Module 5

Calibration of gas flow

meters

Calibration models, uncertainty of

calculation

Peter Lucas
 Contents

• (recapture of) Definitions

• Calibration of a flow meter
− Method
− Calibration curve
− Uncertainty

Page 2 GFM2012 Module 5A

 Definitions (1)

• Calibration: set of operations that establish, under specified

conditions, the relationship between values of quantities
indicated by a measuring instrument and the corresponding
values realized by standards or references of established
uncertainty
• Uncertainty (of measurement): parameter, associated with the
result of a measurement, that characterizes the dispersion of the
values that could reasonably be attributed to the measurand
• Purpose of calibration: to reduce the measurement error and
uncertainty as far as possible for the meter in its final application
• Calibration is determination of
 the error or deviation w.r.t. the reference value
 the uncertainty of this value
• Readjust: update calibration curve of investigated flow meter

Page 3 GFM2012 Module 5A

 Definitions (2)

• The output (error(%), meter factor, impulse factor) is given

as a function of the flow rate (Q)
• Error = Qindicated – Qref
• Error(%) =(Qindicated - Qref)/Qref *100%
• Meter factor = Qref /Qindicated
• Calibration is typically performed at laboratory conditions
(approaching steady and fully developed flow)
• Meter is typically often not used at laboratory conditions
• Important to understand the effect of the operating
conditions on the meter performance, e.g.
− Pressure
− Temperature
− Installation effects (velocity profile, interaction with
standard, pulsation, … )

Page 4 GFM2012 Module 5A

 Contents

• (recapture of) Definitions

• Calibration of a flow meter
− Method
− Calibration curve
− Uncertainty

Page 5 GFM2012 Module 5A

 Calibration method
Selecting a calibration method

• Time of test must be long enough to establish stable conditions

temperature stabilizes after roughly 1 minute
• Time must be long enough such that time uncertainty is negligible
1 ms uncertainty, 10 seconds sufficient
• Time must be sufficient to have sufficient pulses
1 pulse uncertainty for stopping and starting, 10000 pulses
sufficient
• Response time of meter must be suitable to follow changes
introduced by the standard
• Uncertainty standard must be three times better than required
uncertainty

Page 6 GFM2012 Module 5A

Calibration method
Possibilities

Gas
• Directly against standard (typically volume)
• Master method method

Liquid
• Directly against standard (volume or mass)
• (compact) Prover
• Master method method

Page 7 GFM2012 Module 5A

 Calibration method
Master meter method

pref , T ref ref pmut , Tmut mut

in out
Reference Meter Meter-under-Test

• Traceable reference flow meter

− calibrated with acceptable uncertainty
− reliable calibration history
− characterized for performance across a range of fluids and process conditions
• Stationary gas-flow conditions: kg “in” = kg “out”
• Density may differ for the reference meter and meter under test
(pressure drop, temperature increase, … )
• Correction required for volumetric calibrations

Page 8 GFM2012 Module 5A

Master meter method
Correcting for operating conditions (1)

Stationary conditions, constant mass flow rate: mut Qmut ref Qref

The equation of state of real gas

p
ZRsT
Indicated flow rate and flow rate according to master meter

N ref N mut
Qref 3600 , Qmut 3600
I ref 1 I mut 2

Q = volume flow rate [m3/s] N = number of pulses [-]

= density [kg/m3] I = impulse factor [m–3]
p = pressure [N/m2] = measuring time [s]
Rs = specific gas constant
T = temperature = [K]
Z = compressibility [-]

Page 9 GFM2012 Module 5A

 Master meter method
Correcting for operating conditions (2)

Qind Q Qmutind Qmut Qmutind

emut 1
Q Qmut Qmut

ref
mut Qmut ref Qref Qmut Qref

eref
Qref ind
Qref
Qref
Qref
mut

Qref ind
e ref 1
} Qmut ref

mut
Qref ind
e ref 1

Qmutind mut
Qmutind
emut 1 e ref 1 1
ref
Qref ind ref Qref ind
mut e ref 1

Page 10 GFM2012 Module 5A

Master meter method
Correcting for operating conditions (3)
compressibility flow rate according to standard
pressure time
B pm Z s N m I s τ s T0 t s
emut 1 e ref 1
B p s Z m I m N s τ m T0 t m
deviation standard
flow rate according
temperature
to MUT
B = atmospheric pressure [bar]
p = gauge pressure [bar]
Z = real gas constant or compressibility factor [-]
N = number of counted pulses [-]
I = impulse factor of a gas-meter [m-3]
τ = time [s]
T = absolute temperature [K]
eref = relative deviation of the Reference Meter [-]
em = relative deviation of the Meter under Test [-]
Page 11 GFM2012 Module 5A
 Contents

• (recapture of) Definitions

• Calibration of a flow meter
− Method
− Calibration curve
− Uncertainty

Page 12 GFM2012 Module 5A

 Error curve turbine meter
Example and typical shape

Error = f(flow rate)

0,40
Reynolds effect nominal usage

0,20 specs from

manufacture
0,00
0 200 400 600 800 1000 1200

-0,20
Error (%)

-0,40
- Qmin minimal flow rate
Measuring data

- Qn nominal flow rate

-0,60 - Qmax maximal flow rate
Flow Rate (m3/h)

Page 13 GFM2012 Module 5A

 Error curve turbine meter
Fitting (least squares)

Error = f(flow rate)

0,40

0,20

0,00
0 200 400 600 800 1000 1200

-0,20
Error (%)

Measuring data

-0,40 Straatsma FIT

Polynomial 3rd
order

-0,60
Flow Rate (m3/h)

Page 14 GFM2012 Module 5A

Straatsma polynomial

Mathematical description of the error curve based on the

theory of turbine meter
Straatsma polynomial:
0.2 0.33 2
Q Q Q
e a0 a1 a2 a3
Qmax Qmax Qmax

a0 ... a1 constants
a1 influence of drag forces on the rotor blades
a2 effect of the boundary layer in the meter annulus
a3 effect of friction in the meter bearing

Page 15 GFM2012 Module 5A

 Contents

• (recapture of) Definitions

• Calibration of a flow meter
− Method
− Calibration curve
− Uncertainty

Page 16 GFM2012 Module 5A

 Uncertainty sources
Standard

1. Drift, e.g. 0,02% per year

2. (unsteady) Environmental conditions (barometric pressure,
temperature, humidity, radiation)
3. Hysteresis (influence of previous reading or measurement)
4. Uncertainty in reading (half the value of the smallest digit)
5. Repeatability and reproducibility
6. Unsteady operating conditions
7. Bootstrapping
8. Uncertainty due to buffer/ linepack effects:
− buffer: volume between reference meter and meter-under-test
− unsteady effects (temperature) cause a violation of the steady mass
flow assumption

Page 17 GFM2012 Module 5A

 Uncertainty sources
Other

MUT
1. Repeatability
2. Uncertainty in reading (especially for the lower flow rates)

Connection standard and MUT

1. Uncertainty due to leak flow in volume between standard and meter-
under-test
2. Installation effects
3. Interaction between measurement system and quantity measured

Human
1. Human errors and mistakes

Page 18 GFM2012 Module 5A

 Uncertainty
Setting up an uncertainty budget

1. Evaluation of all relevant variables influencing the uncertainty

2. Mathematical model for calibration
measurand (em) = f (B,pm,ps,Zs,Zm,Nm,Ns,Im,Is, m, s,T0,es)

3. Evaluation of all (standard) uncertainties u pm

4. Computation of sensitivity coefficients
f (...)
c pm
pm
5. Computation of the contribution of each parameter u pm Q pmc pm

6. Computation of the total uncertainty by means of ‘quadratic summation’

Page 19 GFM2012 Module 5A

 Uncertainty
Quantification for a calibration

Uncertainty budget Calibration and

1. Uncertainty of the standard Measurement
Capability (CMC):

2. Uncertainty of the meter-under-test, including

− Repeatability
− Uncertainty in reading
umut

Uncertainty of calibration as stated on the certificate

CMC (calibration facility) and umut (meter-under-test)

Total uncertainty: utot= CMC 2 u mut

2

Page 20 GFM2012 Module 5A

 Uncertainty
Example of an uncertainty budget

Sensitivity coefficient
Quantity Value ci u(x i) ui(em ) ui2(em )/utot 2(em)
c i = em / x i

xi [x i] [x i]-1 [x i]-1 [x i] [-] %

Pm [bara] 60.00 (1+em )/Pm 0.0167 0.001 0.0000 0.01

Ps [bara] 61.02 -(1+em )/Ps -0.0164 0.001 0.0000 0.01
Nm [-] 12000 (1+em )/Nm 8.34E-05 2 0.00017 0.74
Ns [-] 12000 -(1+em )/Ns -8.34E-05 1 -0.00008 0.19
Zm [-] 0.970 -(1+em )/Zm -1.032 0.00097 -0.0010 26.66
Zs [-] 0.970 (1+em )/Zs 1.032 0.00097 0.0010 26.66
m [s] 300.00 -(1+em )/ m -3.34E-03 0 0.0000 0.00
s [s] 300.00 (1+em )/ s 3.34E-03 0 0.0000 0.00
-3
Im [m ] 210.00 -(1+em )/Im -4.77E-03 0 0.0000 0.00
-3
Is [m ] 212.00 (1+em )/Is 4.72E-03 0 0.0000 0.00
0
tm [ C] 35.83 -(1+em )/(T0+t m ) -3.24E-03 0.100 -0.0003 2.79
0
ts [ C] 35.86 (1+em )/(T0+t s ) 3.24E-03 0.100 0.0003 2.79
es [-] 0.0085 (1+em )/(1+es ) 0.993 0.0012 0.00119 37.75
em [-] 0.001183 1 1 0.0003 0.00030 2.39
e m [-] 0.0012 0.0019 100.00

Page 21 GFM2012 Module 5A

Uncertainty sources temperature
transmitter

• Manufacturer’s specifications
• Uncertainty of the calibration
• Repeatability (type A uncertainty)
• Vibration
• Power supply effects
• Mounting effects
• Pipe wall temperature different from gas temperature (heat
conduction through transducer and radiation)
• Gas temperature may not be uniform across the diameter of pipe
(important at low flow rates)

Page 22 GFM2012 Module 5A

Uncertainty sources pressure transmitter

• Manufacturer’s specifications
• Uncertainty of the calibration
• Repeatability (type A uncertainty)
• Uncertainty in reading
• Atmospheric pressure
• Ambient temperature effects
• Vibration
• Power supply effects
• Mounting effects

Page 23 GFM2012 Module 5A

VSL
Thijsseweg 11
2629 JA Delft
Postbus 654
2600 AR Delft

T 015 269 15 00
F 015 261 29 71
E [email protected]
I www.vsl.nl