For Horizontal Vessel with boot, Aws becomes,
The RV set pressure is 100 barg & over
(𝐿−𝐷)𝛽
𝐴𝑤𝑠 = (𝜋𝐷 × [𝐸1,𝑒𝑓𝑓 + ] + pressure/maximum accumulated pressure for
180
𝜋 2 ℎ𝑏𝑜𝑜𝑡 %𝑝𝑖𝑝𝑖𝑛𝑔 the fire case is taken to be 21%. For
𝑑𝑏𝑜𝑜𝑡 [ + 1]) × [1 + ] (84)
2 2 100 preliminary sizing, the back pressure at the
For a vertical vessel, Aws becomes, RV discharge is considered to be fairly
𝐴𝑤𝑠 = 𝜋 × 𝐸1,𝑒𝑓𝑓 × 𝐷 × [1 +
%𝑝𝑖𝑝𝑖𝑛𝑔
] (85) constant at 3.5 barg. A rupture disc is not
100
installed upstream of the relief valve
Page 118
considering no corrosive fluid exists and a Therefore, since the PSV back pressure of 3.5
conventional type of relief valve is barg is lower than the critical flow nozzle
recommended. The pressure relief valve pressure [Pcf] of 938 psia (62.8 barg), the flow
requires capacity certification as per ASME behaviour across the PSV is critical.
Sec VIII, Division I. Therefore, to estimate the 1.39+1
relief rate due to the external fire, the tank 2
𝐶 = 520√1.39 × [1.39+1]1.39−1 ≈ 355 (103)
elevation with boot, H for a hemispherical
21
head is calculated as, 𝑃1 = ([100 × 14.5] × [1 + (100)]) + 14.7 (104)
1.2
𝐻𝑤𝑖𝑡ℎ 𝑏𝑜𝑜𝑡 = 5.5 − [2 + ] = 2.9 𝑚 (87) 𝑃1 = 1769 𝑝𝑠𝑖𝑎 (105)
2

𝐾1 = 2.9 + 4 = 6.9 𝑚 (88) 𝑇 = [(1.8 × 277) + 32] + 459.67 = 990°𝑅 (106)

𝑆𝑖𝑛𝑐𝑒 𝐾1 < 7.62 𝑚 → 𝐾1,𝑒𝑓𝑓 = 6.9 𝑚 (89) The coefficient of discharge for the pressure
𝐸1 = 6.9 − 2.9 = 4 𝑚 (90) relief device when installed with or without a
rupture disc in combination, Kd is 0.975. The
𝑆𝑖𝑛𝑐𝑒 𝐸1 > 0 𝑡ℎ𝑒𝑛 𝐸1,𝑒𝑓𝑓 = 4 𝑚 (91)
correction factor due to back pressure Kb is
The effective liquid level angle, is, 1.0 for conventional relief valve. The
𝛽 = 𝐶𝑜𝑠 −1 [1 −
2×4
] ≈ 141° (92) combination correction factor for use of
4.5
rupture disc [Kc] when no rupture disc is
Therefore, the wetted surface area, Aws is,
installed is 1.0. Therefore, for the critical flow
(21−4.5)×141
𝐴𝑤𝑠 = [𝜋 × 4.5 × [4 + ] + behaviour the required effective orifice area is
180
𝜋×1.22 2 20 estimated as,
( × [2 + 1])] × [1 + 100] (93)
2 9,594 990×0.5305
𝐴𝑐𝑎𝑙𝑐 = √ (107)
𝐴𝑤𝑠 = 292.64 𝑚2 ≈ 3,150 𝑓𝑡 2 (94) 355×0.975×1769×1×1 79.3

The Fenv factor is computed as, 𝐴𝑐𝑎𝑙𝑐 = 0.04 𝑖𝑛2 (108)

𝐹𝑒𝑛𝑣 = (0.01319 × 3) + 0.000302 = 0.04 (95) From Table 1, the PSV chosen for a calculated
Therefore, the total heat of absorption or heat orifice area of 0.04 in2 is a ‘D’ designated
load for the insulated vessel with adequate orifice which has an orifice area of 0.11 in2.
drainage and firefighting measures is, References & Further Reading
𝑄 = 21000 × 0.04 × 31500.82 = 618,730 𝐵𝑡𝑢/ℎ(96) 1. “Sizing Selection and Installation of Pressure
The vapour flow rate produced due to the Relieving Devices in Refineries”, API
external fire for of 150 kJ/kg (64.5 Btu/lb), Recommended Practice 520, 7th Edition,
618,730 January 2000, Part -1
𝑊= = 9,594 𝑙𝑏/ℎ 𝑜𝑟 4,352 𝑘𝑔/ℎ (97)
64.5 2. “Control Valve Handbook”, Emerson’s
Based on calculated mass flow rate, the PSV Fisher’s Handbook, 5th Edition
sizing commences & is similar to blocked gas 3. “Pressure Relieving and Depressuring
outlet case. Hence proceeding on similar lines, Systems”, API Standard 521, 6th Edition, Jan
1.39
2 1.39−1 2014
𝑃𝑐𝑓𝑟 = [ ] = 0.5305 (98)
1.39+1 4. “API 521 7th Ed Ballot Item 6.4 Work Item
21
𝑃1 = ([100 × 14.5] × [1 + (100)]) + 14.7(99) 30 – Thermal Expansion Equation
Definitions”
𝑃1 = 1769 𝑝𝑠𝑖𝑎 (100)
5. “Crosby Engineering Handbook”, Technical
𝑃𝑐𝑓 = 1769 × 0.5305 (101)
Doc No. TP-V300, May 1997, Crosby Valve
𝑃𝑐𝑓 = 938 𝑝𝑠𝑖𝑎 ≈ 62.8 𝑏𝑎𝑟𝑔 (102) Inc.
Page 119
 Module 18
Key Process Considerations for Pipeline Design Basis
Prior to venturing into an oil & gas pipeline hydrocarbon. It is to be noted that, in the case
project, the project team would require a of finished products, (e.g., gas oil or Jet A1
design basis, based on which the project is to fuel), many of the auxiliary measures like
proceed. Oil & Gas Pipeline design begins addition of water or mixing with lighter
with a route survey including engineering & hydrocarbons becomes infeasible, since they
environmental assessments. The following affect the product specification.
document provides a few key considerations Vapour Pressure of a liquid is its capacity to
for process engineers to keep in mind, the vaporize/evaporate into its gaseous phase. In
factors that matter when preparing a pipeline pipeline operations, slack flow is a situation
design basis from a process standpoint. where due to the elevational & pipeline
1. Well Production Data/Profile pressure drops, a portion of the hydrocarbon
Well production profiles are required as this experiences pressure below its vapour
determines the size of the pipeline required pressure. As a result, a portion of the liquid
to transport volume/time of fluid. Gases are vaporizes & reaches the high points in the
highly compressible and cannot be treated pipeline. Upon restarting the pipeline, the
the same as liquids such as, crude oils & vapour pockets experience a compressive rise
petroleum distillates. The operating in pressure due to the upstream &
pressures & temperatures are required to be downstream liquid pockets, only to collapse &
known as they determine the design release energy that can rupture pipelines.
conditions of the pipeline. Reid vapour pressures are critical to liquid
petroleum pipeline design, since the pipeline
2. Fluid Physical Properties
must maintain pressures greater than the
The physical properties of the materials being
Reid vapour pressure of the material in order
transported dictate the design and operating
to keep the material in a liquid state.
parameters of the pipeline. Specific gravity,
Pipelines that handle finished products are
compressibility, kinematic & dynamic
preferred to be operated with single phase
viscosity, pour point, and vapour pressure of
flow regime & fully filled pipes. This ensures
the material are the primary considerations.
there is no scope for volatilization that
The pour point of a liquid is the temperature
reduces the scope for fire hazards.
at which it ceases to pour. The pour point for
oil can be determined under protocols set 3. Pumping Costs
forth in the ASTM Standard D-97. Viscous fluids require more power to deliver
In general, crude oils have high pour points. required motive force to the hydrocarbons to
When transported hydrocarbons operate transport them across the pipeline. Waxy
below their pour point, auxiliary measures crudes can be pumped below their pour point
such as heating, diluting with lighter However if the flow is stopped, for e.g., after a
hydrocarbons that are miscible & allows pipeline shutdown, the energy required to
lowering the viscosity & pour point restart the pipeline would be much higher
temperature, mixing with water to allow the than what was required to keep it flowing.
waxes to slide through the pipe walls, or Pipelines also suffer from the formation of
modifying the chemical composition of the hydrates & asphaltenes. Waxes can form
Page 120
crystalline structures that tend to losses are – Hydrostatic pressure drop due to
agglomerate & is referred to as gelling. the pipeline elevation & frictional pressure
Gelling is also a phenomenon that is found in drop which depends on the flow rate. In
storage tanks in production facilities where multiphase pipelines across hilly terrains,
the fluid sits motionless for hours or even hydrostatic pressure drop decreases while
days, resulting in operational difficulties. frictional pressure drop increases with flow.
Hence to attend to these limitations, pour The sum of both these pressure losses gives
point estimation becomes vital to determine the total pressure loss. The pipeline size
if external heating is required. In some cases, chosen should be preferably, the point at
if the waxy crude does not gel enough, it can which the total pressure loss is the least.
get transported to the pump where shear
forces & rise in temperature allow the waxy
crudes to stay above the pour point.
4. Thermal Stresses
Petroleum pipelines are normally buried
unless local regulation prevents them. To do
so, trenches are dug & are laid below
grade/frost line level. Such measures also
provide the advantage of maintaining
relatively constant temperature in line with Fig 1. Pipeline Total Pressure Loss – Hilly Terrain
the ambient/season soil temperature, thereby From the above figure it can be seen that
ensuring the pipeline expansion does not operating a multiphase line at a lower
occur to the point of deflection. Expansion flowrate can actually cost more to pump.
joints as well as in some cases, trenches are
6. Max Hydro Test – ASME B31.8
dug extra wide to accommodate any lateral
For a class1 div 1, the test pressure is limited
movement. In case of river crossings, the
to 125% of the design pressure. For class 3 or
pipeline is to be laid above ground. In
locations that are prone to landslides, buried 4, test pressure is up to 1.4 MAOP.
pipelines option is preferred to avoid direct Therefore, for an 8 hour min. test pressure,
impact of rock structures. But this does not with a design factor of 0.72 for class 1, the
necessarily mean buried pipelines are free test would cause the hoop reaching 72% of
from structural damage, since the weight of Pipeline’s SMYS. Testing at 125% of MAOP
the soil/rock structures deposited above the will result in 1.25 0.72 = 0.9 or 90% of
pipeline can also crush the buried pipelines. SMYS. Hence by hydro testing the pipeline at
1.25 times the operating pressure, the
5. Pipeline Pressure Drop
pipeline is stressed out to 90% of its SMYS.
Pipelines are designed keeping in mind, the The hydro test pressure is based on the
material & construction costs as well location class and maximum test pressure, &
operating costs. Material costs are
becomes the lower value of 8 hour minimum
determined by the pipeline weight, whereas test pressure & test pressure at low point.
operational costs are largely impacted by the
pressure drop experienced which is a 7. Valve Spacing
function of the flow regime. The two key Pipelines need valves to placed & spaced
forces dictating the pipeline total pressure taking into consideration – Rapid
Isolation/Shutdown of pipeline sections to
Page 121
minimize inventory breach, maintain pipeline addition to ASME, several other
design integrity, and facilitate maintenance, organizations, including the API and NACE
repairs & hot tapping operations. Pipelines International, also developed standards used
would also be subjected to pigging & hence by the pipeline industry. Some of the
the valve placement must enable recovery of ASME/API/ANSI standards are,
stuck pigs. 1. “Gas Transmission and Distribution Piping
8. Hydrocarbon Flares Systems,” ASME B31.8, 1999.
Pipelines would sometimes have to be blown 2. “Pipeline Transportation Systems for Liquid
down of any hydrocarbons (liquid or liquid Hydrocarbons and Other Liquids,” ASME
mixed vapours) during events of over B31.4, 1998.
pressure. Burn pit lines serve this purpose. It 3. “Power Piping”, ASME B31.1, 1998;
is important to monitor pilot flames and Addenda B31.1A, 1999; Addenda B31.1B,
provide pilot flame failure alarms. Since burn 2000
pit lines are a source of open flame, they are 4. “Process Piping” ASME B31.1, 1999;
to be located at least 150 m away from Addenda B31.3A, 1999
roadways, process & storage facilities. In
5. “Slurry Transportation Piping Systems”
cases of pipeline in remote locations
ASME B31.11, ‘89; Addenda B31.11A, 1991
requiring maintenance or repair, mobile flare
units can be used. 6. “Pipeline Transportation Systems for Liquid
Hydrocarbons and Other Liquids” ASME
However not all occasions would allow open
B31.4, 2002
flaring, as a result of which, enclosed ground
flares can be used. These conform to the 7. “Gas Transmission and Distribution
requirements of flaring & disposal in Systems,” ASME 31.8, 2003
populated areas or process facilities that are 8. “Specification for Line Pipe”, API 5L, Mar
in close proximity to the flare system. The 2004 / Errata 1, Jan 2005
flaring is smokeless with no visible flame & 9. “Steel Pipelines Crossing Railroads and
noiseless due to insulation of the combustion Highways” API 1102 (1993)
chamber. To attend to the flare capacities 10. “Specification for Pipeline Valves (Gate,
required, a flare study report is to be made Plug, and Check Valves)”, 21st edition, API
part of the design basis. 6D1, June 1998 Supplement 2
9. Pipeline Standards/Codes1 11. Pipeline wall thickness (API B31.G)
ASME has been a pioneer in developing Velocity Considerations3
industry codes & standards for oil & gas
Gas line velocities should be less than 60 to
pipelines. The scope of the first draft of the
80 ft/s to minimize noise & allow for
ASME Code for Pressure Piping, which was
corrosion inhibition. A lower velocity of 50
approved by the American Standards
ft/s should be used in the presence of known
Association in 1935, included the design,
corrosives such as CO2. The minimum gas
manufacture, installation, and testing of oil
velocity should be between 10 and 15 ft/s,
and gas pipelines (ASME B31.4). As the needs
which minimize liquid fallout. The minimum
of the industry evolved over the years, rules
fluid velocity in multiphase systems must be
for new construction have been enhanced,
relatively high to keep the liquids moving in
and rules for operation, inspection, corrosion
order to prevent/minimize slugging. The
control, and maintenance have been added. In
recommended minimum velocity is 10 to 15
Page 122
ft/s. The maximum recommended velocity is The pipeline specification requirement as per
60 ft/s to inhibit noise & 50 ft/s for API 5L plain end line pipe specifications,
CO2 corrosion inhibition. In two-phase flow, it ranges from 6” ND to 80” ND. The product
is possible that the flow stream’s liquid pipeline specification (PSL) with its
droplets can impact the pipe wall causing respective Specified Minimum Yield Strength
erosion of the corrosion products. Erosion of (SYMS) to be used as per API 5L are PSL 1 and
the pipe wall itself could occur if solid PSL 2. The pipeline grades are as follows,
particles, particularly sand, are entrained in Table 2. Product Specification Level (PSL)
the flow stream. SMYS SMYS
Grade Grade
Pipeline Mechanical Design
MPa MPa
As an example to perform Gas Pipeline
mechanical design, ASME B31.8 is used. The PSL 1 Gr A25 172 PSL 2 Gr B 241

requirement to be met for pipeline wall PSL 1 Gr A 207 PSL 2 X42 290
stresses as per ASME B31.8 is Design factor PSL 1 Gr B 241 PSL 2 X46 317
[F], Temperature De-rating [T], Longitudinal
PSL 1 X42 290 PSL 2 X52 359
Joint Factor [E] for the chosen pipeline joining
methods. This is shown below as follows, PSL 1 X46 317 PSL 2 X56 386

Table 1. Reference Mechanical Design Parameters PSL 1 X52 359 PSL 2 X60 414

Design Factors [F] - Gas Pipeline Location PSL 1 X56 386 PSL 2 X65 448

Class Description F PSL 1 X60 414 PSL 2 X70 483

Class 1, Div 1 Deserted 0.80 PSL 1 X65 448 PSL 2 X80 552
Class 1, Div 2 Deserted 0.72 PSL 1 X70 483 - -
Class 2 Village 0.60
Location of the Gas Pipelines
Class 3 City 0.50
1. Class 1 location - A Class 1 location is any
Class 4 Densely Populated 0.40
1-mile pipeline section that has 10 or
Temperature De-rating [T] for Gas Pipelines fewer buildings intended for human
T [0F] T [0C] T occupancy including areas such as,
250 120 1.00
wastelands, deserts, rugged mountains,
grazing land, farmland, sparse populations.
300 150 0.97
2. Class 1, division 1 Location – A Class 1
350 175 0.93
location where the design factor, F, of the
400 200 0.91 pipeline is greater than 0.72 but equal to,
450 230 0.87 or less than 0.80 and which has been
Abbreviation Joining Method E
hydrostatically tested to 1.25 times the
maximum operating pressure.
SMLS Seamless 1.0
3. Class 1, division 2 Location - This is a
ERW Electric Resistance Weld 1.0 Class 1 location where the design factor, F,
EFW Electric Flash Weld 1.0 of the pipeline is equal to or less than 0.72,
SAW Submerged Arc Weld 1.0 and which has been tested to 1.1 times the
maximum operating pressure.
BW Furnace Butt Weld 0.6
4. Class 2 Location - This is any 1-mile
EFAW Electric Fusion Arc Weld 0.8
section of pipeline that has more than 10
Page 123
 but fewer than 46 buildings intended for Gas Pipeline Wall Thickness Estimation
human occupancy including fringe areas The B31.8 code is often used as the standard
around cities and towns, industrial areas, of design for natural gas piping systems in
and ranch or country estates. facilities, such as compressor stations, gas
5. Class 3 Location - This is any 1-mile treatment facilities, measurement &
section of pipeline that has 46 or more regulation stations & tank farms. The B31.8
buildings intended for human occupancy wall-thickness formula is stated as,
except when a Class 4 Location prevails, 𝐷𝑃×𝑂𝐷
including suburban housing developments, 𝑡= (1)
2×𝐹×𝐸×𝑇×𝑆𝑀𝑌𝑆
shopping centres, residential areas, Where,
industrial areas & other populated areas
t = Minimum design wall thickness [in]
not meeting Class 4 Location requirements
6. Class 4 Location - This is any 1-mile DP = Pipeline Design Pressure [psi]
section of pipeline where multi-storey OD = Pipeline Outer Diameter [in]
buildings are prevalent, traffic is heavy or SMYS = Specific Minimum Yield Stress [psi]
dense, and where there may be numerous
F = Design Factor [-]
other utilities underground. Multi-storey
E = Longitudinal Weld Joint Factor [E]
means four or more floors above ground
including the first, or ground, floor. The T = Temperature De-rating Factor [-]
depth of basements or number of References & Further Reading
basement floors is immaterial.
1. “Standard for Gas Transmission and
Line Specification of Gas Pipelines – API 5L Distribution Piping Systems”, ANSI/ASME
1. PSL1 pipes are available through size 2/5” Standard B31.8, 1999
to 80” whereas the smallest diameter pipe
2. “Overview of the Design, Construction, and
available in PSL2 is 4.5” & the largest Operation of Interstate Liquid Petroleum
diameter is 80”. PSL1 pipelines are Pipelines”, ANL/EVS/TM/08-1, Argonne
available in different types of ends, such as
National Laboratory
Plain end, Threaded end, Bevelled end,
3. https://petrowiki.org
special coupling pipes whereas PSL2
pipelines are available in only Plain End.
2. For PSL2 welded pipes, except continuous
welding & laser welding, all other welding
methods are acceptable. For electric weld
welder frequency for PSL2 pipeline is
minimum 100 kHz whereas there is no
such limitation on PSL1 pipelines.
3. Heat treatment of electric welds is
required for all Grades of PSL2 pipes
whereas for PSL1 pipelines, grades above
X42 require it. All kinds of welding method
are acceptable to manufacture PSL1;
however, continuous welding is limited to
Grade A25.

Page 124
 Module 19
Natural Gas Pipeline Transmission Cost and Economics
In any pipeline project, an economic analysis Table 1. Pipeline & Booster Station Details
has to be performed to ensure the project is a Description Value Units
viable investment. The major capital
Upstream Pipeline PSL 1 X 65
components of a pipeline system consists of
the pipeline, Booster station, ancillary Size 16 Inch

machinery such as mainline valve stations, Length incl. fittings [L] 20.4 km
meter stations, pressure regulation stations, Outer Diameter [OD] 406.4 mm
SCADA & Telecommunications. The project
Inner Diameter [ID] 393.7 mm
costs would additionally consist of
environmental costs & permits, Right of Way Wall Thickness [WT] 6.35 mm
(ROW) acquisitions, Engineering & Construction Material (MOC) Carbon Steel
Construction management to name a few.The
MOC Density [ ] 7,850 kg/m3
following module is aimed at conducting a
pipeline economic analysis using the method Downstream Pipeline PSL 1 X 65

of Weight Average Capital Cost (WACC) to Size 16 Inch

estimate gas tariffs, project worth in terms of Length incl. fittings [L] 30.6 Km
Net Present Value (NPV), Internal Rate of
Outer Diameter [OD] 406.4 mm
Return (IRR), Profit to Investment Ratio (PIR)
and payback period. The cost of equity is Inner Diameter [ID] 392.1 mm
estimated using the Capital Asset Pricing Wall Thickness [WT] 7.14 mm
Model (CAPM).
Construction Material (MOC) Carbon Steel
Problem Statement MOC Density [ ] 7,850 kg/m3
A pipeline project is proposed to be built to
Booster Compressor
transport natural gas from a gas processing
facility to a city power station. The scope of Absorbed Power 2,097 kW
gas supply is expected for a period of 25 years Margin on Absorbed Power 20 %
& availability of 8,040 hours/year. Prior Installed Capacity [P] 2,517 kW
calculations indicate a 16” pipeline & a
booster station will be required. The The production profile is as follows,
undertaking requires a CAPEX & OPEX
estimation to explore project viability &
profitability. Project funding is through a
60:40 debt [D] to equity [E] ratio. The cost of
debt is at a lender’s average prime lending
rate [r] of 17%. The cost of equity is taken at
an average risk free interest rate [rf] of 17%
and expected market portfolio return [rm] of
24%. Considering inflation, Annual O&M costs
is taken as 2%. Beta ratio [ ] & linear Figure 1. Production Profile at Processing Facility
depreciation rate [dr] is taken as 0.9 and 8%. The associated costs are as follows,
Page 125
 Table 2. Associated Pipeline Construction Costs Cost Estimation Aspects & Assumptions
Parameter Description 1. The weight average capital cost (WACC)
Total Material Costs [CostPMC] $700/ tonne comprises of a firm’s cost of capital in
Coating & Wrapping [CostPCW] $15/m
which each category of capital is
proportionately weighed. All sources of
Labour Costs [CostLC] $50/m
capital, including common stock &
Install. Cost per inch-km [CostPIC] $11,926/in.km preferred stock, bonds & any other long-
term debt, are included in a WACC
The pipeline installation cost per inch-dia-km
calculation. For the following module, the
is taken by fitting the typical installation cost
cost of equity [ke] & cost of debt [kd], (Sum
data from Table 10.1, page 332 of Ref [1] as,
of which is the cost of capital), expressed
𝑃𝐼𝐶 [$] = −2.103𝐷23 + 153.39𝐷22 − 2883.42𝐷2 +
as a percentage for the pipeline proposal
26606.36 (1)
are included to estimate the WACC.
Table 3. Associated Station Construction Costs [2]
2. The revenue generated comes from the gas
Parameter Description tariffs delivered to the power station.
Station Material Cost [SMC] $2,877/hP Capital is necessary to initiate a project for
Station Labour Cost [SLC] $916/hP which some of the capital can come from
private equity (where investors lend for
Station Misc Cost [SMiC] $367/hP
the project) and debt, where the money is
Station Land Cost [SLaC] $13/hP borrowed from an institution (e.g., bank).
The station material costs is taken as, Ref [2], 3. In order to set the minimum gas tariff &
pay for the equity, lenders/investors or
𝑆𝑀𝐶 [$/ℎ𝑃] = 13035 × 𝑃−0.186 (2)
financial institutions impose a hurdle rate
The station labour costs is taken as, Ref [2], or minimum acceptable rate of return
𝑆𝐿𝐶 [$/ℎ𝑃] = 2274.4 × 𝑃 −0.112 (3) (MARR) (i.e., rate at which NPV = 0 at IRR =
WACC) to offset the cost of the investment
The sales gas tariff allowed for the gas
(i.e., break-even). Generally the hurdle rate
pipeline project is $2/Mscf. The sales tax rate
is equal to the company’s cost of capital. In
[Tsale] on the sold gas is 35% and the effective
the current module cost of capital is taken
corporate tax [TC] is 30%. The Booster station
as a weighted average, the WACC becomes
consists of a gas turbine driven centrifugal
the hurdle rate i.e., at IRR = WACC.
compressor (Single GT/CC). In case of other
configurations, the installed cost can be taken 4. The internal rate of return (IRR) is the
as a factor of Single GT/CC configuration as, expected annual cash flows, (expressed as
a percentage) that the investment can be
Table 4. Associated Station Construction Costs
expected to produce over and above the
Parameter Comparison Factor hurdle rate. Therefore for a project’s
Single GT/CC 1.00 acceptability, IRR > MARR.
Multiple GT/CC 1.29 5. The prime lending rate [r] is the rate
EM/High Speed RC 1.30
imposed by the bank to service the project.
The average risk free interest rate [rf]
High Speed Engine/RC 1.32
represents the rate of return where there
Slow Speed Engine/RC 1.54 is no risk of defaulting/loss. This can be the
interest offered as government/treasury
Page 126
 bonds. In actual practice, inflation is considered to be annually uniform, where
deducted from the interest rate to get a the total life of the asset is divided by the
real risk free interest rate. operational years. Whereas in declining
6. The expected market portfolio return [rm] balance method, the asset is considered to
is a measure of the expected returns on the lose more value in the early years and less
investments made. This is however in later years of operation. For this module,
correlated to market risks that is a straight line depreciation method is
quantified by the beta [ ] ratio and is applied for the stipulated period of 25
useful in determining the volatility when years. It must be noted that the salvage
arriving at the equity costs using CAPM. value at the end of the operational period
<1 indicates that the security issued by becomes zero. However the asset is
the lender providing the equity, is less considered to have a salvage value which is
volatile (i.e., less risky) than the systematic estimated at a depreciation rate of 10% for
this exercise.
market risks. Whereas a >1 indicates that
the security issued by the lender is more 9. Depreciation also relates to taxable
volatile (i.e., more risky) than the income, since the annual depreciation cost
systematic market risks. For this module, is neglected from the annual gross tax on
of 0.9 is taken assuming the venture is by revenue. Annual Cash flow is estimated by
an established firm with a history of low addition of the annual tax benefit from
volatility. In actual practice, low values depreciation [ATBD] to the net annual
need not necessarily show lower volatility revenue [NAR] before subtracting from the
if the security’s value changes minimally in annual gross tax on revenue [AGTR] (i.e.,
the short run but indicates a downward Net Annual revenue – Annual Gross Tax on
trend in the long run. Therefore the Revenue + Annual Tax Benefit from
estimation of requires historical data to Depreciation). It may be noted that when
arrive at a suitable value. companies tend to depreciate the asset
value more, the tax payable decreases and
7. The term [rm – rf] is expressed as the equity
increases the annual savings from
risk premium which indicates the
depreciation.
additional compensation for the risk taken
by the investors/institutions towards 10. The profit to Investment ratio (PIR) is a
providing equity. When multiplied by the measure of investment efficiency. It is
value, it accounts for the responsiveness of estimated as the ratio of NPV to the
the venture to market changes. With discounted capital costs (Disc. CAPEX). For
a project to be viable, the PIR > 1.
higher value, the gains/return on
investment risks is higher. Similarly, with 11. The key differences between using NPV,
lower value, the return on the investment IRR & PIR is that NPV and PIR measure a
risks is lower. project value, while PIR allows screening a
project against a company benchmark. For
8. Assets over time depreciate in value due to
project to be viable, NPV>0, PIR>1 and
wear and tear. Two methods used to
IRR>WACC. The IRR value indicates the
depreciate an asset are the straight line
returns possible above which break even
depreciation method and declining balance
cannot be achieved and NPV<0.
depreciation method. In the straight line
method, the depreciation rate is 12.
Page 127
CAPEX Estimation ECM[$] = [𝑃𝐶 + 𝑆𝐶] × %𝑓𝑎𝑐𝑡𝑜𝑟 (19)

To estimate CAPEX, the pipeline and booster Con[$] = [𝑃𝐶 + 𝑆𝐶] × %𝑓𝑎𝑐𝑡𝑜𝑟 (20)
station tonnage and installed power is
The total project costs are calculated as,
required to be estimated. The pipeline
tonnage is estimated as, 𝑃𝑟𝑜𝑗𝑒𝑐𝑡[$] = 𝐸𝐶𝑃 + 𝑅𝑂𝑊𝐴 + 𝐸𝐶𝑀 + 𝐶𝑜𝑛 (21)

𝜋×𝐿×𝜌 In addition to the pipeline costs, total station

𝑃𝑖𝑝𝑒 𝑇𝑜𝑛𝑛𝑎𝑔𝑒 [𝑡𝑜𝑛𝑛𝑒𝑠] = 4×1000 [𝑂𝐷 2 − 𝐼𝐷 2 ] (4)
costs, other machinery costs & project costs, a
𝑃𝑀𝐶 [$] = 𝐶𝑜𝑠𝑡𝑃𝑀𝐶 × 𝑃𝑖𝑝𝑒 𝑇𝑜𝑛𝑛𝑎𝑔𝑒 (5) working capital [WC] & AFUDC is included.
𝑃𝐶𝑊 [$] = 𝐶𝑜𝑠𝑡𝑃𝐶𝑊 × 𝑃𝑖𝑝𝑒 𝑇𝑜𝑛𝑛𝑎𝑔𝑒 (6) AFUDC which stands for “Allowance for funds
used during construction” represents costs
𝐿𝐶 [$] = 𝐶𝑜𝑠𝑡𝐿𝐶 × 𝑃𝑖𝑝𝑒 𝑇𝑜𝑛𝑛𝑎𝑔𝑒 (7)
associated with financing the project during
𝑃𝐼𝐶 [$] = 𝐶𝑜𝑠𝑡𝑃𝐼𝐶 × 𝑃𝑖𝑝𝑒 𝑇𝑜𝑛𝑛𝑎𝑔𝑒 (8) various stages of construction.
The total pipeline [PC] cost becomes, Total Capex is now calculated as,
𝑃𝐶 [$] = 𝑃𝑀𝐶 + 𝑃𝐶𝑊 + 𝐿𝐶 + 𝑃𝐼𝐶 (9) 𝐶𝐴𝑃𝐸𝑋[$] = 𝑃𝐶 + 𝑇𝑜𝑡𝑎𝑙 𝑆𝐶 + 𝑂𝑀𝐶 + 𝑃𝑟𝑜𝑗𝑒𝑐𝑡 +
The Station Compressor cost is estimated as, 𝑊𝐶 + 𝐴𝐹𝑈𝐷𝐶 (22)

𝑆𝐶[$] = 𝑆𝑀𝐶 + 𝑆𝐿𝐶 + 𝑆𝑀𝑖𝐶 + 𝑆𝐿𝑎𝐶 (10) OPEX Estimation

To estimate annual OPEX, the basis is made
Total Cost of Booster Station based on
on the annual salaries payable [S] and
Comparison factor [Cf] for various GT/CC,
multiplying the annual salary by a %factor in
𝑇𝑜𝑡𝑎𝑙 𝑆𝐶[$] = 𝑆𝐶 × 𝐶𝑓 (11) this module.
The other machinery costs [OMC] including Salary[$] = 𝑆 (23)
mainline valve stations [MVS], meter stations
Payroll[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (24)
[MS] & pressure regulator stations [PRS] can be
computed by accounting them as a percentage of Administrative Expenses[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (25)
the pipeline costs as, Vehicle Expenses[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (26)
𝑀𝑉𝑆[$] = 𝑃𝐶 × %𝑓𝑎𝑐𝑡𝑜𝑟 (12) Office Expenses[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (27)
𝑀𝑆[$] = 𝑃𝐶 × %𝑓𝑎𝑐𝑡𝑜𝑟 (13) Misc Materials & Tools[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (28)
𝑃𝑅𝑆[$] = 𝑃𝐶 × %𝑓𝑎𝑐𝑡𝑜𝑟 (14)
Similarly the related costs for compressor
For SCADA & Telecommunications, the cost is station maintenance is estimated as,
taken as a percentage of pipeline & station, Consumables[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (29)
𝑆&𝑇[$] = [𝑃𝐶 + 𝑆𝐶] × %𝑓𝑎𝑐𝑡𝑜𝑟 (15) Periodic Maintenance[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (30)
The total machinery costs are calculated as, ROW Payments[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (31)
𝑂𝑀𝐶[$] = 𝑀𝑉𝑆 + 𝑀𝑆 + 𝑃𝑅𝑆 + 𝑆&𝑇 (16) Utilities[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (32)
The project costs involving environmental Gas Control[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (33)
costs & permits [ECP], Right of Way
SCADA Maintenance[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (34)
Acquisition [ROWA], Engineering &
Construction Management [ECM], Corrosion Inspection[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (35)
Contingency [Con] is estimated similarly as, Cathodic Protect. Survey[$] = 𝑆 × %𝑓𝑎𝑐𝑡𝑜𝑟 (36)

ECP[$] = [𝑃𝐶 + 𝑆𝐶] × %𝑓𝑎𝑐𝑡𝑜𝑟 (17) The Total O&M [OPEX] is computed as,
ROWA[$] = [𝑃𝐶 + 𝑆𝐶] × %𝑓𝑎𝑐𝑡𝑜𝑟 (18) 𝑂𝑃𝐸𝑋[$] = 𝑆 + 𝐸𝑞. 23 + ⋯ + 𝐸𝑞. 36 (37)
Page 128
CAPEX & OPEX Cost Factors Right of Way (ROW) Acquisitions
Mainline Valve Stations Right of Way Acquisitions can be applied
Block Valves on mainlines are installed to from private and government entities. The fee
isolate pipeline sections for safety & may be lump sum based at the time of
maintenance such as blow down. In case of acquisition, with additional fees paid on an
any pipeline rupture, block valves close to annual basis. The initial cost for acquiring the
isolate the pipeline section and shut-off flow. ROW will be included in the capital cost,
For mainline valve stations a lump sum figure whereas annual ROW lease would be
may be obtained from a construction included as an O&M cost. For most gas
contractor based on the size. pipelines, the initial ROW would be typically
in the range of 6% to 10%.
Meter & Pressure Regulating Stations
Meter Stations are installed for the purpose of Engineering & Construction Management
measuring gas flow rates. Similar to mainline E&C management costs are associated with
valve stations, a lump sum can be quoted for a preparing design documents & drawings at
given size that includes meters, valves, both front end & detailed design phases.
fittings, instrumentation & controls. Pressure Details would also include specifications,
regulation stations are installed for the operating manuals, purchase documents &
purpose of reducing gas delivery pressure equipment acquisition. Construction
prior to delivering to the buyer. The price can management costs would also include costs
be quoted as a lump sum figure. for field personal, transportation, rentals & all
associated costs to manage the pipeline
SCADA & Telecommunications
construction efforts. Engineering &
SCADA allows pipeline process conditions
Construction costs can typically vary from
data to be transmitted by electronic signals
15% to 20% of total project costs.
from remote control units installed on valves
& meters. The signal transmission can be Other Project Costs
either by telephone lines, microwave or As the pipeline project progresses, it is
satellite communications. The pricing of these expected that there could always be some
provisions can be expressed as a percentage unforeseen issues during project execution.
of the total project cost, typically 2% to 5%. To account for these uncertainties,
contingencies & allowances for funds used
Environmental Costs & Permits
during construction (AFDU) are also allotted.
Costs associated with environment and
These costs can vary from 15% to 20% of the
permits pertain to the modification of
total project costs.
pipeline and booster station to prevent
pollution. The costs also include Salvage & Depreciation Costs
compensation for acquisition of land to Pipeline infrastructures experience wear and
compensate for the areas disturbed due to tear thereby rendering depreciation to the
pipeline construction. Permitting costs also project value. The salvage value [SV] based on
include environmental study, environmental the depreciation rate over the operational life
impact report and permits for rail road, of the project is computed as,
stream & river crossings. These costs for a gas 𝑆𝑉[$] = [𝑃𝐶 + 𝑆𝐶 + 𝑂𝑀𝐶 + 𝑅𝑂𝑊𝐴] × [1 − 𝑑𝑟 ]𝑛
pipeline project can vary from 10% to 15% of
(38)
the total project costs.
Where, n = Project Operational Life [Years]
Page 129
The pipeline salvage deduction [SD] is Gas Tariff & Economic Viability
computed as the difference between the The minimum gas tariff required for
Salvage value [SV] & the initial investment economic viability is estimated based on the
cost i.e., [PC+SC+OMC+ROWA]. Therefore the total cost of capital for an NPV = 0 at IRR =
salvage costs is as follows, WACC and yearly gas production. The pipeline
SD[$] = [𝑃𝐶 + 𝑆𝐶 + 𝑂𝑀𝐶 + 𝑅𝑂𝑊𝐴] − 𝑆𝑉 (39) operational availability is 8,040 hours/year.
Therefore, the yearly gas production is,
Annual depreciation of the Asset [ADA]
1000×𝑄𝑀𝑀𝑠𝑐𝑓𝑑
Gas Rate[𝑀𝑠𝑐𝑓] = × 𝐴𝑣𝑎𝑖𝑙𝑎𝑏𝑖𝑙𝑖𝑡𝑦[ℎ𝑟]
ADA[$] = 𝑆𝐷⁄𝑛 (40) 24
(47)
Annual Tax Benefit from Depreciation [ATBD]
The Annual cost of service [ACOS] is
ATBD[$] = 𝐴𝐷𝐴 × 𝑇𝑠𝑎𝑙𝑒 (41) computed as,
Cost of Capital ACOS[$] = 𝑂𝑃𝐸𝑋 + 𝐴𝐴𝐶 + 𝐴𝑅𝐸𝐶 + 𝐴𝐷𝐴 (48)
Based on the cost of capital structure of the The Gas Transportation Tariff is estimated as,
project (that involves both debt & equity), the 𝐴𝐶𝑂𝑆[$]
appraisal is made using the Weight average Tariff𝑇𝑟𝑎𝑛𝑠 [$/𝑀𝑠𝑐𝑓] = 𝐺𝑎𝑠 𝑅𝑎𝑡𝑒[𝑀𝑆𝑐𝑓] (49)
Cost of Capital [WACC] from which all the
Setting an initial guess sales tariff [MinTariff],
cash flows generated will be discounted to
the Net Annual Cash Flow [CF] is calculated
arrive at the Net Present Value [NPV]. WACC
for the condition NPV=0 and IRR=WACC. The
can be expressed as
Annual Revenue [AR] without tax is found as,
𝐸 𝐷
WACC[%] = 𝑘𝑒 + 𝑘𝑑 (42) AR[$] = 𝑀𝑖𝑛 𝑇𝑎𝑟𝑖𝑓𝑓 × 𝐺𝑎𝑠 𝑅𝑎𝑡𝑒[𝑀𝑠𝑐𝑓] (50)
𝐶 𝐶

Where, C = E +D The Net Annual Revenue [NAR] becomes,

The cost of debt [kd] is taken as after-tax cost NAR[$] = 𝐴𝑅 − 𝐴𝐶𝑂𝑆 (51)
of debt and the cost of equity [ke] is based on
The Annual Gross Tax on Revenue [AGTR] is,
CAPM. The cost of debt is estimated as
AGTR[$] = 𝑁𝐴𝑅 × 𝑇𝑠𝑎𝑙𝑒𝑠 (52)
𝑘𝑑 [%] = 𝑟 × [1 − 𝑇𝐶 ] (43)
The Tax Payable is calculated as,
The cost of equity is estimated as,
Tax Payable[$] = AGTR − ADA (53)
𝑘𝑒 [%] = 𝑟𝑓 + [𝛽 × [𝑟𝑚 − 𝑟𝑓 ]] (44)
The Annual Cash Flow [ACF] is calculated as,
The loan annual amortization cost [AAC] is CF[$] = NAR − AGTR + ATBD (54)
computed as,
The Net Present Value (NPV) is calculated as,
[𝐷×𝐶𝐴𝑃𝐸𝑋]×𝑟
AAC[$] = 1 𝑛
(45)
NPV[$] = {∑𝑛=𝑛 −𝑛
1−[ ] 𝑛=1 [𝐶𝐹 ]𝑛 [1 + 𝑖 ] } − 𝐶𝐴𝑃𝐸𝑋
1+𝑟
(55)
where, n=Loan period during project life [Yrs] Where,
The Annual Revenue on Equity Capital n = Operational Life [years]
[AREC] adjusted to effective corporate tax i = Internal Rate of Return (IRR) at WACC
rate [TC] is, In MS-Excel, the NPV can be calculated as,
𝐸×𝑘𝑒 = (𝑁𝑃𝑉(𝑊𝐴𝐶𝐶, 𝐶𝐹1 : 𝐶𝐹𝑛 )) − 𝐶𝐴𝑃𝐸𝑋 (56)
AREC[$] = (46)
1−𝑇𝑐

Page 130
Similarly in MS-Excel, the Internal Rate of Booster Compressor Station
Return (IRR) can be calculated as,
Station Material Cost 69% $9,705,175
= 𝐼𝑅𝑅(𝐶𝐹1 : 𝐶𝐹𝑛 ) (57)
Station Labour Cost 22% $3,089,159
To calculate the “MinTariff” at NPV=0 and Station Misc Cost 9% $1,238,196
IRR=WACC, an MS-Excel Macro is written as,
Station Land Cost 0.3% $43,860
Sub salestariff()
Sub-Total - $14,076,390
Range("NPV").GoalSeek Goal:=0.000001, Total [Incl. Installed
ChangingCell:=Range("MinTariff") 1.00 $14,076,390
Cost Comparison]

End Sub Other Machinery Costs

The Profit to Investment Ratio (PIR) is then Mainline Valve Stations 0.5% $77,234
calculated based on the discounted cash flow Meter Stations 0.5% $77,234
[DCF] as follows, Pressure Regulator
0.1% $15,447
𝐶𝑎𝑠ℎ 𝑂𝑢𝑡𝑓𝑙𝑜𝑤 [𝐶𝐹]𝑛 [1+𝑖]−𝑛 Stations
PIR[−] = = (58)
𝐶𝑎𝑠ℎ 𝐼𝑛𝑓𝑙𝑜𝑤 𝑁𝑃𝑉 SCADA & Telecomm 5.0% $1,476,158
In the above PIR expression, after the Total - $1,646,072
discount factor [1+i]-n is applied to each year,
Project Costs
the negative cash flow, i.e., cash outflow of
Environmental Costs &
each year is summed up and divided by the Permits
15.0% $4,428,473
NPV value. At NPV=0 and WACC=IRR, PIR=0.
Right of Way (ROW)
10.0% $2,952,315
Calculations & Results Acquisition
Based on the estimation methods, the Engineering &
Construction 15.0% $4,428,473
following results are arrived at, Management
Table 5. Pipeline & Booster Station Costs Contingency 10.0% $2,952,315
Description % Cost Total - $14,761,576
Upstream Pipeline
Working Capital [WC] - $5,000,000
Total Material Costs 15% $894,614
AFUDC 20.0% $5,904,630
Coating & Wrapping 5% $306,000
TOTAL CAPEX - $56,835,429
Labour Costs 17% $1,020,000
The Total O&M [OPEX] Costs estimated are as
Install. Cost per inch-km 64% $3,892,719
follows,
Sub-Total - $6,113,333 Table 6. Total O&M Costs
Downstream Pipeline Description % Cost
Total Material Costs 16% $1,505,351 Salaries - $1,000,000
Coating & Wrapping 5% $459,000 Payroll Overhead 20.0% $200,000
Labour Costs 16% $1,530,000 Administrative Expenses 50.0% $500,000
Install. Cost per inch-km 63% $5,839,078 Vehicle Expenses 8.0% $80,000
Sub-Total - $9,333,429 Office Expenses 10.0% $100,000
Total Pipeline Costs - $15,446,761 Misc Materials & Tools 10.0% $100,000

Page 131
 0.4×56,835,429×0.233
Compressor Station Maintenance AREC = = $7,567,231 (67)
1−0.3
Consumable Materials 6.0% $60,000
Gas Tariff, NPV, IRR & PIR Estimates
Periodic Maintenance 18.0% $180,000
Taking the annual availability of 8,040 hours,
ROW Payments 40.0% $400,000 the yearly gas production is as follows for
Utilities 18.0% $180,000 Year 7. It is to be noted calculations are
shown for a set of sample data.
Gas Control 12.0% $120,000
1000×122.3
Gas Rate = × 8,040 = 40,981,711𝑀𝑆𝑐𝑓 (68)
SCADA Maintenance 25.0% $250,000 24

Internal Corrosion The Annual O&M Costs considering inflation is,

30.0% $300,000
Inspection
Annual O&M = $3,590,000 × 1.027−1 = $4,042,923
Cathodic Protection
12.0% $120,000 (69)
Survey
The Annual Cost of Service [ACOS] is,
TOTAL O&M COSTS - $3,590,000
ACOS[$] = 4,042,923 + 5,913,957 + 7,567,231 +
Salvage Costs 1,195,121 = $18,719,233 (70)
The Salvage Costs estimated are as follows, The Gas Transportation Tariff is estimated as,
𝑆𝑉[$] = [15,446,761 + 14,076,390 + 1,646,072 + 18,719,233
Tariff𝑇𝑟𝑎𝑛𝑠 = 40,981,711 = $0.4568/𝑀𝑠𝑐𝑓 (71)
2,952,315] × [1 − 0.08]25 = $4,243,501 (59)
The pipeline salvage deduction is, With a guess value of $0.998/Mscf (MS Excel
𝑆𝐷[$] = [15,446,761 + 14,076,390 + 1,646,072 + Rounded-off value) as minimum gas tariff, the
2,952,315] − 4,243,501 = $29,878,037 (60) Annual Revenue without tax rate included is,
The Annual depreciation of the asset is, AR = 0.998 × 40,981,711 ≈ $40,894,302 (72)

𝐴𝐷𝐴[$] = $29,878,037⁄25 = $1,195,121 (61) The Net Annual Revenue [NAR] becomes,
The Annual Tax benefit from depreciation is, NAR = 40,894,302 − 18,719,233 = $22,175,069 (73)

ATBD[$] = 1,195,121 × 0.35 = $418,293 (62) The Annual Gross Tax on Revenue [AGTR] is,
Cost of Capital AGTR = 22,175,069 × 0.35 = $7,761,274 (74)
The cost of debt is first estimated as, The Tax Payable is calculated as,
𝑘𝑑 [%] = 0.17 × [1 − 0.3] = 11.90% (63) Tax Payable = 7,761,274 − 1,195,121 = $6,566,153
The cost of equity is estimated as, (75)

𝑘𝑒 [%] = 0.17 + [0.9 × [0.24 − 0.17]] = 23.3% (64) The Annual Cash Flow [ACF] for Year 7 is
calculated as,
The weighted average capital cost [WACC] is,
CF[$] = 22,175,069 − 7,761,274 + 418,293 =
WACC[%] = 0.4 × 23.3 + 0.6 × 11.9 = 16.46% (65)
$14,832,088 (76)
Loan Amortization Costs Upon calculating CF for all years, using MS-
The loan annual amortization cost [AAC] is Excel for NPV=0, IRR is calculated as 16.46%.
computed as, The Discounted Cash Flow [DCF] at
[0.6×56,835,429]×0.17 IRR=WACC is,
AAC = 1 25 = $5,913,957 (66)
1−[ ] DCF = 14,832,087 × [1 + 0.1646]−7 = $5,104,636
1+0.17

The Annual Revenue on Equity Capital is, (77)

Page 132
With the calculations performed for NPV=0 at Note: Estimates shown were computed on an
IRR=WACC, the minimum sales gas tariff MS-Excel sheet which takes decimal places
comes to $0.998/Mscf or, ~ $1/Mscf [Refer to until the 10th decimal, while hand calculations
Appendix A] and PIR is 0. were shown by rounding off decimal places.
For end of operational life [Year 25], the References & Further Reading
salvage value of the asset is treated as part of 1. “Gas Pipeline Hydraulics”, E. Sashi Menon,
the Annual Revenue [AR]; therefore the cost 2005, Taylor & Francis
estimation for Year 25 alone changes to,
2. “Pipeline Compressor Station Construction
AR 𝑌𝑒𝑎𝑟 25 [$] = [𝑀𝑖𝑛𝑇𝑎𝑟𝑖𝑓𝑓 × 𝐺𝑎𝑠 𝑅𝑎𝑡𝑒[𝑀𝑠𝑐𝑓]] + 𝑆𝑉 Cost Analysis”, Y. Zhao, Z. Rui, Int J. Oil, Gas
(78) and Technology, Vol. 8No. 1, 2014
The calculations are similarly repeated for (https://www.researchgate.net/publicatio
Sales Tariff rate of $2/Mscf to check for NPV, n/275590240)
IRR, PIR and estimate payback period. The
3. “Economic Analysis of Gas Pipeline Project
results are as follows,
in Nigeria”, Ahmed Adamu, Journal of
Table 7. Economic Analysis for $2/Mscf
Economics and Sustainable Development,
Description Cost Vol.8, No.2, 2017
Total Annual Cash Flow $664,457,040

Net Present Value [NPV] $127,053,012

IRR 45.80%

PIR 2.24

Payback Period
With detailed NPV calculations [Refer to
Appendix A], the payback period for $2/Mscf
can be computed as,
Payback Period =
𝑌𝑒𝑎𝑟 [𝐶𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑣𝑒 𝐴𝑛𝑛𝑢𝑎𝑙 𝐶𝐹 = 𝑇𝑜𝑡𝑎𝑙 𝐶𝐴𝑃𝐸𝑋]
(79)
The payback period can be estimated
graphically as follows,

Figure 2. Payback Period – Gas Tariff at $2/Mscf

Payback Period = 4.212 𝑌𝑟𝑠 (80)

Page 133
Appendix A

Page 134
Appendix B

Page 135
 Module 20
Evaluating Pipeline Operational Integrity - Sand Production
Piping systems associated with production, to follow a procedure that meets pipeline
transporting oil & gas, water/gas injection integrity requirements while also lowering
into reservoirs, experience wear & tear with maintenance, repair & replacement costs.
time & operations. There would be metal loss 2. To assess pipeline integrity, standard
due to erosion, erosion-corrosion and corrosion assessment procedures are
cavitation to name a few. The presence of classified on three levels – Level 1, Level 2
corrosion defects provides a means for and Level 3. Level 1 procedure represents
localized fractures to propagate causing pipe longitudinal area of metal loss based on the
ruptures & leakages. This also reduces the maximum defect depth and overall defect
pipe/pipeline maximum allowable operating length. The ASME B31G, RSTRENG 085dL,
pressure [MAOP]. and DNVGL RP F-101 method for single
The following document covers methods by defect can be classified as Level 1 methods.
DNV standards to quantitatively estimate the Level 2 procedure represents longitudinal
erosion rate for ductile pipes and bends due area of metal loss based on the defect
to the presence of sand. It is to be noted that depth profile. The RSTRENG Effective Area
corrosion can occur in many other scenarios method and DNVGL RP F-101 method for
such as pipe dimensioning, flow rate complex shaped defects can be classified as
limitations, pipe performance such as Level 2 methods. Level 3 assessment
pressure drop, vibrations, noise, insulation, methods involve using Finite element
hydrate formation and removal, severe slug methods (FEM) provided the FEM model is
flow, terrain slugging and also upheaval validated against experimental results.
buckling. However these aspects are not 3. Corrosion failures are caused by two main
covered in this document. mechanisms – Leakage resulting in a
Based on the erosional rates of pipes and relatively small loss of product and
bends, the Maximum Safe Pressure/Revised Rupture causing a sudden release of
MAOP is evaluated based on a Level 1 pressure which propagates in isolation.
Assessment procedure for the remaining 4. To understand corrosion assessment
strength of the pipeline. The Level 1 procedures, two terms come into play –
procedures taken up in this module are Folias Bulging factor [MT] and flow stress
RSTRENG 085dL method, DNVGL RP F-101 [ f]. Folais factor represents the bulging
(Part-B) and PETROBRAS’s PB Equation. effect of a shell surface that is thinner in
General Notes & Assumptions wall thickness [WT] than the surrounding
shell. It takes into account the work-
1. In evaluating corrosion defects, the
hardening effect, i.e., the increase in the
generally accepted or traditional approach
stress concentration levels as the corrosion
is the ASME B31G code which gives overly
defect begins to bulge before eventually
conservative results in terms of lower
causing a failure. The flow stress is the
burst pressures with which operators
stress at which the corrosion defect is
repair/replace the corroded pipe/pipeline
predicted to cause a failure.
segments. This represents higher
maintenance costs necessitating the need
Page 136
5. In pipeline assessment literature, SMYS 8. ASME B31G assumes a parabolic profile
and yield strength are used differently. across the area of the defect, i.e., Area of
Specific Minimum Yield Strength (SMYS) is defect = 2/3 d L, where, d = Defect depth
the absolute minimum yield strength for a and L = Defect longitudinal length.
particular material grade specified by Whereas with the RSTRENG 085dL
ASTM standards. Whereas, yield strength is method, the defect area is approximated as
obtained from mill conducted tensile tests. 85% of the peak depth, i.e., by using a
Wherever possible yield strength should factor of 0.85, i.e., Defect Area = 0.85 d L.
be used. In cases, where the yield strength
value is not available, SMYS can be used
instead.
6. When a corrosion defect occurs inside a
pipe/pipeline, the defect tends to
propagate longitudinally. ASME B31G
mandates a maximum allowable Figure 1. Corrosion Shape Approximation
longitudinal length [LM] for a given defect 9. The potential for sand particles to get
depth [d]. As per Modified ASME B31G carried from the formation to well bore in
method i.e., 085dL method, defects are oil & gas wells is subjected to the reservoir
classified as Long defect and short defect geology. With the onset of water formation
based on the condition, LS2/Dt = 50, Where or rapid change in well conditions, there is
D = Pipeline Outer diameter (OD) and t = sand formation. Employing a zero rate of
pipeline nominal wall thickness. When sand production would be economically
field measured defect’s longitudinal length, infeasible. Therefore sand management
L < LS, the defect is termed as short defect. programmes are put in place whereby
When L > LS, the defect is termed as long upstream facilities are equipped with sand
defect. The DNVGL RP F-101 method does traps with necessary safeguards that aid in
not classify defects in relation to their achieving an acceptable sand rate. The
longitudinal length. The pressure strength standard used for this module is DNVGL RP
of long defects is a function of the O501 which provides empirical models that
longitudinal defect length [L]. The Longer cover plain erosion & not the combined
the defect, lower is the failure pressure effects of corrosion-erosion, droplet erosion
However a limit exists in the value of L, & cavitation. The module therefore
beyond which any large increase in the considers plain erosion which leads to
longitudinal defect length, L produces very corrosion pits in the pipeline & the
little reduction in the failure pressure. associated MAOP is computed using the
7. Long Internal defects are one of the standard corrosion assessment methods.
various causes for geometry corrosion 10. When applying the original ASME B31G
induced damage that occur in oil & gas method in simplified form (Appendix L of
pipelines. These occur on the pipe/pipeline ASME B31.8), the Safe Operating Pressure
bottom due to accumulation of liquids given as P’ must first be calculated using the
including water. Whereas long external pressure corresponding to a hoop stress
defects are caused on the pipeline’s outer equal to 100% of SMYS for the operating
surface due to loss of protective coatings. pressure, P. The resulting P’ is the estimated

Page 137
 failure pressure, which must then be Pipeline Wall Thickness [WT] Estimation
divided by the design factor/desired factor Based on the ASME B31.8 code the wall-
of safety to obtain the correct “Safe thickness formula is stated as,
Operating Pressure”. 𝐷𝑃×𝑂𝐷
Case Study: Problem Statement
𝑡= (1)
2×𝐹×𝐸×𝑇×𝑆𝑀𝑌𝑆
30 MMscfd of well fluids at 40 bara and 400C Where,
is transported through an 8” DN carbon steel t = Minimum design wall thickness [in]
flowline from the well head to a trunk line.
DP = Pipeline Design Pressure [psi]
The process & mechanical details are,
OD = Pipeline Outer Diameter [in]
Table 1. Process & Mechanical Details
SMYS = Specific Minimum Yield Stress [psi]
Parameter Value Unit
F = Design Factor [-]
Operational Life of Pipeline 25 Years
E = Longitudinal Weld Joint Factor [E]
Location of Gas Pipeline Deserted -
T = Temperature De-rating Factor [-]
Class 1,
Location Class -
Div 2 Applying the ASME B31.8 correlation, the
Design Factor [F] 0.72 - calculated wall thickness becomes,
Pipeline Joining Method ERW - [44×14.5]×8.625×25.4
𝑡= = 1.49 𝑚𝑚 (2)
2×0.72×1×1×[448×145.038]
Longitudinal Joint Factor [E] 1.0 -

API 5L Spec
PSL1
-
The total WT including CA of 1.0 mm is,
X65
𝑡𝑡𝑜𝑡𝑎𝑙 = 1.49 𝑚𝑚 + 1.0 𝑚𝑚 = 2.49 𝑚𝑚 (3)
Ultimate Tensile Strength [ u] 530 MPa

SMYS [S] 448 MPa

The Selected WT based on API5L is,
𝑡𝑠𝑒𝑙𝑒𝑐𝑡𝑒𝑑 = 3.175 𝑚𝑚 [0.125 𝑖𝑛] (4)
Design Pressure [DP] 44 bara

Design Temperature [DT] 100 0C The revised design pressure based on the
selected Wall Thickness [WT] is,
De-Rating Factor [T] 1.00 -
2×𝐹×𝐸×𝑇×𝑆𝑀𝑌𝑆×𝑡
Pipeline Diameter [DN] 8.625 in 𝐷𝑃 = (5)
𝑂𝐷

Corrosion Allowance [CA] 1.0 mm 𝐷𝑃 =

2×0.72×1×1×[448×145.038]×3.175
= 93.5 𝑏𝑎𝑟𝑎 (6)
8.625×14.5×25.4
Gas Flow Rate [mg] 31,657 kg/h
The pipeline inside diameter [ID] becomes,
Liquid Flow Rate [ml] 14,928 kg/h
𝐼𝐷 = 𝑂𝐷 − [2 × 𝑊𝑇] (7)
Gas Density [ g] 42.0 kg/m3
𝐼𝐷 = [8.625 × 25.4] − [2 × 3.175] ≈ 212.73 𝑚𝑚 (8)
Liquid Density [ l] 713.2 kg/m3
The pipeline cross section area [At] becomes,
Gas Viscosity [µg] 1.34E-05 kg/m.s
𝜋𝐷 2 𝜋 212.73 2
Liquid Viscosity [µl] 4.72E-04 kg/m.s 𝐴𝑝 = = ×[ ] = 0.0355 𝑚2 (9)
4 4 1000

Mixture Viscosity [µm] 2.58E-05 kg/m.s Therefore the gas velocity is estimated as,
Sand Content [ppmW] 50.0 ppmW 𝑄𝑔 31,657
𝑉𝑔 = = = 5.887 𝑚/𝑠 (10)
𝐴𝑝 3600×42×0.0355
Average Sand Particle Size 300 µm
No. of Pipe Diameter [900 The liquid Velocity is estimated as,
1.5 [-]
Long Elbow] 𝑄 14,928
Inclined Pipe Impact angle [ ] 300 degrees
𝑉𝑙 = 𝐴 𝐿 = 3600×713.2×0.0355 = 0.164 𝑚/𝑠 (11)
𝑝

Page 138
Well Fluids Mixture Properties [𝑚𝑔 +𝑚𝑙 ]×𝑝𝑝𝑚𝑊
𝑚𝑝 = (18)
106
The mixture density and mixture viscosity of
the well fluids can be determined as follows. The erosion rate can be calculated beginning
with estimating the function characterizing
𝜌𝑔 𝑉𝑔 +𝜌𝑙 𝑉𝑙
𝜌𝑚 = (12) pipeline ductility, F( ) as follows,
𝑉𝑔 +𝑉𝑙
𝛼𝜋 30𝜋
𝜇𝑔 𝑉𝑔 +𝜇𝑙 𝑉𝑙 For 300, = = 0.5236 (19)
180 180
𝜇𝑚 = (13)
𝑉𝑔 +𝑉𝑙 𝐹(𝛼) = 0.6 × [𝑆𝑖𝑛(0.5236) + 7.2(𝑆𝑖𝑛(0.5236) −
0.6
Therefore applying the above correlations, 𝑆𝑖𝑛2 (0.5236))] × [1 − 𝑒 −20×0.5236 ] = 0.9890 (20)

𝜌𝑚 =
[42×5.887]+[713.2×0.164]
= 60.2
𝑘𝑔
(14) The sand flow rate based on ppmW is,
5.887+0.164 𝑚3
[31,657+14,928]×50
𝜇𝑚 =
[0.0000134×5.887]+[0.000472×0.164]
= 0.0000258
𝑘𝑔
(15) 𝑚𝑝 = = 2.33 𝑘𝑔/ℎ (21)
5.887+0.164 𝑚.𝑠 106

Inclined Pipe Erosion Rate – DNV RP O501 The pipeline area exposed to erosion is,
The flowline profile over the terrain would 𝐴𝑝 0.0355
𝐴𝑡 = 𝑆𝑖𝑛(𝛼) = 𝑆𝑖𝑛(30) = 0.0711 𝑚2 (22)
have inclined sections. With the onset of
water production from the wells, quartz sand The erosion rate is therefore calculated as,
particles from wells [50 ppmW, 300 µm, 𝐸=
2.33×[2×10−9 ]×[5.887+0.16]2.6 ×0.989×[3.15×1010 ]
(23)
3600×7,800×0.0711
2,650 kg/m3] impinge at an impact angle of
𝐸 = 0.0078 𝑚𝑚/𝑦 (𝑜𝑟) 0.31 𝑚𝑖𝑙𝑠/𝑦 (24)
300. As per DNVGL RP O501 [Rev. 2015], for
ductile materials, the maximum erosion Pipe Bend Erosion Rate – DNVGL RP O501
occurs for impact angles in the range of 150 to Pipeline bends are prone to erosional wear.
300, whereas brittle materials experience When the flow direction in the bend changes,
maximum erosion at normal impact angle. sand particles crash against the bend wall,
The erosive wear can be estimated as, instead of following the flow direction.
𝑚𝑝 ×𝐾×𝑈𝑝𝑛 ×𝐹(𝛼) Assuming a straight length [10D] before the
𝐸= × [3.15 × 1010 ] (16)
𝜌𝑡 ×𝐴𝑡 bend, the erosion rate is estimated as,
Where,
mp = Sand Flow rate [kg/s]
K=Material Constant (2 10-9 for Steel Grades)
n =Material Constant (2.6 for Steel Grades)
Up = Particle Velocity [m/s] (Vg + Vl)
t = Pipeline density [kg/m3]
At = Pipeline Area exposed to Erosion [m2]
F( ) = Function characteristic of ductility [-]
The value of F( ) is calculated as,
Figure 2. Impact Angle [ ] in Pipeline Bends
𝐹(𝛼) = 0.6 × [𝑆𝑖𝑛(𝛼) + 7.2(𝑆𝑖𝑛(𝛼) −
0.6 The characteristic impact angle, for the pipe
𝑆𝑖𝑛2 (𝛼))] × [1 − 𝑒 −20𝛼 ] (17)
bend geometry is calculated from the radius
For the condition, F( ) [0, 1] for [0, /2] of curvature. The radius of curvature, Rc [i.e.,
Note: 1 mil = 1/1000th of an inch bend radius] for a bend is expressed as
The sand flow rate based on ppmW is number of pipe diameters. Considering a 900
calculated as,
Page 139
long elbow, the bend radius in terms of Where, dp = Average Particle diameter
number of pipe diameters is 1.5, i.e., Rc = 1.5. 𝜌𝑚
𝑑𝑝,𝑐 𝑖𝑓 0 < 𝛾𝑐 < 1
1 = 𝛾𝑐 = {𝜌𝑝 [1.88𝑙𝑛(𝐴)−6.04] (30)
𝐼𝐷
𝛼 = 𝑇𝑎𝑛−1 [ ] (23) 0.1 𝑖𝑓 𝛾𝑐 ≤ 0 𝑜𝑟 𝛾𝑐 ≥ 0.1
√2𝑅𝑐

1
The particle size correction function [G] is,
𝛼 = 𝑇𝑎𝑛−1 [ ] = 0.5236 𝑟𝑎𝑑 𝑜𝑟 30° (24)
√2×1.5 𝛾
𝑖𝑓 𝛾 < 𝛾𝑐
The length of the 900 bend is estimated as, 𝐺= {𝛾𝑐 (31)
1 𝑖𝑓 𝛾 ≥ 𝛾𝑐
𝐿𝐵𝑒𝑛𝑑 = 360 × 2𝜋𝑅𝑐 , 𝑊ℎ𝑒𝑟𝑒 𝜃 = 90° (25) The pipeline Bend Area exposed to erosion is,
90 𝑡 𝐴
𝐿𝐵𝑒𝑛𝑑 = × 2𝜋 × 1.5 × 8.625 = 20.3𝑖𝑛 (26) 𝐴𝑡 = 𝑆𝑖𝑛(𝛼) (32)
360

As per DNVGL RP O0501, the erosional rate Applying the expressions to the case study,
[E] for pipe bends is computed as, 𝐴=
60.22 ×𝑇𝑎𝑛(30)×6.1×0.2127
≈ 39,315 (33)
2,650×0.0000258
𝑚𝑝 ×𝐾×𝑈𝑝𝑛 ×𝐹(𝛼)×𝐺×𝐶1 ×𝐺𝐹
𝐸= × [3.15 × 1010 ] (27) 300
𝜌𝑡 ×𝐴𝑡 𝛾 = 0.2127×106 = 0.00141027 (34)
Where, 60.2
𝛾𝑔 = 2,650×[1.88𝑙𝑛(39315)−6.04] = 0.00163961 (35)
E = Erosion Rate [mm/year]
Therefore since < g, the particle size
mp = Sand Flow rate [kg/s]
correction function is,
K=Material Constant (2 10-9 for Steel Grades) 𝛾 0.00141027
𝐺 = 𝛾 = 0.00163961 = 0.8601 (36)
n =Material Constant (2.6 for Steel Grades) 𝑔

Up = Particle Velocity [m/s] (Vg + Vl) The critical particle diameter [dp,c] is
calculated as,
t = Pipeline density [kg/m3]
0.2127×0.00163961
𝑑𝑝,𝑐 = 𝐼𝐷 × 𝛾𝑐 = ≈ 349µ𝑚 (37)
At = Bend Area exposed to erosion [m2] 10−6

F( ) = Function characteristic of ductility [-] The pipeline area exposed to erosion is,
𝐴𝑝 0.0355
G = Particle Size correction function [-] 𝐴𝑡 = 𝑆𝑖𝑛(𝛼) = 𝑆𝑖𝑛(30) = 0.0711 𝑚2 (38)
C1 = Constant accounting for multiple impact
The function characterizing pipeline ductility,
of sand particles at the bend’s outer end [2.5]
F( ) as follows,
GF = Geometry Factor 𝛼𝜋 30𝜋
For 300, = = 0.5236 (39)
The geometry factor [GF], is taken to be 1.0 180 180

based on the assumption that the straight line 𝐹(𝛼) = 0.6 × [𝑆𝑖𝑛(0.5236) + 7.2(𝑆𝑖𝑛(0.5236) −
section, upstream of the bend is greater than 𝑆𝑖𝑛2 (0.5236))]
0.6
× [1 − 𝑒 −20×0.5236 ] = 0.9890 (40)
10D. For straight line section less than 10D,
Therefore the erosion rate is computed as,
the GF increases to 2 or 3. To arrive at the
2.33×[2×10−9 ]×[6.1]2.6 ×0.989×1×2.5×0.86×[3.15×1010 ]
particle size correction term, G, the 𝐸= 3600×7,800×0.0711
dimensionless parameter A is calculated first.
(41)
2 ×𝑇𝑎𝑛(𝛼)×𝑈 ×𝐷
𝜌𝑚 𝑝
𝐴= (28) 𝐸 = 0.0496 𝑚𝑚/𝑦 (𝑜𝑟) 1.95 𝑚𝑖𝑙𝑠/𝑦 (42)
𝜌𝑝 ×𝜇𝑚

The diameter relation [ ] and critical Max Safe Pressure in Corroded Area
diameter relation [ g] is calculated as, With sand erosion occurring due to sand flow,
𝑑𝑝 defects begin to form on the pipeline inner
𝛾= (29)
𝐼𝐷
Page 140
surface. These defects have a certain depth of 2𝑡[𝑆𝑀𝑌𝑆+69]×145.04×𝐹×𝐸×𝑇 1−0.85( )
𝑑
𝑡
𝑃𝑓 = [ ] (46)
penetration [d] for a given wall thickness of
𝑑
𝐷×14.5 1−[0.85( )𝑀−1 ]
𝑡

the pipe [t]. The following section provides Where,

calculations for the maximum safe pressure
SMYS = Specific Min Yield Strength [MPa]
for operation based on RSTRENG 085dL
method, DNVGL RP F101 Single defect D = Pipeline OD [in]
method and PETROBRAS PB method. M = Folias Bulging Factor [-]
As per ASME B31G, for a pit depth of up to For the condition, L2/Dt 50, M is,
10%, the pipeline can be continued to be
𝐿2 𝐿2 2
operated with the existing MAOP. For a pit 𝑀 = √1 + 0.6275 [𝐷𝑡] − 0.003375 [𝐷𝑡] (47)
depth between 10% and 80%, the pipeline
For the condition, L2/Dt > 50, M is,
needs to be operated at the revised/reduced
MAOP based on the corroded wall thickness. 𝐿2
𝑀 = 3.3 + 0.032 [𝐷𝑡] (48)
For a pit depth greater than 80%, the pipeline
For this module, the measured max corroded
would have to repaired or replaced.
area depth [d] and measured longitudinal
As per ASME B31G, for a contiguous corroded
length [L] in the inclined pipe is 0.04” and 3”
area having a maximum depth of more than
respectively. For the selected wall thickness
10% but less than 80% of the nominal pipe
of 3.18 mm, d/t is 0.32, i.e., 32% pit depth.
wall thickness, Lm should not extend along the
longitudinal axis of the pipe for a distance Similarly for the pipe bend, the measured
greater than calculated from the expression, max corroded area depth [d] and measured
longitudinal length [L] is 0.06” and 1.3”
𝐿𝑚 = 1.12𝐵√𝐷𝑡 (43) respectively. For the selected wall thickness
Where, of 3.18 mm, d/t is 0.48, i.e., 48% pit depth.
Lm = Maximum Allowable Longitudinal length Therefore, for the inclined pipeline,
of corroded area [in] 𝐿2 32
= 8.625×0.125 = 8.35 < 50 (49)
𝐷𝑡
D = Pipeline OD [in]
Since L2/Dt < 50, the Folias bulging factor is,
T = Pipeline selected Wall thickness [in]
𝑀 = √1 + [0.6275 × 8.35] − 0.003375[8.35]2 (50)
The constant B is estimated as,
𝑀 = 2.45 (51)
2
𝑑
( ) Therefore the max safe pressure is,
𝐵 = √[ 𝑡
𝑑 ] −1 (44) 2×0.125[448+69]×145.04×0.72×1×1 1−(0.85×0.32)
1.1( )−0.15
𝑡 𝑃𝑓 = [ ] (52)
8.625×14.5 1−[0.85×0.32×2.45−1 ]

As per ASME B31G, B cannot be > 4.0. For 𝑃𝑓 = 88.4 𝑏𝑎𝑟𝑎 (53)
corrosion depth [d/t] between 10% and
Performing similar calculations for Pipe Bend
17.5%, the value of B is to be limited to 4.0.
with d = 0.06” and L=1.3”, the Max safe
For e.g., with d/t = 0.32, the value of B & Lm is,
pressure is 90.0 bara.
2
0.32
𝐵 = √[1.1(0.32)−0.15] − 1 = 1.23 (44) DNV RP F101 Single Defect Method
The max safe pressure with DNVGL RP F101
𝐿𝑚 = 1.12 × 1.23√8.625 × 0.125 = 1.43 𝑖𝑛 (45) single defect method is determined as,
RSTRENG 085dL Method 𝑑
2×𝑡×𝜎𝑢 ×145.04×𝐹×𝐸×𝑇 1−( )
𝑡
The max safe pressure with RSTRENG 085dL 𝑃𝑓 = [ 𝑑 ] (54)
[𝐷−𝑡]×14.5 1−[( )𝑀−1 ]
𝑡
method is determined as follows,
Page 141
Where, Results
u = Ultimate Tensile Strength [MPa] Summarizing, the max safe pressure is 88.4
bara for pipeline and 90 bara for pipe bend,
D = Pipeline OD [in]
Table 2. Max Safe Pressures
M = Folias Bulging Factor [-]
Max Safe Pressure,
Method
𝐿2 Pf [bara]
𝑀 = √1 + 0.31 [𝐷𝑡] (55)
Inclined Pipeline
-
Pipe Bend
Applying the DNVGL RP F101 Single defect
method to the same inclined pipe and pipe RSTRENG 085dL 88.4 90.0
bend data for an ultimate tensile strength of DNV RP F101 Single Defect 91.9 96.3
530 MPa, the max safe pressure is, Petrobras PB 94.3 99.7
𝑀 = √1 + [0.31 × 8.35] = 1.89 (56) Design Pressure [DP] 93.5 93.5
2×0.125×530×145.04×0.72×1×1 1−0.32
𝑃𝑓 = [1−[0.32×1.89−1 ]] (57) Max Safe Pressure, Pf 88.4 90.0
[8.625−0.125]×14.5

𝑃𝑓 = 91.9 𝑏𝑎𝑟𝑎 (58) Based on the erosion rate for an operating

period of 25 years, the pipeline WT lost is,
Performing similar calculations for pipe bend
with d=0.06” and L=1.3”, the Max safe Table 3. Pipeline WT Lost

pressure is 96.3 bara. Inclined Pipeline

Parameter
Pipe Bend
PETROBRAS PB Method
Erosion Rate [mm/y] 0.0078 0.0496
The max safe pressure with PETROBRAS PB
WT Lost in 25 Years [mm] 0.20 1.24
method is determined as,
1−( )
𝑑 References & Further Reading
2×𝑡×𝜎𝑢 ×145.04×𝐹×𝐸×𝑇 𝑡
𝑃𝑓 = [ ] (59)
[𝐷−𝑡]×14.5 𝑑
1−[( )𝑀
𝑡
−1] 1. “Managing Sand Production and Erosion”,
DNVGL-RP-O501, Aug 2015 Edition.
Where,
2. “Manual for Determining Remaining
u = Ultimate Tensile Strength [MPa]
Strength of Corroded Pipelines”, ASME
M = Folias Bulging Factor [-] B31G-1991
𝐿2 1 𝐿2 4
3. “Folias Factor”, Science Direct,
𝑀 = √1 + 0.217 [𝐷𝑡] − 1.15×106 [𝐷𝑡] (60) https://www.sciencedirect.com/topics/en
Applying the PETROBRAS PB method to the gineering/folias-factor
same inclined pipe and pipe bend data for an 4. “Modified Equation for the Assessment of
ultimate tensile strength of 530 MPa, the max Long Corrosion Defects”, Adilson C.
safe pressure is, Benjamin, Ronaldo D Vieria, Jose Luiz F.
Friere, Jaime T.P. de Castro,
0.8354
𝑀 = √1 + [0.217 × 0.835] − 1.15×106 = 1.68 (61) https://www.researchgate.net/publication
/249657141
2×0.125×530×145.04×0.72×1×1 1−0.32
𝑃𝑓 = [8.625−0.125]×14.5
[1−[0.32×1.68−1 ]] (62)

𝑃𝑓 = 94.3 𝑏𝑎𝑟𝑎 (63)

Performing similar calculations for Pipe Bend

with d = 0.06” and L=1.3”, the Max safe
pressure is 99.7 bara.

Page 142
Appendix A

Page 143
 Module 21
ECONOMIC INUSLATION FOR INDUSTRIAL PIPING
Thermal Insulation for Industrial Piping is a Lower heat transfer coefficients & thermal
common method to reduce energy costs in conductivity offer a lower rate of heat
production facilities while meeting process loss/gain. It is for this reason; materials that
requirements. Insulation represents a capital provide low thermal conductivity are chosen
expenditure & follows the law of diminishing to provide insulation. To provide effective
returns. Hence the thermal effectiveness of insulation, the conductive heat transfer from
insulation needs to be justified by an the metal has to be kept lower than the
economic limit, beyond which insulation convective heat transfer on the insulation’s
ceases to effectuate energy recovery. To external side to prevent the outer insulation
determine the effectiveness of an applied temperature from increasing drastically.
insulation, the insulation cost is compared
with the associated energy losses & by
choosing the thickness that gives the lowest
total cost, termed as ‘Economic Thickness’.

Figure 2. Critical Insulation Thickness

From the above, when insulation is applied on
a bare pipe of a given nominal diameter, the
heat transfer rate increases as the insulation
Figure 1. Economic Insulation Thickness Selection radius/thickness increases. As the insulation
thickness increases, until reaching the critical
The following module provides guidance to
radius [Rc], there is a progressive fall in the
estimate the economic thickness for natural
convective resistances causing higher heat
gas piping in winter conditions as an example
losses from the pipe.
case study.
Therefore for insulation to be properly
Design Considerations
effective in restricting heat transmission, the
To estimate the economic insulation
outer pipe radius R2 must be greater than or
thickness, the following factors are to be
equal to the critical radius [Rc] of the
given attention – Energy costs
insulation. If this condition is not satisfied, no
(steam/electricity), annual hours of
useful purpose will be served with the chosen
operation, operating surface temperature,
material of insulation.
pipe dimensions, estimated cost of insulation,
Case Study & Assumptions
and average exposure to the ambient. These
are critical to predict the thermal resistances To demonstrate the economic insulation
and heat transfer coefficients and the total calculations, a case study is made based on a
heat loss or gain, from or to the system. natural gas piping operating in cold winter

Page 144
conditions. In addition, certain assumptions 𝐶𝑝,𝑎𝑖𝑟 = 1.006 (2)
are made for this example case study.
3. The thermal conductivity [W/m.K] of air is,
1. Ambient temperature [Ta] is taken as 00C &
𝑘𝐴𝑖𝑟 = −2.69 × 10−8 𝑇 2 + 9.04 × 10−5 𝑇 + 9.56 × 10−4 (3)
wind velocity is taken as 18 km/h (5 m/s).
4. The thermal diffusivity [m2/s] of air is,
2. The pipe inside heat transfer [HT]
coefficient [hi] is neglected since it is small 𝛼𝐴𝑖𝑟 = 1.99 × 10−10 𝑇 2 + 1.5 × 10−8 𝑇 − 7.96 × 10−7 (4)

compared to outer/ambient HT coefficient. 5. The dynamic viscosity [kg/m.s] of air is,

3. Radiation is accounted for with the 𝜇𝐴𝑖𝑟 = −4.22 × 10−11 𝑇 2 + 7.19 × 10−8 𝑇 + 8 × 10−7 (5)
emissivity of the outer bare pipe taken as
6. The kinematic viscosity [m2/s] of air is,
0.9 while insulation emissivity is 0.13.
= 1.02 × 10−10 𝑇 2 + 3.1 × 10−8 𝑇 − 2.69 × 10−6 (6)
4. Heat transfer through pipe & insulation 𝐴𝑖𝑟

material is assumed to be perfectly radial 7. The Prandtl Number of air is computed as,
& critical thickness is estimated at steady 𝑃𝑟 = −5.12 × 10−7 𝑇 2 + 3.7 × 10−5 𝑇 + 0.7642 (7)
state conditions, i.e., at equilibrium. Based on the above correlations, for an
Design Data ambient temperature of 273.15 K, the air
properties are as follows,
For estimation of heat transfer coefficients,
the process data used is as follows, Table 2. Air Properties at Ambient Conditions

Table 1. Natural Gas Composition Parameter Value Unit

Component MW Mol% Density [ air] 1.293 kg/m3

- [kg/kmol] [%] Specific Heat [Cp,air] 1.006 kJ/kg.K

Methane [CH4] 16.04 76.23 Thermal Conductivity [kair] 0.0236 W/m.K

Ethane [C2H6] 30.07 10.00 Thermal Diffusivity [ air] 0.000018 m2/s

Propane [C3H8] 44.01 5.00 Dynamic Viscosity [ air] 0.000017 kg/m.s

i-Butane [i-C4H10] 58.12 1.00 Thermal Exp. Coefficient [ air] 0.0037 1/K

n-Butane [n-C4H10] 58.12 1.00 Kinematic Viscosity [ air] 0.000013 m2/s

i-Pentane [i-C5H12] 72.15 0.30 Natural Gas Pipe Construction Details

n-Pentane [n-C5H12] 72.15 0.10 The construction details of the natural gas
Water [H2O] 18.02 0.25 pipe is as follows,
Table 3. Pipe Construction Details
Carbon dioxide [CO2] 44.01 3.00
Parameter Value Units
Hydrogen Sulphide [H2S] 34.08 0.07
Nitrogen [N2] 28.01 3.00 Pipe Material Carbon Steel

Total 100.0 Design Pressure 11.0 bara

The air properties between -250C & 500C are Design Temperature 100 0C

computed using fitted equations as follows, Pipeline DN 6.625 in

1. Air Density [kg/m3] is computed as, Pipe WT 3.58 mm

𝜌𝐴𝑖𝑟 = 0.0000158𝑇 2 − 0.0134𝑇 + 3.7622 (1) Pipe ID 161.1 mm

2. Air Specific Heat [kJ/kg.K] is computed as, Pipe Length incl. Fittings [Le] 1,000 m

Pipe Total OD [D3] 269.875 mm

Page 145
Pipe Thermal Cond. [kpipe] 45 W/m.K

Pipe Surface Emissivity[ ] 0.90 -

Ambient Temperature [Tamb] 0 0C

Wind Velocity [Va] 18 km/h

Insulation Material Urethane Foam

Insulation Thermal Cond. [kins] 0.018 W/m.K

Insula. Surface Emissivity [ ] 0.13 -

Figure 4. Hydrate P-T Curve

Results
With the methodology employed, the pipe
process results computed with Weymouth &
DAK-EoS are as follows,
Table 5. Pipe Process Results
Figure 3. Pipe Construction
Parameter Value Units
The Process data used for the case study is,
Pipe Inlet Velocity [V] 10.0 [m/s]
Table 4. Pipe Inlet Process Data

Parameter Value Units

Pipe Exit Velocity [Ve] 10.5 [m/s]

Pipe Exit Temperature [Te] 39.2 [0C]

Pipe Gas Flow Rate [Q] 12.0 MMSCFD
Pipe Exit Pressure [Pe] 19.07 [bara]
Pipe Inlet Pressure [P1] 20.0 bara
Pressure Drop [ΔP] 0.93 [bar]
Pipe Temperature [T1] 40.0 0C

ΔP per km [ΔP/L] 0.93 [bar/km]

Gas MW 21.16 kg/kmol

Pipe Inlet Cp 2.0967 kJ/kg.K The dQ vs. Insulation radius plot shows a
decreasing trend between heat loss from a
Compressibility Factor [Z1] 0.9539 -
bare pipe [Qbare] and heat loss from an
Gas Flow [Act_m3/h] 742 m3/h insulated pipe [QIns] with increase in
Gas Density [r] 17.04 kg/m3 insulation thickness.
Mass Flow [m] 12,643 kg/h

The gas compressibility factor, Z is predicted

using DAK EoS. Gas line pressure drop is
estimated using Weymouth equation. Due to
the presence of water in the natural gas
stream, ice & hydrate formation tendencies
exist. For a flow pressure of 16.14 bara, the
hydrate temperature is 9.520C.
Therefore insulation is to be provided to
ensure hydrate & ice formation does not take Figure 5. dQ vs Insulation Thickness
place. A Hydrate P-T plot is therefore A plot between the total annual costs &
presented as follows, insulation thickness shows that the annual
Page 146
total cost of the energy losses is the least at 2” temperature difference rather than to the
insulation thickness representing 53,694€. difference between two temperatures to the
fourth power.
Forced Convection
To calculate the external heat transfer
coefficient [ho], Nusselt number for forced
convection over circular cylinder with cross
flow can be estimated using Churchill and
Bernstein correlation [1]. This equation is
valid for all Re.Pr 2 and the correlation is
expressed as,
4⁄
1 1 5⁄ 5
0.62 𝑅𝑒 ⁄2 𝑃𝑟 ⁄3 𝑅𝑒 8
𝑁𝑢 = 0.3 +
2
1⁄ [1 + (282000) ] (10)
4
0.4 ⁄3
[1+( 𝑃𝑟 ) ]
Figure 6. Costs per Year vs. Insulation Thickness
Appendix A: Design Methodology Prandtl Number [Pr] of ambient air is,
To estimate the thermal insulation required, 𝐶𝑃,𝑎𝑖𝑟 𝜇𝑎𝑖𝑟
𝑃𝑟 = (11)
the heat losses & heat transfer coefficients are 𝑘𝑎𝑖𝑟

accounted based on 3 modes of heat transfer Reynolds number [Re] becomes,

driven by temperature differences – namely, 𝑉𝑎𝑖𝑟 𝐷3 𝜌𝑎𝑖𝑟
𝑅𝑒 = (12)
pipe wall conduction, free convection, forced 𝜇𝑎𝑖𝑟

convection & ambient radiation. For the bare The above correlation is valid for all ranges of
pipe & insulation cases, air flows over the Reynolds number (Re) and Pr 0.2, where all
pipe surface thereby forming a film with a properties are evaluated at film temperature.
certain temperature. This film temperature It is to be noted that as per [1], Churchill &
determines the rate of heat losses through the Bernstein correlation is reasonable over a
pipe surface/insulation. The air film certain range of conditions but for most
temperature [Tairfilm] on the insulation surface engineering calculations, the accuracy is not
is estimated iteratively. Therefore for the first expected to be much better than 20% because
iteration, these are based on more recent results
encompassing a wide range of conditions.
𝑇𝑎𝑖𝑟𝑓𝑖𝑙𝑚,𝑖𝑛𝑠,1 = 𝑇𝑎𝑚𝑏 [°𝐶] + 1°𝐶 (8)
Natural/Free Convection
Radiation Heat Transfer To estimate the heat transfer due to natural
To estimate the radiation heat transfer convection, the correlation by Churchill & Chu
between the ambient & concrete insulation [1] can be used and is of the form,
on the tank, the expression is written as, [1], 2

2 1
ℎ𝑟 = 𝜀 × 𝜎(𝑇𝑎𝑖𝑟𝑓𝑖𝑙𝑚, + 𝑇𝑎 )(𝑇𝑎𝑖𝑟𝑓𝑖𝑙𝑚 + 𝑇𝑎2 ) (9) 0.387 𝑅𝑎 ⁄6
𝑁𝑢 = {0.6 + 8⁄ } (13)
9 27
0.559 ⁄16
Where, = 5.67 10-8 W/m2/K [1+(
𝑃𝑟
) ]

= Surface emissivity
Where, Rayleigh number (Ra) is computed as,
The radiation mode expressed above is 𝑔×𝛽𝑎𝑖𝑟 ×[𝑇𝑎𝑖𝑟𝑓𝑖𝑙𝑚 −𝑇𝑎 ]𝐷33
written in a manner similar to convection, i.e., 𝑅𝑎 = (14)
𝛼𝑎𝑖𝑟 𝛾𝑎𝑖𝑟
the radiation rate equation is linearized Where, = Thermal expansion coefficient
making the heat rate proportional to the = Thermal diffusivity
air
Page 147
Therefore the combined heat transfer Insulation Economics
coefficient is computed as, The economic thickness of insulation depends
1⁄
4
𝑁𝑢𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑 = [𝑁𝑢𝑓𝑟𝑒𝑒 4
+ 𝑁𝑢𝑓𝑜𝑟𝑐𝑒𝑑 ] 4
(15) on the insulating & maintenance costs and
also the annual value of heat loss. This would
𝑘𝑎𝑖𝑟 𝑁𝑢𝑐𝑜𝑚𝑏
ℎ𝑐𝑜𝑚𝑏 = (16) depend on the cost of producing energy &
𝐷3
thermal conductivity of the lagging. Generally
Therefore the external heat transfer
thicker insulation will represent higher
coefficient, hair, overall, is computed as,
investment costs and lower heat loss costs.
ℎ0 = ℎ𝑐𝑜𝑚𝑏 + ℎ𝑟 (17) The annual heat losses are computed as,
Bare Pipe & Insulation Resistance 𝑄
𝐶𝐿𝑜𝑠𝑠 = ( ) [𝑘𝑊] × 𝑛 × 𝑄𝑐𝑜𝑠𝑡 /𝑘𝑊ℎ (27)
𝐴 𝐿𝑜𝑠𝑠
The resistance offered by the bare pipe &
Where, n = number of annual hours
insulation is estimated as follows,
Insulation Costs is the product of insulation
𝐷
𝐷2 𝑙𝑛[ 2 ]
𝑅𝑏𝑎𝑟𝑒 =
𝐷1
(18)
volume and insulation cost per m3.
2𝑘𝑝𝑖𝑝𝑒
𝜋
𝐷 𝑉𝐼𝑛𝑠 [𝑚3 ] = 4 [𝐷32 − 𝐷22 ] × 𝐿𝑒 (28)
𝐷3 𝑙𝑛[ 3 ]
𝐷2
𝑅𝐼𝑛𝑠 = (19)
2𝑘𝑖𝑛𝑠 𝐶𝐼𝑛𝑠 = 𝑉𝑖𝑛𝑠 × [𝐶𝑖𝑛𝑠 ⁄𝑚3 ] (29)
Total Resistance – Bare Pipe & Insulation Labour Costs is the product of cost per unit
For bare pipe, the total resistance is metre & length of pipe
calculated as,
𝐶𝐿𝑎𝑏𝑜𝑢𝑟 = 𝐿𝑒 × [𝐶𝑙𝑎𝑏 ⁄𝑚] (30)
1
𝑅𝑇𝑜𝑡𝑎𝑙 = 𝑅𝑏𝑎𝑟𝑒 + (20)
ℎ0 The cost of energy losses is quantified by the
For Insulated Pipe, the total resistance is Net Present Value (NPV) of the future energy
calculated as, costs during an insulation life of typically 5
𝑅𝑇𝑜𝑡𝑎𝑙 = 𝑅𝐵𝑎𝑟𝑒 + 𝑅𝐼𝑛𝑠 +
1
(21) years. For this module, a discount rate [i] of
ℎ0
15% is used. The number of annual working
Piping Heat Losses
hours is taken as 8,000 hours, cost of energy
𝑈 = 1⁄𝑅 (22) (electricity to run the gas compressor) is
𝑇𝑜𝑡𝑎𝑙
taken as 0.10€/kWh and the insulation cost is
𝑄
( ) [𝑊 ⁄𝑚2 ] = 𝑈 × (𝑇1 − 𝑇𝑎 ) (23) taken as 50€/m3. The annual value of the
𝐴 𝐿𝑜𝑠𝑠

𝑄 𝑄 energy losses for 5 years is calculated as,

(𝐴 ) [𝑊 ⁄𝑚] = ( ) [𝑊 ⁄𝑚2 ] × 𝜋𝐷3 (24)
𝐿𝑜𝑠𝑠 𝐴 𝐿𝑜𝑠𝑠 1−[1+𝑖]−𝑛
𝑁𝑃𝑉 = 𝑅 × (31)
𝑄 𝑖
𝑄𝐿𝑜𝑠𝑠 [𝑘𝑊] = ( ) [𝑊 ⁄𝑚] × 𝐿𝑒 (25)
𝐴 𝐿𝑜𝑠𝑠
Where, R is the cost of energy losses
𝑄
𝑇𝑎𝑖𝑟𝑓𝑖𝑙𝑚,𝑖𝑛𝑠,2 = 𝑇1 − [(𝐴 ) [𝑊 ⁄𝑚] × 𝑅𝑖𝑛𝑠 ] (26) The annual total cost is computed as,
𝐿𝑜𝑠𝑠
𝐶𝑇𝑜𝑡𝑎𝑙 = 𝑁𝑃𝑉 + 𝐶𝐼𝑛𝑠 + 𝐶𝐿𝑎𝑏𝑜𝑢𝑟 (32)
The above set of heat transfer calculations are
performed first for a bare pipe & then The Insulation thickness corresponding to the
performed for various insulation thicknesses lowest total cost will be the economic
to estimate the heat losses, QLoss [W/m2] and thickness of insulation.
QLoss [kW], QLoss [kWh/year] which is References & Further Reading
computed by multiplying QLoss [kW] with the 1. “Fundamentals of Heat and Mass Transfer”,
annual working hours. Incropera, DeWitt, Bergman, Lavine, 6th
Edition.
Page 148
Appendix A: Heat Transfer Coefficients

Appendix B: Economic Analysis

Page 149
 Module 22
FRONT END LOADING FOR PIPELINE PROJECT MANAGEMENT
Abstract: With growing clamour for clean pressure and temperature profiles
energy globally, the midstream industry determine the pipeline wall thickness.
becomes crucial for any hydrocarbon When your reservoir engineer and
exporting country. To have an effective production technologist are indecisive
midstream network, would mean about the Stock Tank Oil in place (STOIP),
construction & production costs also have a how much & at what rate the recoverable
role to play in ensuring globally competitive volumes from the wells are going to be
& affordable prices of oil & gas products. One extracted, chances are that you are going to
could argue that short term and long term underestimate/overestimate the pipeline
barrel prices apart from supply and demand sizes. In case of multiphase flow, whether 2
is a motivating factor for operators to invest or 3 phase, the pipeline sizes significantly
in hydrocarbon projects, but it can also be affect your flow regimes and carry the risk
equally said, that despite what the global of slug formation. The slug volumes decide
price or supply & demand is, effective project the size of your slug catcher &
management & execution also determines the underestimating its size can cause
economic success for all project stakeholders. equipment failure. Hence work it out with
A key stage in midstream project your subsurface team to arrive at a
management is Front End Loading (FEL) conclusive and accurate production profile
where strategic information that addresses prior to performing pipeline FEED.
internal and external risks, resource 2. Material Costs Overrun: Nothing can be
availability, allocation and commitment is more disastrous than realizing as the
made before sanctioning or making a final project progresses that your pipeline
investment decision (FID) on the project. actually costs more because of
Simply put, the more homework you do in the underestimating the pipeline’s wall
early stages of a project decides how much thickness. Wall thickness is a key value
success can be achieved. Although project that depends on the design pressure &
management is a vast subject, the following eventually determines the pipeline weight.
module focuses on the some of the Since pipeline weight is proportional to the
repercussions of poorly executed front end square of the outer diameter (OD), for
loading (FEL) steps in midstream activities. every millimetre increase, so does the
weight increase. When the parabolic
Front End Engineering (FEED)
increase of per unit pipeline weight is
1. Choosing Pipeline Sizes: The starting
multiplied with the total kilometres of
point to design any oil & gas pipelines is
pipeline length, the pipeline material costs
the well production, pressure &
are going to probably overrun the project
temperature profile in addition to the
budget. Therefore, it is not just the
composition of the contents that the
engineering standard chosen, but following
pipeline will carry. Production profiles are
this crucial step as part of a check list prior
needed to estimate the peak flow rates
to finalizing the pipeline sizes is a must.
which the pipeline experiences and in turn
determine the pipeline size, whereas the
Page 150
3. Pipeline Corrosion: Various Engineering starts during detailed engineering stage
design practices offer solutions as to what because of material cost overruns.
should be the corrosion allowance for a Contract Management
given pipeline application. Produced water 1. Vendor Contracts: As much as the top
and Hydrogen sulphide (H2S) are the management works on the terms and
popular enemies that contribute to conditions of a production sharing contract
pipeline corrosion. In addition to pipeline (PSC), taxation, governmental regulations,
contents, sand from well fluids that escape etc. that shows its effects on the company’s
sand traps, hydrate particles and fluid flow balance sheets, so must the procurement
rates exacerbate metal erosion. department spend time due diligently to
However, engineers sometimes fail to ensure that the right vendors are available
account for the effect of external forces. In to deliver material and equipment which
offshore pipelines, sea waves and sand affects project schedule and costs. A classic
underneath the soil act upon the pipeline case of project cost overrun is when the
components such as risers thereby procurement department realizes that
inducing stress. Human error also needs to there is only one particular vendor to meet
be taken into account where, when ships your project requirements after the project
collide with platforms, can cause dents has been sanctioned. If a procurement
from where corrosion propagates. Pigging strategy and supply chain is not in place, it
operations and depressurization to can leave the project to the mercy of the
dislodge hydrates can also contribute to sole vendor. Hence ITT (Invitation to
pipeline corrosion simply because pigs can Tender) documents must be prepared at
cause dents when their velocities are not the earliest to receive competing offers
regulated properly resulting in the pig from various bidders which in turn allow
getting stuck. When hydrates get dislodged project managers to prepare realistic
during a depressurization, there are good schedules and costs incurred.
chances that a high velocity column of 2. Interface Management: Project
hydrate can collide with pipeline bends management also includes interface
thereby cracking and exposing metal to management. In midstream projects when
corrosion effects. facilities such as booster compressors,
Therefore, a key step during FEED is for sectionalizing valves & burn pit lines are
engineering teams to take time out and vendor items, it is important for
allocate resources to do a pipeline stress contractors to keep a constant open line of
analysis, on-bottom stability analysis, a communication with vendors to ensure
basic corrosion management plan covering that the engineering & hook-up drawings
pipeline coatings & cathodic protection, a and datasheets have been followed to meet
risk assessment report and Pipeline project specifications prior to execution.
Integrity Management (PIM) report to There is nothing more upsetting for project
ensure that the wall thicknesses & managers to watch their tables pile up with
supporting structures chosen is adequate Change Orders (CO). When equipment that
to meet all internal and external risks that is already manufactured & delivered to the
the pipeline can experience. If one argues site but do not conform to the project
that this is a far fetched vision during early specifications because of poor
FEED, wait till you see blame game that communication with the vendor, it is the
Page 151
 homework of developing a contracting assess, if the scope of work (SoW), quality
strategy, regular project review by and execution schedule has been met to
engineering teams at the early stages satisfaction and avoid feeling like cab
which can minimize the damages to the passengers who constantly suspects if the
project’s cost & schedule. Otherwise even cab meter is functioning correctly. This
your legal department might be left out in also allows project owners to assess if re-
the open to dry under the sun. work through change orders will become a
3. Man-hours Billing & LSTK: Between habit during the detailing phase of
Projects awarded on man-hour billing vs. engineering.
Lump sum Turn key (LSTK) contracts, in 4. Under/Over Quoting: When contractors
reality, it depends on factors such as - how lack experience with similar projects in the
well the project owner defines the scope & past to determine what it takes to execute
shares the project vision. During FEL stage a project, it either ends up working for free
of long term projects, if project charters on all that ‘extra’ SoW or losing the E&C
that have requirements changing contract to a brighter guy. Therefore if you
dynamically and frequently, chances are are a small contractor, start small.
that the engineering contractor would 5. Take Benchmarking Seriously: High FEL
hesitate to engage with the project owner projects or projects that have clear vision,
on LSTK basis (unless the contractor is clarity and scope are expected to have
desperate for the money to keep his shorter schedules, predictable costs, and
company afloat). Basically, if the project completion in all respects. Benchmarking
requirements are not expected to change with similar projects that had sound
much during the course of the project, the contracts with reputed suppliers gives a
project owner can negotiate to put the good idea where your project is heading
budget and schedule risks on the towards, quicker and confident final
contractor on LSTK terms. investment decisions (FID) and also aids in
But to be practical, there is no project, eliminating uncertainties that warrant
where complete clarity is always available excess contingencies.
prior to Sanctioning/Final Investment 6. Project Economics & Standardization:
Decision (FID). Hence it is prudent for When projects run on low profit margins,
project owners to keep their options open instead of cutting corners and getting into
to enter into a mixed contract where both trouble, it is more sensible to first
man-hour billing and LSTK methods understand how project economic factors
provide flexibility, transparency, such as direct & indirect costs, revenue,
accountability and ease of management to margin, overheads, taxation affect the
the project. Typically FEED follows a man- project’s profitability. Contracts made with
hour billing cycle & Detailing work follows vendors who employ Product
LSTK terms. For such mixed contracts, the Standardization, maintain sound balance
onus is on the engineering contractor to sheets, ready availability of credit better
prove transparency and accountability bet to ensure your project’s vendor items
during and after the man-hour based FEED are delivered on time to meet project
by maintaining clear open book records on schedules and quality. Therefore, it is
the project progress & work delivered. preferable during FEL for procurement
This is important for the project owner to teams to refer back to previously approved
Page 152
 contractors who meet project owner’s managers have the mandatory task of
business objectives because they better keeping track of Local environmental
understand the Project Owner’s regulations from the earliest stage of Front
requirements. End Loading.
Local Laws & Regulations 3. Planning ROW Path: Not all projects are
1. Pipeline Location Markers: It is no doubt expected to receive the kind of budgets to
a momentous joy in meeting project build oil & gas facilities in one go and
requirements, executing, completing, hence projects are implemented in phases.
running a guarantee test, handing over the Sometimes, though budgets are sanctioned,
keys of the facilities to the project owner & projects are not implemented considering
closing the business deal. But if a buried unforeseeable poor market demand. In the
natural gas pipeline that runs through event where the project is expected to go
large localities of human occupation through a later stage expansion or when an
without any pipeline location markers & underestimation of market demand causes
the local government body in-charge of downsizing the infrastructure but market
laying roads & electrical cables hit the gas demand increases at a later stage, the
pipeline while digging up, in all likelihood existing pipeline capacity becomes
the incident will hit the tabloids when insufficient to transport. In such cases, it is
there is an explosion. Hence always have prudent to plan early during the FEL stage
an emergency response plan of action as to acquire and accommodate additional
part of the project plan with constant ROW for future pipeline expansions.
communication with local civic authorities. However to do so, local landowners and
governmental authorities must be
2. Right of Way (ROW): Project Owners,
consulted early to acquire the requisite
Project Managers, project engineers and all
land and approvals for gaining ROW rights.
relevant stake holders have the duty to
follow all local laws and meet References & Further Reading
environmental regulations. When project 1. Lessons Learned from UKCS Oil and Gas
owners skip such an early FEL step & Projects 2011-2016, Oil & Gas Authority
Engineers are busy proving their calibre
laying an above-ground gas pipeline laden
with high H2S content through a forest area
with no cognizance of the local habitat or
environmental regulations, one wouldn’t
want to see an elephant stepping on it.
When the pipeline ruptures, with all that
hydrogen sulphide laden gas spewing out
killing the surrounding habitat due to
poisoning & explosions, the project owner
can be sure to become the next subject of a
Hollywood movie or a Greenpeace
Activist’s Documentary.
In protected habitats, the ROW of local
flora & fauna gains first priority over
Project Owner’s ROW. Therefore, project
Page 153
 Module 23
Flash Steam and Steam Condensates in Return Lines
In power plants, boiler feed water is 4. Proper sizing of stem condensate return
subjected to heat thereby producing steam lines requires consideration of all
which acts as a motive force for a steam operating scenarios, chiefly start up,
turbine. The steam upon doing work loses shutdown and during normal running
energy to form condensate and is conditions. During plant start up, steam is
recycled/returned back to reduce the not generated instantly. As a result, the
required make up boiler feed water (BFW). condensate lines would be filled with
liquids which gradually turn two-phase
Recycling steam condensate poses its own
until reaching normal running conditions.
challenges. Flash Steam is defined as steam
During shutdown conditions, with time,
generated from steam condensate due to a
flash steam in the lines condense leaving
drop in pressure. When high pressure and
temperature condensate passes through behind condensates due to natural cooling.
process elements such as steam traps or 5. Condensate return line design must also
pressure reducing valves to lose pressure, the consider the effects of water hammering.
condensate flashes to form steam. Greater the When multiple steam return lines are
drop in pressure, greater is the flash steam connected to a header pipe that is routed
generated. This results in a two phase flow in to a flash drum, flash steam in the presence
the condensate return lines of cooler liquid from other streams would
condense rapidly to cause a water hammer.
General Notes
1. To size condensate return lines, the Fraction of Flash Steam
primary input data required to be Taking an example case, condensate flows
estimated is A. Fraction of Flash Steam and across a control valve from an upstream
condensate, B. Flow Rates of Flash Steam pressure of 5 bara to 2 bara downstream. The
& condensate, C. Specific volume of flash saturation temperature at 5 bara is 151.84 0C
steam & condensates, D. Velocity limits & 120.20C at 2 bara. The specific volume of
across the condensate return lines. water at 5 bara is 0.001093 m3/kg & 0.00106
m3/kg at 2 bara. The latent heat of saturated
2. Sizing condensate return lines also require
steam upon reaching 2 bara is 2201.56 kJ/kg.
lower velocity limits for wet steam since
liquid droplets at higher velocities cause The % flash steam generated is estimated as,
% 𝐹𝑙𝑎𝑠ℎ
internal erosion in pipes and excessive ℎ𝑓,1 = ℎ𝑓,2 + [ × ℎ𝑓𝑔 ] (1)
100
piping vibration. A rule of thumb, for
Where,
saturated wet steam is 25 – 40 m/s for
short lines of the order of a few tens of hf,1 = Upstream specific enthalpy [kJ/kg]
metres and 15 - 20 m/s for longer lines of hf,2 = Downstream specific enthalpy [kJ/kg]
the order of a few hundred metres. hf,g = Latent Heat of Saturated Steam [kJ/kg]
3. Condensate return lines work on the The upstream specific enthalpy, hf1 of
principle of gravity draining. To effectuate saturated water at 5 bara is 640.185 kJ/kg
this, drain lines are to be sloped downward and hf2 of 504.684 kJ/kg at 2 bara. The steam
at a ratio of atleast 1:100. specific volume at 2 bara is 0.8858 m3/kg.

Page 154
The fraction of flash steam is calculated as, The dynamic viscosity for calculating the
[640.185−504.684] Reynolds number can be chosen as the
% 𝐹𝑙𝑎𝑠ℎ = × 100 = 6.15% (2)
2201.56
viscosity of the liquid phase or a quality
Therefore the condensate fraction is, averaged viscosity, µh.
% 𝐶𝑜𝑛𝑑 = 100 − 6.15 = 93.85% (3)
𝜇ℎ = 𝑥𝜇𝑣 + [1 − 𝑥]𝜇𝐿 (8)
The steam volume is calculated as,
The homogenous model for gravitational
𝑚3
𝑉𝑆𝑡𝑒𝑎𝑚 = 0.8858 × 0.0615 = 0.05448 𝑘𝑔
(4) pressure drop is applicable for large drop in
pressures and mass velocities < 2000
The condensate volume is calculated as,
kg/m2.s, such that sufficient turbulence exists
𝑚3
𝑉𝐶𝑜𝑛𝑑 = 0.00106 × 0.9385 = 0.000995 𝑘𝑔 (5) to cause both phases to mix properly and
ensure the slip ratio (uv/uL) between the
Condensate Return Pipe Sizing
vapour and liquid phase is ~1.0. For more
To size the condensate return line, the bulk
precise estimates capturing slip ratios and
properties and mixture properties can be
varying void fraction, correlations such as
used to estimate the pipe size. It must be
Friedal (1979), Chisholm (1973) or Muller-
remembered that as the two-phase mixture
Steinhagen & Heck (1986) can be used.
travels through the pipe, there is a pressure
profile that causes the flash % to change The total pressure drop is the sum of the
along the pipe length. Additionally due to the static head, frictional pressure drop &
pipe inclination, a certain amount of static pressure drop due to momentum pressure
head is added to the total pressure drop. gradient.
To estimate the pipe pressure drop across the ∆𝑃𝑇 = ∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 + ∆𝑃𝑚𝑜𝑚 + ∆𝑃𝑓𝑟𝑖𝑐 (9)
pipe length, a homogenous model for
modelling the two phase pressure drop can The Static Head [ Pstatic] is computed as,
be adopted. The homogenous mixture acts as 𝐻×𝜌ℎ× 𝑔×𝑆𝑖𝑛𝜃
∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 [𝑏𝑎𝑟] = (10)
a pseudo-fluid, that obeys conventional 105

design based on single phase fluids Where,

characterized by the fluid’s average
H = Pipe Elevation [m]
properties.
= Pipe inclination w.r.t horizontal [degrees]
The mixture properties can be estimated as,
The pressure drop due to momentum
𝜌ℎ = 𝜌𝐿 [1 − 𝜀ℎ ] + 𝜌𝑣 𝜀ℎ (6) pressure gradient [ Pmom] is,
Where,
𝑑𝑃 𝑑(𝑚⁄𝜌ℎ )
L = Condensate Density [kg/m3] = (11)
𝑑𝑍 𝑑𝑍

v = Steam Density [kg/m3] If the vapour fraction remains constant across

the piping, the pressure drop due to
h = Homogenous void fraction for a given
momentum pressure gradient is negligible.
steam quality [x] [-]
The frictional pressure drop is calculated as,
The homogenous void fraction [ h] for a given
𝑓×𝐿×𝜌ℎ ×𝑉 2
steam quality [x] can be estimated as, ∆𝑃𝑓 = (12)
2𝐷
1
𝜀ℎ = 𝑢 1−𝑥 𝜌𝑣
(7) Where, P = Pressure drop [bar]
1+[ 𝑣 × × ]
𝑢𝐿 𝑥 𝜌𝐿
f =Darcy Friction Factor [-]
Page 155
L = Pipe Length [m] The Colebrook equation was developed
taking into account experimental results for
h = Mixture Density [kg/m3]
the flow through both smooth and rough pipe.
V = Bulk fluid Velocity [m/s] It is valid only in the turbulent regime for
D = Pipe Inner Diameter, ID [m] fluid filled pipes. Due to the implicit nature of
DVρh this equation it must be solved iteratively. A
Re = (13)
µh result of suitable accuracy for almost all
Where, µh = Dynamic Viscosity [kg.m/s] industrial applications will be achieved in less
than 10 iterations. For Reynolds number up
h = Homogenous Density [kg/m3]
greater than ~4000,
The Darcy Friction Factor [f] depends on the
1 ε⁄DH 2.51
Reynolds number follows the following = −2 log10 [ + ] (18)
√f 3.7 Re√f
criteria,
Homogenous Property Calculations
If Re <= 2100 ; Hagen Poiseuille’s Equation
The two phase mixture flows through the
If Re <= 4000 ; Churchill Equation condensate return line. The associated
If Re > 4000 ; Colebrook Equation density and viscosity of flash steam and
The Laminar Flow equation also referred to condensate at 2 bara and 120.20C is,
as the Hagen Poiseuille’s equation is, 𝜌𝑣 =
1
= 1.129
𝑘𝑔
(19)
0.8858 𝑚3
64
f= (14) 1 𝑘𝑔
Re 𝜌𝐿 = = 943.4 (20)
0.00106 𝑚3
The Churchill equation combines both the 𝑘𝑔
𝜇𝑣 = 0.000229 (21)
expressions for friction factor in both laminar 𝑚.𝑠

& turbulent flow regimes. It is accurate to 𝜇𝐿 = 0.0000128

𝑘𝑔
(22)
𝑚.𝑠
within the error of the data used to construct
The homogenous void fraction [ h] for a slip
the Moody diagram. This model also provides
an estimate for the intermediate (transition) ratio (uv/uL) of 1.0, i.e., uv = uL, and a steam
region; however this should be used with quality [x] of 6.15% is,
caution. 𝜀ℎ =
1
1−0.0615 1.129 = 0.9821 (23)
1+[1× × ]
The Churchill equation shows very good 0.0615 943.4

agreement with the Darcy equation for The two phase homogeneous density is,
laminar flow, accuracy through the 𝜌ℎ = 943.4 × [1 − 0.9821] + [1.129 × 0.9821] (24)
transitional flow regime is unknown & in the
𝑘𝑔
turbulent regime a difference of around 0.5- 𝜌ℎ = 18.014 (25)
𝑚3
2% is observed between the Churchill
The two phase homogeneous viscosity is,
equation and the Colebrook equation. For
0.0615×1.28 [1−0.0615]×2.29
Reynolds number up to ~4000, 𝜇ℎ = + (26)
105 104
1⁄ 𝑘𝑔
8 12 1 12 𝜇ℎ = 0.000216 (27)
f = 8 [(Re) + (A+B)1.5
] (15) 𝑚.𝑠

16 Pressure Drop Calculations

A = [2.457ln (
1
)] (16) The return condensate line from the control
7 0.9 ε
( ) +0.27
Re D valve discharge is sloped at a ratio of 1:100
37,530 16 for gravity drain. The layout of the return
B = [( )] (17)
Re condensate line is,
Page 156
 The condensate return line mixture fluid
velocity is calculated as,
1000 1
Q 4×[ ]×
18.01 3,600
V= = = 11.7142 𝑚/𝑠 (30)
A π×[0.04094]2

The Reynolds number is estimated as,

ID×V×𝜌ℎ 0.04094×11.7×18.01
Re = = (31)
𝜇ℎ 0.000216

Re ≈ 39,971 (MS-Excel computed) (32)

Figure 1. Condensate Return Line to Receiver
Since the Reynolds number is much higher
The condensate receiver operates at 1.1 bara
than 4,000, the flow is fully turbulent and the
pressure. The mechanical details of the piping
friction factor is calculated based on
for a flow rate of 1,000 kg/h, pipe size of 1.5”,
Colebrook equation. The friction factor is
100m length & pipe roughness of 45.2 m is,
estimated as,
Table 1. Condensate Return Line Details
𝑓 = 𝑓𝐶𝑜𝑙𝑒𝑏𝑟𝑜𝑜𝑘 = 0.0251 (33)
Parameter Value Unit
The frictional pressure drop is now calculated
Mass Flow rate [m] 1000.0 kg/h
using the Darcy-Weisbach expression as,
Volumetric Flow [Q] 55.51 m3/h
0.0251×100×18.01×11.72
Pipe Length [L] 100 m ∆𝑃𝑓 = (34)
2×0.04094×105
Pipe Roughness [ε] 45.2 μm ∆𝑃𝑓 = 0.757 𝑏𝑎𝑟 (35)
Pipe Outer Diameter [OD] 48.3 mm
The slope angle is calculated as,
Pipe SMYS [Carbon Steel] 30,000 psi 1 180
𝜃 = [𝑇𝑎𝑛−1 (100)] × = 0.6° (36)
Pipe Design Pressure [DP] 7 bara 𝜋

Pipe Wall Thickness [WT] 0.08 mm The static pressure drop [ Pstatic] becomes
Corrosion Allowance [CA] 1.0 mm 18.01×9.81×[(1+5)×𝑠𝑖𝑛(0.6°)]
∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 = (37)
105
Calculated WT 1.08 mm
Selected WT 3.68 mm ∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 = 0.000106 𝑏𝑎𝑟 (38)

Pipe Inner Diameter [ID] 40.94 mm Therefore the total P with negligible P due
to momentum pressure gradient [ Pmom].
The pipe wall thickness chosen is based on
ASME/ANSI B36.10M and is calculated based ∆𝑃𝑡𝑜𝑡𝑎𝑙 = ∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 + ∆𝑃𝑓 (39)
on the hoop stress created by internal ∆𝑃𝑡𝑜𝑡𝑎𝑙 = 0.757 + 0.000106 = 0.757 𝑏𝑎𝑟 (40)
pressure in a thin wall cylindrical vessel as,
The condensate exit pressure is 2 – 0.757 =
48.3
[7×14.5]×[ ]
WT =
DP×OD
= 25.4
× 25.4 (28) 1.243 bara which is higher than the receiver’s
2×SMYS 2×30,000 operating pressure of 1.1 bara.
WT = 0.08𝑚𝑚 (29)
References & Further Reading
Adding CA of 1 mm, the WT becomes 1.08
1. “Engineering Data Book III”, Ch 13, Two
mm. Based on ASME/ANSI B36.10M, the Phase Pressure Drop, Wolverine Tube, Inc.
selected WT is 3.68mm. The inner diameter
2. “Steam Handbook”, Dr. Ian Roberts, Philip
calculated for the selected WT is 40.94 mm.
Stoor, Michael Carr, Dr. Rainer Hocker,
Oliver Seifert, Endress+Hauser
Page 157
Appendix A

Page 158
 Module 24
Single Phase Liquid Vessel Sizing for HYSYS Dynamics
Process Facilities often have intermediate to a peak flowrate rise of 1,474 m3/h in a 2 min
storage facilities that store liquids prior to interval. The liquid level rises to ~57% from
transporting to downstream equipment. The 50%.
period of storage is short, i.e., of the order of
minutes to hours & is defined as Holdup time.
The Holdup time can also be explained as the
reserve volume required to ensure safe &
controlled operation of downstream equipment.
The intermediate vessel also acts as a buffer
vessel to accommodate any surge/spikes in
flow rates, and is termed as surge time. Vessel
volume is an input data required in process
dynamic simulation and the following covers
estimation of volume required for single phase
liquid flow into an intermediate
Figure 1. Surge Flow rate increase on %
vertical/horizontal/flat bottomed vessel.
Liquid Level
Problem Statement Thumb Rules
Water at 1,341 m3/h, flows into a vessel & held The different arrangement types are as follows,
for a holdup time of 1 min before discharging
into downstream equipment. The vessel’s liquid
percent level is desired to be held at 50% (half
full) for a certain drain rate. Estimate the size of
the vessel required for an L/D ratio of 1.
Design Methodology & Results
Based on the above data, the vessel volume [V]
for a flow rate of 1,341 m3/h is, Figure 2. Intermediate Storage Vessel Types
& L/D Ratio
𝑻𝒉𝒐𝒍𝒅𝒖𝒑 [𝒎𝒊𝒏] 𝟏𝟎𝟎
𝑽 = 𝑸[𝒎𝟑 ⁄𝒉] × × % 𝑳𝒊𝒒𝒖𝒊𝒅 𝑳𝒆𝒗𝒆𝒍 (1) 1. Holding time for most intermediate tanks is
𝟔𝟎
10 min (half full tanks)
Based on the vessel volume estimated for the
2. Holding time for feed tanks to furnace is 30
holdup time, the dimensions of the vessel are,
min (half full).
𝟒𝑽
𝑫 = √𝝅𝑳 ; 𝑾𝒉𝒆𝒓𝒆 𝑳 = 𝒏 × 𝑫 (2) 3. The optimum ratio is 3 for commonly used
L/D ratios of 2 to 5.
Substituting the values for vessel dimensions,
4. Vessels < 4 m3, are vertically mounted with
𝟏 𝟏𝟎𝟎
𝑽 = 𝟏𝟑𝟒𝟏 × 𝟔𝟎 × = 𝟒𝟒. 𝟕𝒎𝟑 (3) L/D ratio of 2 to 5 on leg supports/ brackets.
𝟓𝟎
5. Vessels 4 m3 < V < 40 m3 is horizontal &
Taking L = D, i.e., L/D ratio of 1.0
saddle supported with L/D ratio of 2 to 5.
𝑫=√
𝟑 𝟒×𝟒𝟒.𝟕
≈ 𝟑. 𝟗 𝒎, For L = D; L = 3.9 m (4) 6. Vessels V > 40 m3 is flat bottom tank on
𝝅
concrete foundation with L/D ratio of 0.5-1.5.
Surge Study References & Further Reading
From the estimates, the vessel chosen is flat Perry’s Chemical Engineers Handbook, 7th
bottomed on concrete foundation & is subjected Edition

Page 159
 Module 25
Key Thermo-Physical Properties of Light Crude Oils
Process facilities are equipped with 3. The ASTM D-86 distillation of an oil
protection measures, such as pressure safety fraction is conducted in laboratory room
valves (PSV) & as a minimum, PSVs are sized conditions at 1 atm and the D-86
for a fire case. To do so for a pressure vessel distillation curve ends at ~6500F (3440C).
containing crude oil a key parameter is the 4. The ASTM D-1160 distillation of an oil
Latent heat of Vaporization [Hv]. fraction is conducted at much lower
For pure components, the Joback’s Method pressures, typically 10 mmHg for heavier
can be employed which uses basic structural oils with high boiling points to prevent
information of the chemical molecule to decomposition of the oil sample. With this
estimate thermo-physical data. However it method, oil fractions can be distilled upto
can be complex for equipment that contains ~9500F to ~10000F (5100C to 5380C),
crude oil because the plus fractions [C7+] can reported on a 760mm Hg basis.
contain thousands of straight chain, cyclic & 5. The boiling point of all compounds in a
functional groups. Therefore by splitting and crude mixture can be represented by a
lumping the crude fractions, a smaller single characteristic boiling point called
number of components are arrived at, to Volume Average Boiling Point [VABP].
characterize and be able to apply Equation of Since the individual mole fractions of the
State (EoS) correlations to estimate the petroleum stream is not known, VABP is
fraction’s thermo-physical properties. calculated from standard distillation data
To estimate properties such as MW, Specific [ASTM D-86] followed by calculating the
gravity [ ], Critical Pressure [Pc], Critical mean average boiling point [MeABP].
Temperature [Tc] and Latent heat of 6. The Molecular weight [MW], Specific
Vaporization [Hv], the following module gravity [ ] & boiling point [Tb] are taken as
provides few correlations applicable for light the key properties to define the makeup of
crudes with boiling points < 4550C based on a petroleum fraction. In this module, the
D-86 Distillation curves. Katz-Firoozabadi [1978], Riazi-Daubert
General Notes [1980, 1987] & Ahmed [1985] correlations
1. Latent heat of Vaporization [Hv], can be are shown to predict MW, specific gravity
estimated using critical properties of the [ ], Critical Pressure [Pc] & Critical
plus fractions in the hydrocarbon mixtures. Temperature [Tc]. To estimate Hv, Riedel
2. Oil fractions tend to decompose at ~6500F correlation is employed to estimate the
(3440C) at 1 atm. As a result, it becomes Latent Heat of Vaporization [Hv,NBP] at
necessary to lower the pressure to as low Normal Boiling Point [MeABP/NBP/Tb].
as 40mm Hg to obtain the True Boiling Watson relation is used to estimate HV,T at
Point (TBP) distillation curves. ASTM desired temperature.
methods can be used to convert the Selected Correlations
resulting boiling point curve into TBP The below table gives a summary of the two
curves using correlations from API generalized correlations to estimate MW, Pc,
Technical Data Book – Petroleum Refining. Tc and Hv of the petroleum fraction.

Page 160
Table 1. Generalized Correlations for Pc, Tc and MW Kreglewski and Zwolinski [1961] generalized
Katz-Firozabaadi Correlation [1978] expression which is of the form,
𝑐 )]
𝐤𝐠 𝟔.𝟗𝟕𝟗𝟗𝟔−𝐥𝐧[𝟏𝟎𝟖𝟎−𝑻𝒃 ]
𝟑⁄
𝟐 𝜃 = 𝜃 − 𝑒 [𝑎−(𝑏×𝑀𝑊 (1)
𝐌𝐖 [𝐤𝐦𝐨𝐥] = [ ]
𝟎.𝟎𝟏𝟗𝟔𝟒
Where, represents the properties such as
[𝟑.𝟓𝟔𝟎𝟕𝟑−(𝟐.𝟗𝟑𝟖𝟖𝟔×𝐌𝐖 𝟎.𝟏 )]
𝛄[−] = 𝟏. 𝟎𝟕 − 𝐞 Tb, , Pc and Tbr.

𝑷𝒄 [𝒃𝒂𝒓𝒂] = 𝐞[𝟔.𝟑𝟒𝟒𝟗𝟐−(𝟎.𝟕𝟐𝟑𝟗×𝐌𝐖
𝟎.𝟐𝟗𝟗 )]
The value of the constants, a, b, c in the above
expression is based on a tabulated set
𝟎.𝟓𝟓 )]
𝑻𝒃𝒓 [−] = 𝟏. 𝟐 − 𝐞[−𝟎.𝟑𝟒𝟕𝟒𝟐−(𝟎.𝟎𝟐𝟑𝟐𝟕×𝐌𝐖 generated from the physical properties of 26
condensates and crude oil systems. The value
𝐓
𝐓𝐜 [°𝐊] = 𝐓 𝐛 of based on MW in the Katz-Firoozabadi
𝐛𝐫

𝟑⁄
correlation predicts within 0.4% for straight
[𝟔.𝟗𝟗𝟓𝟓−𝐥𝐧(𝟏𝟎𝟗𝟎−𝑻𝒃 )] 𝟐
𝐍𝐜 [−] = [ ] chain numbers [SCN] groups from C6 to C50.
𝟎.𝟏𝟏𝟏𝟗𝟑
Similarly, the absolute average deviation
𝐓𝐛 = °𝐊
(AAD%) of the Kreglewski and Zwolinksi
Riazi-Daubert Extended Correlation [1980] [1961] correlation gives an AAD% of 0.4%,
𝐤𝐠 0.07%, 0.15% and 1% in the properties of Tb,
𝐌𝐖 [𝐤𝐦𝐨𝐥] = [𝟒𝟐. 𝟗𝟔𝟓 × 𝐓𝐛𝟏.𝟐𝟔𝟎𝟎𝟕 ×
, Tbr, Pc respectively between correlations
𝛄𝟒.𝟗𝟖𝟑𝟎𝟖 ] × and physical properties of the 26 condensates
𝐞[(𝟎.𝟎𝟎𝟎𝟐𝟎𝟗𝟕×𝐓𝐛)−(𝟕.𝟕𝟖𝟕𝟏𝟐×𝛄)+(𝟎.𝟎𝟎𝟐𝟎𝟖𝟒𝟕𝟔×𝐓𝐛×𝛄)] and crude oil systems.

𝐓𝐛 = °𝐊; 𝛄 = 𝐔𝐧𝐢𝐭𝐥𝐞𝐬𝐬 VABP and MeABP Calculation

For petroleum fractions usually, there would
Riazi-Daubert Correlation [1987]
be no information available about the weight,
𝑷𝒄 [𝒑𝒔𝒊𝒂] = [𝟒𝟓𝟐𝟎𝟑 × 𝑴𝑾−𝟎.𝟖𝟎𝟔𝟑 × mole or volume fractions considering the
𝜸𝟏.𝟔𝟎𝟏𝟓 ] × 𝒆[(−𝟎.𝟎𝟎𝟏𝟖𝟎𝟕𝟖×𝑴𝑾)+(−𝟎.𝟑𝟎𝟖𝟒×𝜸)] large number of compounds present. In such
cases, the ASTM based D-86 distillation data
𝐓𝐜 [°𝐑] = [𝟓𝟒𝟒. 𝟒 × 𝐌𝐖 𝟎.𝟐𝟗𝟗𝟖 × 𝛄𝟏.𝟎𝟓𝟓𝟓 ] × for light oils (API Gravity > 310API and D-86
𝐞[(−𝟎.𝟎𝟎𝟎𝟏𝟑𝟒𝟕𝟖×𝐌𝐖)+(−𝟎.𝟔𝟏𝟔𝟒𝟏×𝛄)] Temperatures < 4550C) can be used to
estimate the Volume Average Boiling Point
The Latent Heat of Vaporization [Hv] is
(VABP) and Mean Average Boiling Point
calculated as,
(MeABP) which can be calculated as follows,
Table 2. Riedel Correlation and Watson Relation
𝑇10% +𝑇30% +𝑇50% +𝑇70% +𝑇90%
Riedel Correlation 𝑉𝐴𝐵𝑃[℃] = (2)
5

𝟏.𝟎𝟗𝟐×𝟖.𝟑𝟏𝟒𝟓×𝐓𝐛 ×[𝐥𝐧𝐏𝐜 −𝟏.𝟎𝟏𝟑] It is to be noted that when the average boiling

𝐇𝐯,𝐍𝐁𝐏 [𝐤𝐉/𝐦𝐨𝐥] = 𝐓
[𝟎.𝟗𝟑− 𝐛 ]×𝟏𝟎𝟎𝟎
𝐓𝐜
point (ABP) of a crude sample is estimated
based on weight (W), moles (M) and volume
𝐏𝐜 = 𝐛𝐚𝐫𝐚; 𝐓𝐜 = °𝐊 ; 𝐓𝐛 = °𝐊
(V) basis, there would exist a difference in
Watson Relation each of these average boiling points. To relate
𝐓 −𝐓 𝟎.𝟑𝟖
the different types of ABPs, the VABP value is
𝐇𝐯 [𝐤𝐉/𝐦𝐨𝐥] = 𝐇𝐯,𝐍𝐁𝐏 × [𝐓 𝐜−𝐓 ] corrected with a slope line and correction
𝐜 𝐛

𝐓 = °𝐊 ; 𝐓𝐜 = °𝐊 ; 𝐓𝐜 = °𝐊 ; 𝐇𝐯 = 𝐤𝐉/𝐦𝐨𝐥 factor line to find other ABPs. The Slope Line

(S) is estimated as,
It is to be noted that, the Katz-Firoozabadi 𝑇90% −𝑇10%
𝑆[℃⁄% 𝑅𝑒𝑐𝑜𝑣𝑒𝑟𝑒𝑑 ] = (3)
[1978] correlation was originally based on 80

Page 161
With the S value, the correction factor, TMeA Or, 𝑇𝑏 ≅ 243℃ ≅ 516°𝐾 ≅ 929°𝑅 (11)
is estimated using the empirical expression, Applying the MeABP/Tb value, the critical
∆𝑇𝑀𝑒𝐴 = −1.53181 − [0.0128 × 𝑉𝐴𝐵𝑃0.6667 ] + properties, and MW is estimated as follows,
[3.646064 × 𝑆 0.333 ] (4) kg
3
6.97996−ln[1080−516] ⁄2
MW [kmol] = [ ] = 188.4 (12)
Where, 0.01964
0.1 )]
VABP = Volume Average Boiling Point [0C] γ = 1.07 − e[3.56073−(2.93886×188.4 = 0.8238 (13)
141.5
With the correction factor, TMeA, the MeABP API Gravity = 0.8238 − 131.5 = 40.26°𝐴𝑃𝐼 (14)
is estimated as, 0.299 )]
𝑃𝑐 = e[6.34492−(0.7239×188.4 = 17.8 𝑏𝑎𝑟𝑎 (15)
𝑀𝑒𝐴𝐵𝑃 [℃] = 𝑉𝐴𝐵𝑃 − ∆𝑇𝑀𝑒𝐴 (5) 0.55 )]
𝑇𝑏𝑟 = 1.2 − e[−0.34742−(0.02327×188.4 = 0.733 (16)
To estimate the critical properties, MW and
516
latent heat of vaporization [Hv], MeABP Tc = 0.7335 = 704°K (17)
becomes the normal boiling point, Tb. [6.9955−ln(1090−516)]
3⁄
2
Nc = [ ] = 13.78 (18)
Case Study 0.11193

Light Crude Oil is present in a process vessel Similarly applying Riazi-Daubert correlations
at 3250K [51.850C]. To size a PSV for fire case,
from Table 1 with =0.8238 and Tb = 5160K.
the latent heat of vaporization [Hv] value is
kg
required to be computed. The D-86 MW [kmol] = 189.8 (19)
distillation curves are as follows,
𝑃𝑐 = 266 psia = 18.3 𝑏𝑎𝑟𝑎 (20)
Table 3. ASTM D86 Vol% vs Temperature
Tc = 1,254°R = 697°K (21)
Vol% [ASTM D86] D86 Temperature [0C]

0 [IBP] 155.1 The Latent Heat of Vaporization [Hv,NBP]

based on Katz-Firoozabadi Pc, Tc, Tb data is,
10 179.1
30 222.4 Hv,NBP =
9.079434×516×[ln17.8−1.013] 𝑘𝐽
≅ 44.5 𝑚𝑜𝑙(22)
516
[0.93− ]×1000
704
50 260.3
70 289.0 At 3250K, Hv,T is,
90 315.7 704−325 0.38 𝑘𝐽
Hv,T = 44.49 × [704−516] = 58.07 𝑚𝑜𝑙 (23)
100 [FBP] 352.9
Similarly, using Riazi-Daubert Pc, Tc, Tb data,
With the available data, VABP is estimated as, 𝑘𝐽
Hv,NBP = 46.96 𝑚𝑜𝑙 (24)
179.1+222.4+260.3+289+315.7
𝑉𝐴𝐵𝑃[℃] = (6) 𝑘𝐽
5
At 3250K, Hv,T = 61.8 𝑚𝑜𝑙 (25)
𝑉𝐴𝐵𝑃[℃] = 253.3℃ (7)
Additional Correlation – Ahmed [1985]
The slope, S is estimated as,
Based on Ahmed [1985] correlation of the
𝑆 [℃⁄% 𝑅𝑒𝑐𝑜𝑣𝑒𝑟𝑒𝑑] =
315.7−179.1
= 1.7075 (8) Katz-Firoozabadi [1978], physical properties
80
are tabulated with the number of carbon
The correction factor TMeA becomes, atoms using a regression model of the form,
𝑎5
∆𝑇𝑀𝑒𝐴 = −1.53181 − [0.0128 × 253.30.6667 ] + 𝜃 = 𝑎1 + 𝑎2 𝑛 + 𝑎3 𝑛2 + 𝑎4 𝑛3 + (26)
𝑛
[3.646064 × 1.70790.333 ] = 10.1℃ (9) Where,
The MeABP is estimated as, = Tc, Pc
n = number of carbon atoms
𝑀𝑒𝐴𝐵𝑃[℃] = 𝑇𝑏 = 253 − 10.1 ≅ 243℃ (10) a1, a2, a3, a4, a5 =coefficients
Page 162
 Table 4. Ahmed [1985] Constants - and MW Table 7. Average of Estimates

Property MW Average
Property
Properties
Coefficients [-] [kg/kmol]
MW [kg/kmol] 189.0
a1 0.86714949 -131.11375
[-] 0.8242
a2 0.00341434 24.96156
Pc [bara] 18.4
a3 -0.00002840 -0.34079022
Tc [0K] 701
a4 2.4943308 108 0.00249412
Hv,NBP [kJ/mol] 46.05
a5 -1.16279840 468.32575
Hv,3250K [kJ/mol] 60.36
Table 5. Ahmed [1985] Constants - Pc and Tc
References & Further Reading
Property Pc Tc 1. “Physical Properties of Heavy Petroleum
Coefficients [psia] [0R] Fractions and Crude Oils”, Mohammad. R.
Riazi, Taher A. Al-Sahhaf, Fluid Phase
a1 275.56275 915.53747
Equilibria, 117 (1996) 217-224
a2 -12.522269 41.421337
2. “Equation of State and PVT Analysis”, Tarek
a3 0.29926384 -0.7586859 Ahmed, Gulf Publishing Company
a4 -0.00284521 0.00586754 3. “Petroleum Refinery Process Modelling:
Integrated Optimization Tools and
a5 1711.7226 -1302.8779
Applications”, Y.A. Liu, Ai-Fu Chang, Kiran
Based on Ahmed [1985] correlation, Pashikanti, First Edition, 2018 Wiley–VCH
Verlag GmBH & Co.
MW[𝑘𝑔⁄𝑘𝑚𝑜𝑙 ] = 188.7 (27)
4. “Evaluation of Different Correlation
𝑃𝑐 = 266 psia = 19.1 𝑏𝑎𝑟𝑎 (28) Performance for the calculation of Critical
Tc = 1,254°R = 702°K (29) Properties and Acentric Factor of Petroleum
𝑘𝐽 Heavy Fractions”, Dacid B. L, Rafel B. S,
Hv,NBP = 46.71 𝑚𝑜𝑙 (30)
Andre P.C.M.V, Adolfo P. P, Viatcheslav I. P,
𝑘𝐽
At 3250K, Hv,T = 61.15 𝑚𝑜𝑙 (31) http://dx.doi.org/10.5772/intechopen.71166

Results
Summarizing the results,
Table 6. Results Summary
Katz- Riazi- Ahmed
Property
Firoozabadi Daubert [1985]
MW [kg/kmol] 188.4 189.8 188.7

[-] 0.8238 - 0.8245

Pc [bara] 17.8 18.3 19.1

Tc [0K] 704 697 702

Hv [kJ/mol] 44.49 46.96 46.71

Hv,T [kJ/mol] 58.07 61.80 61.15

Taking an average of the estimates made, the

critical properties, MW and Hv is estimated as,
Page 163
Appendix: MS-Excel Calculations

Page 164
 Module 26
Evaporation Pond Process Design in Oil & Gas Industry
In the upstream oil & gas industry, produced The following module focuses on estimating
water is a by-product of well production. the rate of evaporation, water surface
Hydrocarbon wells initially produce less temperature and rate of heat transfer to the
water but in late field life, the water content water in an evaporation pond.
increases. Produced water can contain oil Methodology
carryover and a host of salts with TDS
Similarities exist between mass, momentum
ranging anywhere from 2,000 mg/L to 40,000
& heat transfer phenomenon. Therefore, the
mg/L for which evaporation ponds are used
empirical correlations for heat transfer are
to concentrate by evaporating the associated
also applicable for mass transfer. Schmidt
water.
number plays a similar role to Prandtl
The energy requirement consists of pumping number in convection heat transfer. The heat
concentrate to the pond and in some cases transfer to the water from the air supplies the
aeration is provided to enhance the rate of energy required to evaporate the water,
evaporation. The ponds are lined with
synthetic liner material to prevent seepage of 𝑞 = 𝑚ℎ𝑓𝑔 = ℎ𝐴[𝑡∞ − 𝑡𝑠 ] = ℎ𝑚 𝐴ℎ𝑓𝑔 [𝜌𝑠 − 𝜌∞ ](1)
water into the soil. In case of any corrosive Where,
compounds in the water, the number of layers h = Convective heat transfer coefficient [W/m2.k]
is increased. Landscape and topography play hm = Convective Mass Transfer Coefficient [m2/s]
a role in setting up evaporation ponds and it A = Surface Area [m2]
is necessary to have a flat terrain to avoid any
m = Evaporation Rate [kg/s]
overflow of the contents.
ts, s = Surface temperature & vapour density [K,
Evaporation ponds must also ensure that the kg/m3]
amount of water entering is minimized and
t , = Air Temperature & vapour density [K,
avoid any flooding. As part of waste disposal,
kg/m3]
the ponds maybe designed to accumulate
The energy balance can be arranged as,
sludge over the life time of the operating
wells or can be periodically removed. The − =ℎ [
ℎ 𝑡 −𝑡s
] (2)
𝑠 ℎ𝑓𝑔
below figure depicts an evaporation pond. 𝑚

The heat transfer coefficient, h can be

estimated based on Nusselt Number (Nu) as,
1⁄ 1
𝑁𝑢 = 0.664𝑅𝑒 2 𝑃𝑟 ⁄3 , For Laminar Flow (3)
4⁄ 1
𝑁𝑢 = 0.037𝑅𝑒 5 𝑃𝑟 ⁄3 , For Turbulent Flow (4)
The mass transfer coefficient, hm can be
calculated using Sherwood Number (Sh),
1⁄ 1
𝑆ℎ = 0.664𝑅𝑒 2 𝑆𝑐 ⁄3 , For Laminar Flow (5)
4⁄ 1
𝑆ℎ = 0.037𝑅𝑒 5 𝑆𝑐 ⁄3 , For Turbulent Flow (6)
Where,
𝑆ℎ×𝐷𝑣
Figure 1. Evaporation Pond [2] h𝑚 = (7)
𝐿
Page 165
ℎ=
𝑁𝑢×𝑘
(8) The Reynolds Number (Re) is estimated as,
𝐿
𝑢∞ 𝑎𝑖𝑟 𝐿
Dividing both heat and mass transfer 𝑅𝑒 = (16)
𝜇
coefficients and substituting in Eq. (2) yields,
1⁄ Where,
ℎ 𝑁𝑢×𝑘 𝑃𝑟 3 𝑘
= 𝑆ℎ×𝐷 = [ 𝑆𝑐 ] (9) L = Pond Length along direction of air [m]
ℎ𝑚 𝑣 𝐷𝑣

𝑃𝑟
1⁄
3 𝑘 𝑡 −𝑡s
For the range 00C to 800C, the surface
𝑠
− = [ 𝑆𝑐 ] [ ] (10) temperature [Ts] from curve fit data is,
𝐷𝑣 ℎ𝑓𝑔

The above expression is solved for s and hfg T𝑠 [𝐾] = [(19.45777 × 𝑙𝑛[𝜌𝑠 ]) + 100.4106] + 273.15 (17)
is evaluated at surface temperature [ts]. Air Case Study
properties, k, Dv, Sc, Pr are evaluated at film Air at 250C & 101.325 kPa flows at 10 m/s
temperature [tf], as an average of t and ts. along the length of an evaporation pond of L
𝑡∞ +𝑡𝑠
𝑡𝑓 = 2
(11) W of 10m 2m. The relative humidity is 60%.
The rate of heat transfer to water, rate of
The solution is arrived beginning with a guess
evaporation & the water surface temperature
value of surface temperature, ts in Eq. (10) &
is to be estimated. Evaluating the saturated
iteratively solved until convergence. Relating
vapour pressure,
saturated vapour pressure [Ps] with moist air
temperature [T , 0C] using Arden Buck
25 25
[(18.678− )×( )]
𝑃𝑠 = 6.1121𝑒 234.5 257.14+25 × 100 (18)
equation,
𝑃𝑠 = 3,169 𝑃𝑎 (19)
𝑇∞ 𝑇∞
𝑃𝑠 [𝑇∞ > 0℃] = 6.1121𝑒
[(18.678−
234.5
)×(
257.14+𝑇∞
)]
× 100 (12) The vapour density at 250C is estimated as,
Where, 3,169×18.02 60 𝑘𝑔
∞ = [8314.447×298.15] × [100] = 0.013822 𝑚3 (20)
Ps is saturated vapour pressure [Pa]
The vapour density [ ] for a given relative Taking an initial guess of 150C, the tf is,
humidity [RH] is calculated as, 𝑡𝑓 =
15+25
= 20℃ = 293.15 𝐾 (21)
2
𝑃 ×𝑀𝑊
𝑠 𝑤𝑎𝑡𝑒𝑟 𝑅𝐻%
∞
= [8314.447×𝑇 ] × [ 100 ] (13) Evaluating air properties at tf = 200C, air
∞
=1.1975 kg/m3, =0.0000181 kg/m.s, k =
The mass diffusivity of moisture in air [Dv] is
0.0257 W/m0C, hfg = 2,465 kJ/kg, Pr = 0.7094,
estimated using Sherwood and Pigford, 1952
293.152.5 𝑚2
expression, valid for mass diffusivity of water 𝐷𝑣 = [
0.926
]×[ ]×
1
= 2.5 × 10−5 (22)
101.325 293.15+245 106 𝑠
vapour in air up to 1,1000C 0.0000181
𝑆𝑐 = 1.1975×2.5×10−5 = ~0.6063 (23)
0.926 𝑇 2.5 1
𝐷𝑣 = [𝑃 ] × [𝑇+245] × 106 (14) 1⁄
𝑎𝑚𝑏 0.7094 3 0.0257 298.15−288.15
𝑠
=[ ] × [ ] + 0.01382 (24)
0.6063 2.5×10−5 2,465×1000
Where,
= 0.01822 𝑘𝑔⁄𝑚3 (25)
Dv = Mass diffusivity of moisture in air [m2/s] 𝑠

Pamb = Atmospheric pressure [kPa] Estimating the surface water temperature

[Ts] for s =0.0182 kg/m3,
T = Ambient Temperature [K]
The Schmidt Number (Sc) is estimated as, T𝑠 = [(19.45777 × 𝑙𝑛0.01822) + 100.4106] + 273.15 (26)
𝜇 T𝑠 = 295.62 𝐾 = ~22.5℃ (27)
𝑆𝑐 = 𝜌 (15)
𝑎𝑖𝑟 ×𝐷𝑣
Recalculating the air properties & iterating
Where, = Dynamic Viscosity [kg/m.s] the calculations, Ts = 200C, s = 0.01601
air = Air density [kg/m3]
Page 166
 𝑘𝑔 𝑘𝑔
kg/m3, air = 1.1865 kg/m3, = 0.00001826 𝑚 = ~0.000987
𝑠
= 85.3
𝑑𝑎𝑦
(32)
kg/m.s, hfg = 2,454 kJ/kg and Dv = 2.54 10-5
The Rate of heat transfer [q] is,
m2/s. The Reynolds number & Sherwood
𝑘𝑔 𝑘𝐽
number is estimated as, 𝑞 = 𝑚ℎ𝑓𝑔 = 0.000987
𝑠
× 2,454
𝑘𝑔
= 2,422𝑊 (33)
10×1.1865×10
𝑅𝑒 = 0.00001826
= ~6,497,679 (28) References & Further Reading
4⁄ 1⁄ 1. “Heat Transfer”, 10th Ed, Holman JP.
𝑆ℎ = 0.037 × 6,497,679 5 × 0.6058 3 = 8,828 (29)
2. https://www.fws.gov/ecological-
The convective mass transfer coefficient is,
services/energy-development/oil-gas.html
8,828×2.54×10−5
h𝑚 = 10
= 0.0224 𝑚/𝑠 (30) 3. 2005 ASHRAE Handbook - Fundamentals
The rate of evaporation [m] is, (SI), Chapter 5

𝑚 = 0.022426 × 20 × [0.01601 − 0.013822] (31)

Appendix A

Page 167
Appendix B

Page 168
 Module 27
Exploring LPG Cylinders for Medical Oxygen – A Preliminary Study
The following module is a study to explore Pcyl = Cylinder Pressure [bara]
the usage of LPG cylinders for medical oxygen Patm = Atmospheric Pressure [bara]
in times of medical emergencies. The study k = ratio of specific heats [Cp/Cv] [-]
aims at understanding how long medical The blowdown time can be estimated as,
oxygen can be supplied to cater to patients 𝑡
requiring supply between 0.5 lit/min to 2 𝑃𝑐𝑦𝑙 = 𝑃0 𝑒𝑥𝑝 [− ] (B)
lit/min.
𝑉𝑐𝑦𝑙
General Notes & Assumptions 𝜏=[ 𝑘+1 1 ] (C)
2 2(𝑘−1) 𝑘×𝑅×𝑇𝑐𝑦𝑙 ⁄2
1. Medical Oxygen composition is taken to 𝐶𝑑 ×𝐴×[
𝑘+1
] ×[
𝑀𝑊
]

contain 90% O2, 5% N2 and 5% Ar. Where,

2. The LPG cylinder considered has a 33.3 = Discharge Time Constant [sec]
litre water capacity, storing 14.2 kg of LPG. Pcyl = Cylinder Pressure [bara]
3. The analysis is performed as a vessel with P0 = Cylinder Initial Pressure [bara]
an orifice discharging the fluid to the
Design Data & Results
downstream and considering patients
The input data and results for 0.5 lit/min is as
requiring oxygen in the range of 0.5 lit/min
follows,
and 2 lit/min. The orifice discharge
Table 1. Input Data and Results for 0.5 lit/min
coefficient [Cd] is taken as 0.62.
Parameter Value Unit
4. Considering a cylinder pressure cap of 16.9
kg/cm2, the pressure cap for the study is Effective Cylinder Volume [V] 0.0333 m3
taken as 16.0 bara at 250C. The pressure at Medical Oxygen MW 32.2 kg/kmol
which medical oxygen is delivered is taken Initial Pressure [P0] 16 bara
as 1.01325 bara.
Initial Temperature [T0] 25 0C

5. For the analysis, an isothermal blowdown

Oxygen k [Cp/Cv] 1.395 -
condition is taken assuming the breathing
Choked Flow Exists or Not Choked Flow
process from the medical oxygen cylinder
takes sufficiently long time and the gas Compressibility Factor [Z] 0.9902 -

temperature also does not change with Oxygen Density [r] 20.99 kg/m3
time. Hence heat is absorbed through the Mass of O2 in Cylinder [m] 0.699 kg
walls such that the cylinder temperature is
Orifice Throat Diameter [d] 0.30 mm
close to ambient temperature.
Orifice Throat CSA [A] 7.0926E-08 m2
Governing Relationships
Discharge Coefficient [Cd] 0.62 -
To estimate the blow down time, a transient
Speed of Sound [C] 327.7 m/s
study is performed. To check if choked flow
exists, the following condition is applied, Discharge Time Constant [t] 3,991 sec
𝑘 Mass Flow Rate [mg] 0.0002 kg/s
𝑃𝑐𝑦𝑙 𝑘+1 𝑘−1
≥ [ ] (A) Volumetric Flow Rate 0.000008 m3/s
𝑃𝑎𝑡𝑚 2
Where, Required Vol. Flow Rate 0.50 lit/min

Page 169
The input data and results for 2 lit/min is, Oxygen-Demand-Tata-Consulting-
Table 2. Input Data and Results for 2 lit/min Engineers-Response.pdf)
Parameter Value Unit 3. “Tank Blowdown Math”, Dean Wheeler,
Brigham Young University, March 13, 2019
Orifice Throat Diameter [d] 0.60 mm
(https://www.et.byu.edu/~wheeler/Tank_
Orifice Throat CSA [A] 2.8334E-07 m2
Blowdown_Math.pdf)
Discharge Time Constant [ t] 999 sec
Appendix A: Derivations of Expressions
Mass Flow Rate [mg] 0.0007 kg/s 𝑑𝑚
= −𝑚𝑔 (1)
Volumetric Flow Rate 0.000033 m3/s 𝑑𝑡

Required Vol. Flow Rate 2.00 lit/min ‘m’ = mass of gas in the cylinder, expressed as,
𝑚 = 𝜌𝑐𝑦𝑙 × 𝑉𝑐𝑦𝑙 (2)
Plotting a graph between cylinder pressure
and Time for both cases of 0.5 lit/min and 2 Where,
lit/min, cyl = Density of gas in cylinder [kg/m3]
Vcyl = Volume of gas in cylinder [m3]
Whereas, mg is mass flow rate at choked flow
conditions is expressed as,
𝑚𝑔 = 𝐶𝑑 × 𝐴 × 𝜌𝑐 × 𝑣𝑐 (3)
Where,
Cd = Orifice Discharge Coefficient [-]
A = Orifice Cross-sectional Area [m2]
c = Density at choked flow at throat [kg/m3]
vc = Speed of Sound [m/s]
Figure 1. Cylinder Pressure vs. Time
The speed of sound can be estimated as,
From the above figure, the cylinder pressure 1⁄
beginning from an initial pressure of 16 bara 𝑘×𝑅×𝑇𝑐 2
𝑣𝑐 = [ ] (4)
𝑀𝑊
reaches 1 atm in about 3 hours for a
discharge rate of 0.5 lit/min and about 45 Where,
min for the case of 2 lit/min. MW = Fluid Molecular Weight [kg/kmol]
References & Further Reading Tc = Temperature at choked conditions [K]
For a reversible adiabatic expansion, the fluid
1. “https://www.gasmartindia.com/news/w
density at the orifice throat can be related to
hat-is-the-composition-of-medical-
the fluid density in the cylinder as,
oxygen/#:~:text=Medical%20oxygen%20c
1
omprises%20of%20minimum,removed%2 𝜌𝑐 2 𝑘−1
=[ ] (5)
0leaving%20behind%20only%20oxygen. 𝜌𝑐𝑦𝑙 𝑘+1
2. Emergency Options for Medical Oxygen 𝑘−1
𝑇𝑐 𝜌𝑐
Storage & Alternate Mode of Oxygen =[ ] (6)
𝑇𝑐𝑦𝑙 𝜌𝑐𝑦𝑙
Generation, Preliminary Assessment
2
Report, Tata Consulting Engineers Limited, Or, 𝑇𝑐 = 𝑇𝑐𝑦𝑙 × [𝑘+1] (7)
April 2021 (https://www.tce.co.in/wp- Therefore the speed of sound at cylinder
content/uploads/2021/04/Meeting- conditions can be expressed as,

Page 170
 2×𝑘×𝑅×𝑇𝑐𝑦𝑙
1⁄
2 Applying ideal gas law to convert densities to
𝑣𝑐 = [ ] (8) pressures,
𝑀𝑊×(𝑘+1)
𝑡
Therefore the mass flow rate at choked flow 𝑃𝑐𝑦𝑙 = 𝑃0 𝑒𝑥𝑝 [− ] (19)
related to cylinder process conditions
Appendix B: MS Excel Calculations
becomes,
1 1⁄
2 𝑘−1 2×𝑘×𝑅×𝑇𝑐𝑦𝑙 2
𝑚𝑔 = 𝐶𝑑 × 𝐴 × 𝜌𝑐𝑦𝑙 [ ] ×[ ] (9)
𝑘+1 𝑀𝑊×(𝑘+1)

Or
𝑘+1 1⁄
2 𝑘×𝑅×𝑇𝑐𝑦𝑙 2
𝑚𝑔 = 𝐶𝑑 × 𝐴 × 𝜌𝑐𝑦𝑙 [𝑘+1]2(𝑘−1) ×[ 𝑀𝑊
] (10)

Where,
R = Gas Constant [8.314 m3.bar/kmol.K]
The cross-sectional area of the orifice is,
2𝜋
𝐴 = 4 × 𝑑𝑜𝑟𝑖𝑓𝑖𝑐𝑒 (11)

Where,
dorifice = Orifice diameter [m]
Therefore solving for blowdown time,
𝑘+1
𝑑𝜌𝑐𝑦𝑙 2 2(𝑘−1)
𝑉𝑐𝑦𝑙 × = 𝐶𝑑 × 𝐴 × 𝜌𝑐𝑦𝑙 [ ] ×
𝑑𝑡 𝑘+1
1
𝑘×𝑅×𝑇𝑐𝑦𝑙 ⁄2
[ ] (12)
𝑀𝑊
Rearranging the above,

𝑉𝑐𝑦𝑙 𝑑𝜌𝑐𝑦𝑙
−[ 𝑘+1 1 ] = 𝜌𝑐𝑦𝑙 (13)
2 2(𝑘−1) 𝑘×𝑅×𝑇𝑐𝑦𝑙 ⁄2 𝑑𝑡
𝐶𝑑 ×𝐴×[ ] ×[ ]
𝑘+1 𝑀𝑊

Simplifying the expression by taking a

discharge time constant [ ],

𝑉𝑐𝑦𝑙
𝜏=[ 𝑘+1 1 ] (14)
2 2(𝑘−1) 𝑘×𝑅×𝑇𝑐𝑦𝑙 ⁄2
𝐶𝑑 ×𝐴×[ ] ×[ ]
𝑘+1 𝑀𝑊

Substituting and solving for the blowdown

time,
𝑑𝜌𝑐𝑦𝑙 𝜌𝑐𝑦𝑙
=− (15)
𝑑𝑡
𝑐𝑦𝑙
𝑑𝜌𝑐𝑦𝑙 𝑡=𝑡 𝑑𝑡
∫ = − ∫𝑡=0 (16)
0 𝜌𝑐𝑦𝑙

𝑐𝑦𝑙 𝑡
𝑙𝑛 [ ]=− (17)
0
𝑡
𝑐𝑦𝑙
= 0
𝑒𝑥𝑝 [− ] (18)

Page 171
 Module 28
Heating Value Estimation for Natural Gas Applications
For natural gas custody transfer applications, Estimating Fuel Calorific Values
the gross calorific or gross heating value is The Gross calorific value (GCV) in mass terms
necessary for both the buyer and seller to can be computed as per ISO 6976:1995 at
estimate the sales price of natural gas. In case 1.01325 bara & 15 0C (referred to as Standard
of fuel suppliers, heat content is expressed in Conditions as per ISO 6976), as follows,
terms of Higher Heating value (HHV) to ∑ 𝑋𝑖 ×𝐺𝐶𝑉𝑖[𝑚𝑜𝑙]
estimate fuel charges in kWh. Whereas Lower 𝐺𝐶𝑉𝑚𝑎𝑠𝑠 = ∑ 𝑋𝑖 ×𝑀𝑖
(1)
Heating Value (LHV) is employed to estimate
Where,
fuel requirements since the total energy input
for a specific power output is already fixed. Xi = Molar fraction of component ‘i’
To understand how fuel heating values are GCVi[mol] = Molar Gross calorific value of
affected, LHV and HHV is explained as, component ‘i', [kJ/mol]
1. The lower heating value (LHV) or net Mi = Molecular mass of component ‘i’ [g/mol]
calorific value (NCV) of a fuel is defined as However in this module, the standard
the amount of heat released by combusting temperature is taken as 250C and the heating
a specified quantity at 25°C and returning values are estimated based on a
the temperature of combustion products to stoichiometric balance instead of ISO
150°C, with the assumption that latent 6976:1995 method. Taking the following
heat of vaporization of water in the natural gas composition,
reaction products is not recovered, i.e.,
Table 1. Natural Gas Composition
when water in the combustion product, is
Component MW Mol%
in its vapour form, it is called LHV/NCV.
LHV is a better indication of a fuel’s useful - kg/kmol %
heat since the combustion products are Methane [CH4] 16.043 85.0
above the boiling point of water. Ethane [C2H6] 30.070 5.0
2. The higher heating value or gross calorific Propane [C3H8] 44.097 3.0
value (GCV) of a fuel is defined as the
n-Butane [n-C4H10] 58.123 1.0
amount of heat released by combusting a
i-Butane [i-C4H10] 58.123 1.0
specified quantity at 25°C and the products
have returned to a temperature of 25°C, n-Pentane [n-C5H12] 72.15 0.5

taking into account the latent heat of i-Pentane [i-C5H12] 72.15 0.3
vaporization of water in the products. i.e., Hydrogen [H2] 2.016 0.1
when water in the combustion product, is Carbon monoxide [CO] 28.011 0.1
in its liquid form, it is called HHV/GCV.
Carbon dioxide [CO2] 44.011 0.2
Presence of water is detrimental to a fuel’s
Nitrogen [N2] 28.0135 3.8
heating value, since with high combustion
Mixture MW [kg/kmol] 19.385
temperatures, water turns into steam & eats
away a portion of the energy released as To estimate the heat of reaction, the
latent heat of vaporization [ HV], i.e., HHV combustion of hydrocarbons follow the below
includes latent heat of vaporization of water. stoichiometric balance as,
Page 172
 3𝑛 + 1 i-Pentane [i-C5H12] -154.40
𝐶𝑛 𝐻2𝑛+2 + [ ] 𝑂2 → 𝑛𝐶𝑂2 + [𝑛 + 1]𝐻2 𝑂
2
Carbon monoxide [CO] -110.52
To estimate the natural gas calorific value, the
Hydrogen [H2] 0.00
calculations require,
Oxygen [O2] 0.00
1. Natural gas Composition
Nitrogen [N2] 0.00
2. Heat of formation [ H0f] at Ref. Conditions Carbon dioxide [CO2] -393.51
3. Heat of Reaction per mole computed as, Water Vapour [H2O (g)] -241.83
0 0 0 Water [H2O (l)] -285.84
𝛥𝐻𝑅𝑥𝑛,25℃ = ∑ 𝑋𝑖,𝑃 ∆𝐻𝑓,𝑃 − ∑ 𝑋𝑖,𝑅 ∆𝐻𝑓,𝑅 (1)

4. Based on per mole heat of reaction for each Performing a stoichiometric balance of the
combustible species/component, the combustion reactions,
individual mole fraction of each Table 3. Combustion Reaction Set
component is multiplied with the
CH4 + 2 O2 → 1 CO2 + 2 H2O(g)
respective heat of reaction and summed up
C2H6 + 3½ O2 → 2 CO2 + 3 H2O(g)
to arrive at the Net Calorific Value (NCV) or
Lower Heating Value (LHV). A negative C3H8 + 5 O2 → 3 CO2 + 4 H2O(g)
sign in the LHV/NCV value indicates, heat nC4H10 + 6½ O2 → 4 CO2 + 5 H2O(g)
is released due to the combustion process.
iC4H10 + 6½ O2 → 4 CO2 + 5 H2O(g)
5. To estimate the HHV, firstly, the hydrogen
nC5H12 + 8 O2 → 5 CO2 + 6 H2O(g)
content in each component of the natural
gas mixture (which forms water) is iC5H12 + 8 O2 → 5 CO2 + 6 H2O(g)
multiplied by its respective component’s H2 + ½ O2 → 1 H2O(g)
mole fraction, summed up, divided by 2
CO + ½ O2 → 1 CO2
and multiplied with the heat of reaction
H2O(g) → 1 H2O(l)
from the conversion of H20(g) H20(l),
i.e., [ HRxn,250C = H0f,H2O(l) - H0f,H2O(g)]. Quoting a sample case for methane at 250C,
Followed by, subtracting the above 0
∑ 𝑋𝑖,𝑅 ∆𝐻𝑓,𝑅 = [(1 × −74.84) + (2 × 0)] (2)
estimate from the modulus (positive) value
of LHV/NCV. 0
∑ 𝑋𝑖,𝑃 ∆𝐻𝑓,𝑃 = [(2 × −241.83) + (1 × −393.51)] (3)

The standard heat of formation [ H0f] at 250C 0

𝛥𝐻𝑅𝑥𝑛 = [−877.162] − [−74.84] = −802.32 𝑚𝑜𝑙 (4)
𝑘𝐽

for the reactants and products are as follows,

For the case of H2 at 250C,
Table 2. Standard Heat of Formation [250C]
0
∑ 𝑋𝑖,𝑅 ∆𝐻𝑓,𝑅 = [(1 × 0) + (0.5 × 0)] (5)
Component H0f,250C
0
- kJ/mol ∑ 𝑋𝑖,𝑃 ∆𝐻𝑓,𝑃 = [1 × −241.83] (6)
𝑘𝐽
Methane [CH4] -74.84 0
𝛥𝐻𝑅𝑥𝑛 = [−241.83 − 0] = −241.83 𝑚𝑜𝑙 (7)
Ethane [C2H6] -84.67
For the case of H2O(g) to H2O(l) at 250C,
Propane [C3H8] -103.85
0
∑ 𝑋𝑖,𝑅 ∆𝐻𝑓,𝑅 = −241.83 (8)
n-Butane [n-C4H10] -124.73
0
i-Butane [i-C4H10] -134.50 ∑ 𝑋𝑖,𝑃 ∆𝐻𝑓,𝑃 = −285.84 (9)
𝑘𝐽
n-Pentane [n-C5H12] -146.40 0
𝛥𝐻𝑅𝑥𝑛 = [−285.84 − (−241.83)] = −44.01 (10)
𝑚𝑜𝑙

Page 173
Similarly, performing calculations for other n-Butane [n-C4H10] 0.010 10 0.1
components to yields, i-Butane [i-C4H10] 0.010 10 0.1
Table 4. Heat of Reaction Summary n-Pentane [n-C5H12] 0.06
0.005 12
Component ΔH0Rxn, 250C [kJ/mol] i-Pentane [i-C5H12] 0.003 12 0.036
CH4 -802.32 Hydrogen [H2] 0.0010 2 0.002
C2H6 -1,427.83 Total 4.238
C3H8 -2,043.98 The HHV/GCV is now calculated as,
nC4H10 -2,658.44 𝐻𝐻𝑉 = 894.43 − [
4.238×−44.014
]
𝑘𝐽
= 987.70 𝑚𝑜𝑙 (11)
2
iC4H10 -2,648.67
Expressing in mass terms [kJ/kg], the heating
nC5H12 -3,272.11 values are as follows,
iC5H12 -3,264.11 894.43×1000
𝐿𝐻𝑉/𝑁𝐶𝑉 = 19.385
= 46,140 𝑘𝐽/𝑘𝑔 (12)
H2 -241.83 987.7×1000
𝐻𝐻𝑉/𝐺𝐶𝑉 = 19.385
= 50,951𝑘𝐽/𝑘𝑔 (13)
CO -282.99
References & Further Reading
H2O(g) -44.014
1. “Principles of Chemical Engineering
Therefore the LHV/NCV is computed as, Processes”, Nayef Ghasem, Redhouane
Table 5. LHV/NCV Estimation Henda, 2nd Edition, Taylor & Francis Group
Component 𝑿𝒊 ∆𝑯𝟎𝑹 2. ISO 6976 (1995) Natural gas - Calculation
- [kJ/mol] of calorific values, density, relative density
Methane [CH4] 0.85 -802.32 -681.974
and Wobbe index from composition

Ethane [C2H6] 0.050 -1,427.83 -71.392 3. https://www.clarke-energy.com/heating-

value/
Propane [C3H8] 0.030 -2,043.98 -61.320
4. https://www.industrialheating.com/articl
n-Butane [n-C4H10] 0.010 -2,658.44 -26.584
es/90561-calculating-the-heat-of-
i-Butane [i-C4H10] 0.010 -2,648.67 -26.487
combustion-for-natural-gas
n-Pentane [n-C5H12] 0.005 -3,272.11 -16.361

i-Pentane [i-C5H12] 0.003 -3,264.11 -9.792

Hydrogen [H2] 0.001 -241.83 -0.242
Carbon monoxide 0.001 -282.99 -0.283
Carbon dioxide 0.000 0.000
Nitrogen [N2] 0.000 0.000

𝑳𝒐𝒘𝒆𝒓 𝑯𝒆𝒂𝒕𝒊𝒏𝒈 𝑽𝒂𝒍𝒖𝒆 (𝑳𝑯𝑽) = ∑ 𝑿𝒊 ∆𝑯𝟎𝑹 -894.43

The higher heating value (HHV) is computed

by initially making a hydrogen balance as,
Component Hydrogen Balance

Methane [CH4] 0.850 4 3.4

Ethane [C2H6] 0.050 6 0.3
Propane [C3H8] 0.030 8 0.24
Page 174
Appendix A

Page 175
 Module 29
Empirical Approach to Hydrate Formation in Natural Gas Pipelines
Natural Gas Pipelines often suffer from transfer coefficient is taken to be 25 W/m2.K.
production losses due to hydrate plugging. The ambient temperature is 120C. The
For an effective hydrate plug to form, factors hydrate formation temperature for the
can vary from pipeline operating pressure composition is experimentally estimated to
and temperature, presence of water below its be 500F at 325 psia. It is required to estimate
dew point, extreme winter conditions & Joule the pipeline exit temperature & the hydrate
Thomson cooling. In the event hydrates form formation temperature along the pipeline. For
in the pipeline section, their consequence the estimates, the Joule-Thomson coefficient
depends on how well the hydrates is assumed to be an average of 5.60C/bar
agglomerate to grow and form a column. If throughout the pipeline. The natural gas
the pipeline section temperature is only at composition is as follows,
par with the hydrate formation temperature, Table 1. Gas Mixture [GPSA, Sec 20, Page 20-15]
the particles do no agglomerate; instead they
Mol% MW [Mi] yiMi
have to cross the metastable region which is Component
of the order of 50C to 60C, before hydrate [%] [kg/kmol] [-]
formation accelerates to block the pipeline. Methane 78.40 16.04 12.58

Ethane 6.00 30.07 1.80

Propane 3.60 44.01 1.58

i-Butane 0.50 58.12 0.29

n-Butane 1.90 58.12 1.10

CO2 0.20 44.01 0.09

N2 9.40 28.01 2.63

Total 100.00 MW [kg/kmol] 20.08

Figure 1. P-T Hydrate Curve [1] Methodology

Although engineering softwares exist to The pipeline temperature profile can be
estimate pipeline process conditions and also estimated based on Coulter & Bardon (1979)
generate a P-T hydrate curve, the following correlation [4]. The steady state temperature
tutorial provides a guidance summary to profile is calculated from the momentum
estimate the expected pipeline temperature equation, while omitting the potential &
profile and the associated hydrate formation kinetic energy terms in the enthalpy equation.
temperatures. 𝑑ℎ 𝑑𝑄
+ =0 (1)
𝑑𝐿 𝑑𝐿
Problem Statement
Where,
A DN 14”, 20 km hydrocarbon line carrying
𝜋×𝑂𝐷×𝑈×∆𝐿
natural gas at the rate of 85,000 kg/h, 40 bara 𝑄= 𝑚
[𝑇0 − 𝑇𝑠 ] (2)
and 250C is fed to a receiving station. The 𝑑ℎ = 𝑐𝑝 𝑑𝑇 − 𝜇𝑐𝑝 𝑑𝑃 (3)
total pipeline pressure drop per km [ P/km]
Where,
is taken to be 1 bar/km. The overall heat

Page 176
U = Overall HTC [W/m2.K] Results
ID = Pipeline OD [m] Substituting the values to arrive at the
pipeline temperature profile, the gas specific
m = mass flow rate [kg/s]
gravity is estimated as,
L = Pipeline length [m] 20.08
𝛾𝑔 = 28.9625 = 0.6933 (7)
T0 = Fluid Temperature [K]
14×25.4
Ts = Surrounding Temperature [K] 𝜋×[
1000
]×25
𝑎= 85,000 = 0.0005711 (8)
[ ]×2.071×1000
= Joule-Thompson Coefficient [0C/bar] 3600

Cp = Specific heat capacity [J/kg.K] 𝑇[𝐿] = 12.0195 × 𝑒 −0.0005711×𝐿 + 286.1305 (9)

g = Gas Specific Gravity, MW/28.9625 [-] The hydrate formation temperature [Th] is,
Solving for pipeline temperature profile, 𝑇ℎ [℉] = [14.0835 × 𝑙𝑛(𝑃, 𝑝𝑠𝑖)] − 32.9023 (10)
𝜇 𝑑𝑃 𝜇 𝑑𝑃 Plotting the above expressions, we get,
𝑇[𝐿] = [𝑇0 − 𝑇𝑠 − (𝑎 ) ( 𝑑𝐿 )] 𝑒 −𝑎𝐿 + 𝑇𝑠 + (𝑎 ) ( 𝑑𝐿 ) (4)

Where,
𝜋×𝑂𝐷×𝑈
𝑎= 𝑚×𝐶𝑝

It is to be noted that the specific heat [Cp] and

Joule-Thompson [J-T] co-efficient [ ] varies
with the pipeline pressure & temperature.
But for computational purposes, is assumed
to be constant. The purpose of including the J-
T coefficient is to account for cooling during
gas expansion along the pipeline. The ideal
mass specific heat [Cp], kJ/kg.K, of natural gas Figure 2. Hydrate Formation Temperature
can be computed as, From the plot, the pipeline temperature stays
𝐶𝑝 = [(−10.9602𝛾𝑔 + 25.9033) + (0.21517𝛾𝑔 − above the hydrate formation temperature. In
0.068687)𝑇 + (−0.00013337𝛾𝑔 ) + 0.000086387)𝑇 2 + practice, to increase the difference, the inlet
(0.000000031474𝛾𝑔 ) − 0.000000028396)𝑇 3 ]/ 𝑀𝑊(5) gas can be either heated or hydrate inhibitors
Where, T = Temperature [K] such as MeOH, MEG or TEG can be added.
References & Further Reading
Hydrate Formation Temperature
1. https://www.sciencedirect.com/topics/en
To estimate the hydrate formation
gineering/hydrate-formation-curve
temperature [Th], Towler & Mokhatab (2005)
2. “Handbook of Natural gas Transmission and
[3], proposed the following correlation,
Processing”, Saied Mokhatab, William A.
𝑇ℎ [℉] = [13.47 × 𝑙𝑛(𝑃)] + [34.27 × 𝑙𝑛(𝛾)] − Poe, John Y. Mak, 3rd Edition.
[1.675 × 𝑙𝑛(𝑃) × 𝑙𝑛(𝛾)] − 20.35 (6) 3. “Hydrate Formation Calculation in the
Where, Natural Gas Purification Unit”, J A Prajaka
P = Pressure [psia] et al 2019 IOP Conf. Ser.: Mater. Sci. Eng.
543 012084
The validity of the above expression is for the
4. Predicting Compositional Two Phase Flow
1. Temperature Range: 260 K to 298 K
Behaviour in Pipelines, H. Furukawa, O.
2. Pressure Range: 1200 kPa to 40,000 kPa Shoham, J.P. Brill, Transaction of the ASME,
3. MW: 16 g/mol to 29 g/mol (0.55 < g < 1.0) Vol 108, September 1986.

Page 177
Annexure: MS-Excel Spreadsheet

Page 178
 MODULE 30
METHODOLOGY FOR SLUG CATCHER SIZING
Oil & Gas Pipelines are often subjected to an operation called ‘Pigging’ for maintenance purposes
(For e.g., cleaning the pipeline of accumulated liquids or waxes). A pig is launched from a pig
launcher that scrapes out the remnant contents of the pipeline into a vessel known as a ‘Slug
catcher’. The term slug catcher is used since pigging operations produces a Slug flow regime
characterized by the alternating columns of liquids & gases. Slug catcher’s are popularly of two
types – Horizontal Vessel Type & Finger Type Slug catcher. However irrespective of the type
used, the determination of the slug catcher volume becomes the primary step before choosing
the slug catcher type. In order to estimate the pigging volume, engineers use various Flow
Assurance (FA) tools to estimate & plot a graph between ‘Pipeline Volumetric Flow vs Time’. Here
pipeline volumetric flow rate refers to the point at which liquid exits prior to entering into the
slug catcher.

METHODLOGY

The methodology that could be adopted to estimate the excess space required in the slug catcher
volume is based on measuring the volumetric flow rate of the fluid that exits the pipeline from
the time a ‘pig’ is launched into the pipeline and measuring the same parameter until the pig
exits the pipeline. The point at which the pipeline is routed to the slug catcher is where using an
FA tool; a transient study is made to plot a graph between Volumetric flow rate vs. Time. Since the
slug catcher would have a provision to drain the accumulated liquids, it is taken as Drain Rate
(QD) which also decides the effective volume of the slug catcher.

Figure 1. Pig Traversing in a Pipeline

Using the Volumetric Flow Rate vs Time graph, the vessel space required to accommodate the
excess liquid from the pigging operation for a drain rate (QD) can be computed as,

Where,

Page 179
The purpose of introducing a factor of ‘margin’ in estimating the Slug catcher volume is due to
uncertainties that Flow Assurance Tools can produce as well as the users’ convergence criteria in
arriving at a well converged volumetric flow rate vs time graph. This can be anywhere from 2%
to 30% and is incumbent upon the Flow Assurance Engineer to thoroughly understanding the
principles of FA & the respective FA solver prior to carrying out slug catcher volume calculations.

Following on the equations provided, a tabulation of the volumetric flow rate vs time can be
made in MS-Excel as,

Volumetric Flow
Time [sec] Pigging Volume [m3]
Rate [m3/s]

ti Qi

ti+1 Qi+1

ti+2 Qi+2

ti+3 Qi+3

.. .. ….

Cumulative Slug/Pigging Volume

Below is an example graph of pigging volume estimated using the above method for a given
drain rate (QD). The cumulative slug/pig volume computed is the area represented under the
total liquid volume flow into the slug catcher curve (‘blue’ line). The ‘red’ line represents the
total liquid volumetric flow that enters the slug catcher while the vessel is constantly drained.

Figure 2. Example of Slug Catcher Volume Required (Illustration Purposes Only)

Page 180
CHOOSING A SLUG CATCHER SIZE

The drain rate of the slug catcher is determined by the downstream liquid processing capacity in
an Oil & gas facility. Therefore in Brownfield projects, the drain rate cannot be inadvertently
increased beyond its design limit in the event when the pigged volume exceeds the existing slug
catcher’s capacity (For e.g., due to change of production). In such instances, it is prudent to
install an additional slug catcher to cater to the additional pigging volume. In Greenfield projects,
the Basis of Design (BOD) becomes the primary document that determines the drain rate of the
slug catcher. Therefore the excess volume required in the slug catcher above the normal
operating liquid level (NLL) can be determined by plotting a Slug Catcher Volume Vs. Drain Rate
Graph. Below is a representative plot for illustration purposes.

Figure 3. Example of Slug Catcher Volume Selected Based on Drain Rate (Illustration Purposes Only)

SUMMARY
The intent of the slug catcher operation is to accommodate the excess liquids generated by the
pigging operation but not to alter the drain rate that can affect the downstream liquid handling
equipment. This is accomplished by providing space for the excess pigging volume capacity over
and above the Normal Liquid Level (NLL) but below High Liquid Level (HLL) at which liquid
levels are maintained in the slug catcher.

Page 181
 About the Author
Vijay Sarathy is a Process Engineer with professional expertise
spanning over 14 years of in the Upstream Oil & Gas industry. He
started his career with GE performing CFD analysis for Gas Turbine
Ventilation systems for GE-Nuovo Pignone’s PGT series for ATEX
certification and subsequently in areas of Front End Engineering
Design, Process Dynamic Simulation including Turbomachinery,
Onshore Pipeline Flow Assurance, Optimization & Cost benefit
studies for Gas-Oil Separation Units.
Vijay holds a Master’s Degree in Chemical Engineering from Birla Institute of Technology
& Science (BITS), Pilani, India and is a Chartered Chemical Engineer from the Institution
of Chemical Engineers (IChemE), UK. He is also an Industry Advisory Board Member for
the International Association of Certified Practicing Engineers (IACPE, Texas, USA) as
well as a Contributing Author to the Engineering Practice Magazine. Vijay has worked as
an Upstream Process Engineer with major conglomerates of General Electric, ENI
Saipem and Shell.

Front Cover Photo: Hans Isaacson, www.unsplash.com