511

A publication of

CHEMICAL ENGINEERING TRANSACTIONS

VOL. 36, 2014
Guest Editors: Valerio Cozzani, Eddy de Rademaeker
Copyright 2014, AIDIC Servizi S.r.l.,
ISBN 978-88-95608-27-3; ISSN 2283-9216

The Italian Association

of Chemical Engineering
www.aidic.it/cet
DOI: 10.3303/CET1436086

Consequence Analysis: Comparison of Methodologies

under API Standard and Commercial Software
Chiara Vianelloa, Ludovica Guerrinia, Giuseppe Maschio*a, Alberto Murab
a

Dipartimento di Ingegneria Industriale, University of Padova, Via F.Marzolo 9, 35131 Padova, Italy
Antea S.r.l., Via San Crispino 82, 35129 Padova, Italy
[email protected]
b

In this study the procedure for conducting the consequence analysis according to API RBI is presented.
The guidelines of API methodology are reported in the document API 581. The calculation of
consequences was carried out by Inspection Manager tool (IM), thanks to the implementation of RBI
analysis in this software developed in a previous work (Vianello et al. 2013), and Phast, a commercial
software package for the estimation consequences modelling.
Through the software, two example of quantitative consequence analysis was conducted by taking into
account the consequences of releases both of flammable-explosive and toxic compound based on real
process operating conditions.
In both methods, the results are determined as an affected impact area and are useful to establish a
comparison between the two software, in order to putting out the differences in the calculation of the
consequences.
Finally, it is important to highlight that API approach is greatly simplified procedure and the consequence
areas calculated using API methodology are not intended to be employed during unit design and other
safety purposes, but only to establish a relative risk ranking of equipment items on the basis of risk, in RBI
assessment.

1. Introduction
Loss of containment from processing equipment and the subsequent release of hazardous materials may
result in damage to surrounding equipment and produce serious injury to personnel, production losses,
and undesirable environment impacts.
A Quantitative Risk Assessment (QRA) is a valuable tool for determining the risk derived from the use,
handling, transport and storage of dangerous substances (Egidi et al., 1995; Marhavilas et al. 2011; TNO,
1999). QRAs are used to evaluate the potential risk caused by the activity and to provide the competent
authorities with relevant information to enable decisions on the acceptability of risk related to
developments on site, or around the establishment or transport route.
The calculation of risk is dependent on the determination of a probability of failure combined with the
consequence of failure.
Consequence analysis quantifies the vulnerable zones for an incident. Once these zones are identified,
the risk analysis suggests measures of mitigation or prevention that can be proposed to eliminate damage
to plant and potential injury to personnel. Estimation of vulnerability zone of such an incident plays an
important role in preparing a realistic emergency plan.
An error of consequences estimation is directly associated an error to estimated of risk and ultimately the
risk reduction requirements.
For these reasons, it is important to determine correctly consequences of a leak or rupture of a
component.
The aim of the study is to compare the consequence estimation approach described in the API Standard
581 and the approach used in commercial software, such as Phast (DNV software).

Please cite this article as: Vianello C., Guerrini L., Maschio G., Mura A., 2014, Consequence analysis: comparison of methodologies under
api standard and commercial software, Chemical Engineering Transactions, 36, 511-516 DOI: 10.3303/CET1436086

512

2. Consequence Analysis calculation procedures

The different approach for the calculation of the consequence of a leak or rupture of a component for use
in API RBI and in Phast is described in this paragraph.
The consequence analysis in the Risk-Based Inspection (API RBI) assessment is performed to aid in
establishing a ranking of equipment items on the basis of risk and provide a suitable inspection plan in the
refining and petrochemical industry.
In API RBI the consequences are calculated using a simplified procedure based on empirical equations
and a predefined set of hole sizes that reflects the range of possible outcomes. The results are expressed
as an affected impact area in quantitative terms.
In particular, impact area from flammable and explosive substances are calculated using event trees to
determine the probabilities of various outcomes, combined with computer modelling to determine the
magnitude of the consequence. Consequence areas are quantified based on the effects of thermal
radiation and overpressure on surrounding equipment and personnel. Additionally, the cloud dispersion
analysis methods are used to quantify the magnitude of flammable release and to determine the extent
and duration of personnel exposure toxic release. Toxic consequences are calculated using computer
modelling to determine the magnitude of the consequence area as a result of overexposure of personnel
to toxic concentrations within a vapour cloud.
Event trees are utilized to assess the probability of each of the various event outcomes and to provide a
mechanism for probability-weighting the loss of containment consequences.
Two levels of assessment are available in API RBI methodology for computing the consequence of failure:
the former, or Level 1, is a simplistic method to estimate the consequence area for a limited number of
representative fluids; the Level 2 is intended to be much more rigorous and can be applied to wider range
hazardous fluids.
In Level 1 consequences analysis contains table lookups and graphs that can readily be used to calculated
the consequence of releases without the need of specialized consequence modelling software or
technique. The simplifying assumptions are listed below:

The fluid phase upon release can only be either a liquid or a gas, depending on the storage
phase and the phase expected to occur upon release to the atmosphere. In general, no
consideration is given to the cooling effects of flashing liquid, rainout, jet liquid entrainment or
two-phase.

Fluid properties for representative fluids containing mixtures are based on average value.

Probabilities of ignition have been pre-determined for each of the representative fluids as a
function of temperature, fluid Auto Ignition-Temperature (AIT) and release type. These values are
constants, totally independent of the release rate.
This limitations can be overcome by Level 2 consequence analysis may be used in cases where more
rigorous calculation are required and the assumptions of the Level 1 are not valid. For example this
occurs if the specific fluid is not represented adequately within the list of reference fluid groups provided in
the Level 1 or if it is necessary to consider the effects of two-phase releases or BLEVEs. On the other
hand, the Level 2 requires a large amount of input data.
In this study is considered the Level 1 consequence analysis procedure as a simplistic method for
approximating the consequence area of a hazardous release.
First of all, a representative fluid that most closely matches the fluid contained in the pressurized system
being evaluated is selected from the list. Because very few refinery and chemical plant streams are pure
materials, the selection of a representative fluid almost always involves making some assumptions.
It is important to note that the flammable consequence results are not highly sensitive to the exact material
selected, provided the molecular weights are similar, because air dispersion properties and heat of
combustion are similar for all hydrocarbons with similar molecular weights. This is particularly true for
straight chain alkanes, but becomes less true as the materials become less saturated or aromatic. If the
mixture contains toxic components and a toxic consequence calculation is desired, the choice of reference
fluid is still required, even if the toxic component only makes up a small traction of the mixture. Once it has
been selected the representative fluid it is necessary to determine its average properties which are
dependent on the stored phase of the fluid. In this analysis the only inputs required are basic fluid
properties (such as molecular weight, density and ideal gas specific heat ratio, k) and operating conditions.
The initial phase of the hazardous material inside the equipment prior to coming into contact with the
atmosphere is required to be defined as either liquid or vapour, and the normal boiling point is used in
determining the phase of the material following the release.
The second step is the calculation of theoretical release rate that depend upon the physical properties of
the material, the initial phase, the process operating conditions and the assigned release hole size.

513

The correct release rate equation must be chosen, based on the phase of the material when it is inside the
equipment item, and its discharge regime (sonic or subsonic), as the material is released. For a two-phase
release a conservative assumption is to utilize the liquid release equation, even if it always should be
preferred a Level 2 analysis.
Different analytical models and methods are available to estimate the effects of an instantaneous versus a
continuous type of release. It is very important to determine the properly release type, because the
calculated consequences can differ greatly according the type of analytical model chosen to represent the
release.
In API RBI, the release is modelled as one of two following type:
a) Instantaneous Release (or puff) is one that occurs so rapidly that the fluid disperses as a single
large cloud or pool;
b) Continuous Release (or plume) is one that occurs over a longer period of time, allowing the fluid
to disperse in the shape of an elongated ellipse (depending on weather conditions).
The transition between a continuous release and an instantaneous release in API RBI is defined as
release where more than 4,536 kg of fluid mass escapes in less than 3 min.
The third step is the calculation of the consequence areas associated with several event outcomes for
release of flammable and toxic fluids.
The flammable and explosive consequences for either component damage and personnel injury are
measured in terms of the area affected by the ignition of a release.
The flammable consequence area is determined from a simple polynomial expression that is a function of
the release magnitude and the representative fluid.
The consequence equations are presented in the following generic form:
CA

a rate

(1)

CA

a mass

(2)

The Equation 1 is used for continuous release and the Equation 2 is used for an instantaneous release.
The coefficient a, b for component damage areas and personnel injury areas, present in the previous
equations, are provided for reference fluids. If the release is steady state and continuous, the release rate
is used as the input to the consequence analysis; otherwise if the release is instantaneous, the release
mass is required into the equation.
An event tree analysis is performed by listing possible events or outcomes and providing estimates for the
probabilities of each event. Probabilities of ignition, probabilities of delayed ignition and other probabilities
in the Level 1 event tree are selected based on expert for each of the reference fluids, temperature
(proximity to the AIT) and release types (i.e. instantaneous or continuous).
These probabilities are constant and independent of release rate or mass.
There are several potential consequence outcomes for any release involving a flammable material and an
equation represents the consequence area referred to a particular event outcome; however, a single
combined result is determined as the average of all possible consequence outcomes, weighted according
to probability. The consequence area for each outcome, computed from Equation 3, is multiplied by the
associated event tree probabilities. If the impact criterion uses only a portion of the consequence area (for
instance, flash fires use only 25 % of the area within the LFL for equipment damage, see Table 1) include
this in the probability equation.
The equation that summarizes the result of the process is as follows:
CA

p CA

(3)

To calculate these consequence areas for a particular event outcome (pool fire, VCE, etc.), API method
uses threshold limits for thermal radiation (for jet fire, pool fire and fireball) and overpressure, sometimes
referred as impact criteria. They are reported in the Table 1.
Table 1: Threshold limits used in API RBI
Parameter
Component damage criteria
Explosion Overpressure (kPa)
34.5
Thermal Radiation (kW/m2)
37.8
25 % of the area within the lower
Flash Fire
flammability limits (LFL) when the
cloud is ignited

Personnel injury criteria

20.7
12.6
The LFL Limits of the cloud when
ignite

514

For pool fires, the API RBI method assumes a dike size of 30.5 m by 30.5 m (929 m2) and estimates the
flammable consequences due to a maximum pool fire of that size.
The predicted results using the above threshold limits are intended to produce a relative risk ranking,
which, while being considered to be reasonably accurate, are not the highest levels of consequence that
could be estimated for a given accident sequence.
The computer modelling necessary to determine consequence areas associated with cloud dispersion
(flash fires, VCEs, toxic releases) require specific input regarding meteorological and release conditions.
For the Level 1 Consequence Analysis, meteorological conditions representative of the Gulf Coast annual
averages are used. The meteorological input assumptions that you have to enter in order to complete the
data for this analysis is small, as shown in the Table 2. They are satisfactory for a wide variety of plant
conditions.
Table 2: Meteorological input
Parameters
Value
Atmospheric Temperature
21C
Relative Humidity
75 %
Wind speed
12.9 km/h
Stability Class
D
Surface Roughness Parameter 30.5 mm
A similar procedure is used for determining the consequence associated with releases of toxic chemicals
such as hydrogen sulphide, hydrogen fluoride, ammonia or chlorine, that typically contribute to toxic risks
for a refinery. Toxic impact areas are based on Probit equations and can be assessed whether the stream
is pure or a percentage of a hydrocarbon stream.
The toxic release rate or mass to be used in the toxic consequence analysis is determined based on the
tox
mass fraction of the toxic component (mfract ), that is present in the release fluid.
rate

mfract

mass

W (4)

mfract

mass

(5)

For pure toxic fluids, the toxic release rate (ratentox) is equal to the release rate (Wn) and the toxic release
mass (massntox) is equal to the release mass (massn). The toxic consequences areas are calculated
through specific equations that depend on toxic fluid and release type. For example, the toxic
consequence area for Ammonia and Chlorine are computed using the Equations 6 for continuous release
and the Equation 7 for an instantaneous release:
CA

e rate

(6)

CA

e mass

(7)

The constants e and f are functions of the release duration and provided for each toxic fluid.
In the event the release can involve both toxic and flammable outcomes, it is assumed that either the
flammable outcome consumes the toxic material, or the toxic materials are dispersed and flammable
materials have insignificant consequences. In this case, the probability for the toxic event is the remaining
non-ignition frequency for the event (i.e., the probability of safe dispersion).
The calculation of consequences according API RBI approach is carried out by Inspection Manager tool
(IM) , thanks to the implementation of RBI analysis in this software developed in a previous study. The
number of values that you have to enter in order to complete the data for this model is small; moreover the
software allows the insertion of data in a faster way, simplifying the analysis.
At this point it is necessary to conduct the consequence analysis with a software package for the
estimation consequences modelling, such as Phast, to establish a comparison between the two
methodologies and highlight the potential limitations of the procedure used in API RBI. The modelling of
consequences used in the Phast code is in agreement with the equations of the Yellow Book of TNO
(TNO). The study carried out by Phast requires a different type of input data for the consequence analysis,
such as the definition of the calculation model, discharge material, inventory, operating conditions, weather
conditions. The user defines a given hazardous event, selecting the most suitable Model and its own set of
values for the input data to represent a particular hazardous event.
Model known as the Vessel/Pipe Source Model is selected. It considers the release of material from its
storage or process conditions in a vessel or pipe, through all the stages in its dispersion to a harmless

515

concentration, and it also performs fire, explosion and toxic calculations to obtain representative effect
zones for the dispersing cloud.
Model performs dispersion and effects calculations for a release from containment, setting the levels of
radiation intensity and concentration on basis the same threshold values adopted in API (see Table 1).
The program processes the consequence calculations for Dispersion and Flammable Models and it
performs the analysis for specific weather conditions, reported in Table 2.

3. Quantitative examples
The present study considers two quantitative examples to compare the results by two software and put out
the differences in the calculation of the consequences.
The types of hazardous event that are selected in the analysis are as follows:

a leak in a piping containing an inflammable material

a leak in a piping containing a toxic material

This allows to take into account either the flammable-explosive consequence area and toxic consequence
area.
In both cases the quantitative analysis by taking as reference a small leak in a pipeline outgoing from a
storage vessels and the release hole diameter is 6.4 mm.
In the first case the pipeline has a diameter equal to 100 mm and the fluid inside is benzene. The phase of
benzene at normal operating conditions is liquid and also the phase of fluid at ambient condition after
release to atmosphere is liquid. The representative fluid selected is aromatic. The main input data to
estimate the consequence area are provided in Table 3.
Table 3: Input data to calculate explosive-flammable consequences for release of benzene
Parameter
Material
Process conditions
Location
Scenario

Input Field
Discharge material
Inventory
Temperature
Pressure
Elevation
Scenario type
Hole diameter
Phase Released

Value
Benzene
450 kg
60C
2 bar
3m
Leak
6.4 mm
Liquid

The release rate calculated by RBI IM and Phast is respectively 0.23 kg/s and 0.38 kg/s.
Subsequently two software compute the damage areas for auto-ignition not likely and continuous release.
The outcomes considered are jet fire, pool fire, flash fire and VCE and the consequence areas of each
individual event is combined into a single probability weighted empirical equation representing the overall
2
2
consequence area of the event tree. The final component damage area is 4.8 m for API RBI and 7.5 m
2
2
for Phast and the final personnel injury damage area is 18 m for API RBI and 36 m for Phast.
The second case considers a gas release from an horizontal pipeline, containing ammonia at saturation
conditions and ambient temperature. The toxic release is constituted by pure ammonia.
The rupture is assumed to occur 3 meters from the ground. The dispersion analysis was performed using
a saturated liquid at ambient temperature, with liquid being released from a low pressure storage tank.
Toxic impact criteria used is for a Probit value of 5.0. For ammonia, set the dispersion concentration of
interest to 25 ppm. The Table 4 shows a information needed to conduct the analysis for this equipment.
Table 4: Input data to calculate toxic consequences for release of ammonia
Parameter
Material
Process conditions
Location
Scenario

Input Field
Discharge material
Inventory
Temperature
Pressure
Elevation
Scenario type
Hole diameter
Phase Released

Value
Ammonia
100 kg
20C
2 bar
3m
Leak
6.4 mm
Gas

516

The release rate calculated by RBI IM and Phast is respectively 1.0210-2 kg/s and 1.5610-2 kg/s.
The toxic consequence area associated with continuous release of ammonia is 14.3 m2 if it is computed
2
using API procedure and 12.3 m using Phast. Through the two software, the quantitative analysis was
been conducted and the estimated consequence areas are summarized in the following table.
Table 5: Results for flammable and toxic consequence area

Damage component
Injury personnel

Injury personnel

Flammable Consequence Area

API RBI
4.8 m2
18 m2
Toxic Consequence Area
API RBI
14.3 m2

Phast
7.5 m2
36 m2
Phast
12.3 m2

The results highlights that API RBI approach underestimates the consequences area for release of
flammable substance compared to Phast, while for toxic consequence area the values are similar in both
cases. It is also interesting to note that the areas for release of benzene calculated by Phast are about
twice those determined by RBI IM.
Phast normally gives conservative results in the consequence calculations.

4. Conclusions
The consequence modelling procedure for API RBI is greatly simplified approach to a relatively complex
discipline. The input process involves examining a large number of assumptions that are implicit in the
procedure, starting the selection of representative fluid and from all the parameters associated with this
choice. API RBI consequence analysis includes also a simplified methodology for assigning the
effectiveness of various type of detection, mitigation and isolation systems typically used in petrochemical
processing plant. These systems affect a release in different ways: some systems reduce magnitude and
duration of the release by detecting and isolating the leak; other ones reduce the consequence area by
minimizing the chances for ignition or limiting the spread of material. In the present study the impact of
these systems on release magnitude and consequence estimation was neglected, according a more
conservative approach; hence the theoretical release rate (or mass rate) is used as the input to the
analysis and the magnitude of final consequence area is not reduced by a specific factor, depending on
the effectiveness of mitigation system.
Finally, it is important to note that the consequences calculated according to the procedures described in
API RBI are not intended to be used in a rigorous consequence analysis of a component and if more
accurate consequence estimates are needed, the analyst should refer to more rigorous analysis
techniques, such as those used in quantitative risk analysis.
Acknowledgements
Regione Veneto is gratefully acknowledged for the financial support of the grant of Ludovica Guerrini.
References
API 581, 2008, Risk Based inspection, Base Resource Document, American Petroleum Institute, 2nd
edition, September 2008, Washington, D.C., USA.
Egidi, D., Foraboschi, F. P., Spadoni, G., & Amendola, A., 1995, The ARIPAR project: analysis of the
major accident risks connected with industrial and transportation activities in the Ravenna area.
Reliability Engineering & System Safety, 49(1), 7589. doi:10.1016/0951-8320(95)00026-X.
Marhavilas, P. K., Koulouriotis, D., Gemeni, V., 2011, Risk analysis and assessment methodologies in the
work sites: On a review, classification and comparative study of the scientific literature of the period
20002009. Journal of Loss Prevention in the Process Industries, 24(5), 477523.
Phast DNV, Tutorial Manual.
TNO. ,1999, Guidelines for quantitative risk assessment (Purple Book). (CPR18E, Ed.). CPR18E.
TNO. ,1997, Methods for the calculation of the physical effects (Yellow Book). (CPR14E, Ed.). CPR14E.
Vianello, C., Maschio, G., Mura, A., Babolin, D., Gambato, F., Attori, C., 2013, Development of a RBI tool
for inspection management in chemical plants Chemical Engineering Transactions, 31, 235-240.