GRI-00/0189

A MODEL FOR SIZING HIGH CONSEQUENCE AREAS

ASSOCIATED WITH NATURAL GAS PIPELINES

TOPICAL REPORT

Prepared by:

Mark J. Stephens

C-FER Technologies
200 Karl Clark Road
Edmonton, Alberta T6N 1H2
CANADA

C-FER Report 99068

Prepared for:

GAS RESEARCH INSTITUTE

Contract No. 8174

GRI Project Manager

Keith Leewis,
Pipeline Business Unit

October 2000
LEGAL NOTICE

This report was prepared by C-FER Technologies as an account of work sponsored by the Gas
Research Institute (GRI). Neither C-FER, GRI, members of GRI, nor any person acting on their
behalf:

a. MAKES ANY WARRANTY OR REPRESENTATION, EXPRESS OR IMPLIED

WITH RESPECT TO THE ACCURACY, COMPLETENESS, OR USEFULNESS OF THE
INFORMATION CONTAINED IN THIS REPORT, OR THAT THE USE OF ANY
INFORMATION, APPARATUS, METHOD, OR PROCESS DISCLOSED IN THIS
REPORT MAY NOT INFRINGE PRIVATELY OWNED RIGHTS, OR
b. ASSUMES ANY LIABILITY WITH RESPECT TO THE USE OF, OR FOR ANY AND
ALL DAMAGES RESULTING FROM THE USE OF, ANY INFORMATION,
APPARATUS, METHOD, OR PROCESS DISCLOSED IN THIS REPORT.

i
 REPORT DOCUMENTATION PAGE Form Approved
OMB No. 0704-0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources,
gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this
collection of information including suggestions for reducing this burden to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis
Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, D.C. 20503.
1. AGENCY USE ONLY 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
October, 2000 Topical Report
4. TITLE AND SUBTITLE 5. FUNDING NUMBERS
A Model for Sizing High Consequence Areas Associated with Natural Gas Pipelines
GRI Contract 8174

6. AUTHOR(S)
Mark J. Stephens

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION

C-FER Technologies REPORT NUMBER
200 Karl Clark Road
Edmonton, Alberta T6N 1H2 C-FER J068
CANADA

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITORING

GRI AGENCY REPORT NUMBER
8600 West Bryn Mawr Ave.
Chicago, IL 60631-3562 GRI-00/0189

11. SUPPLEMENTARY NOTES

12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

13. ABSTRACT (Maximum 200 words)

This report developed a simple and defendable approach to sizing the ground area potentially affected by a worst-case ignited rupture of a high-pressure natural gas
pipeline. Based on this model, a simple equation has been developed that relates the diameter and operating pressure of a pipeline to the size of the area likely to
experience high consequences in the event of an ignited rupture failure. Pipeline incident reports, located in the public domain, were reviewed and provide the
basis for evaluating the validity of the proposed affected area equation. The correlation suggests that the simple equation provides a credible estimate of affected
area.
14. SUBJECT TERMS 15. NUMBER OF PAGES

16. PRICE CODE

$125
17. SECURITY 18. SECURITY 19. SECURITY CLASSIFICATION 20. LIMITATION OF
CLASSIFICATION OF CLASSIFICATION OF THIS OF ABSTRACT ABSTRACT
REPORT PAGE Unclassified
Unclassified Unclassified
NSN 7540-01-280-5500 Standard Form 298 (Rev.2-89)

ii
 RESEARCH SUMMARY

Title A Model for Sizing High Consequence Areas Associated with Natural
Gas Pipelines

Contractor(s) C-FER Technologies

GRI-Contract 8174
Number

Principal Mark J. Stephens

Investigator(s)

Report Type Topical Report

Objective State To develop a simple and defendable approach to sizing the ground area
potentially affected by the failure of a high-pressure natural gas pipeline.

Technical The rupture of a high-pressure natural gas pipeline can lead to outcomes that can pose a
Perspective significant threat to people and property in the immediate vicinity of the failure location.
The dominant hazard is thermal radiation from a sustained fire and an estimate of the
ground area affected by a credible worst-case event can be obtained from a model that
characterizes the heat intensity associated with rupture failure of the pipe where the
escaping gas is assumed to feed a fire that ignites very soon after line failure.

Technical Approach An equation has been developed that relates the diameter and operating pressure of a
pipeline to the size of the affected area in the event of a credible worst-case failure event.
The model upon which the hazard area equation is based consists of three parts: 1) a fire
model that relates the rate of gas release to the heat intensity of the fire; 2) an effective
release rate model that provides a representative steady-state approximation to the actual
transient release rate; and 3) a heat intensity threshold that establishes the sustained heat
intensity level above which the effects on people and property are consistent with the
adopted definition of a High Consequence Area (HCA).

Results For methane with an HCA threshold heat intensity of 5,000 Btu/hr ft2, the hazard area
equation is given by:
r = 0.685 p d 2
where r is the hazard area radius (ft), d is the line diameter (in), and p is the maximum
operating pressure (psi).

Project Implications Natural gas transmission line operators will provide periodic assurances that their
pipelines are safe. The Federal code 49CFR192 mandates increased wall thickness
thereby reducing the corrosion and mechanical damage risks as the population density
increases. The definition of High Consequence Areas is expected to require additional
protection for people with limited mobility such as day care centers, old age homes, and
prisons. This report suggests the definition for the HCA area of increased protection be
set by two parameters, the pipe diameter and it’s operating pressure.

iii
TABLE OF CONTENTS

Legal Notice i
Report Documentation ii
Research Summary iii
Table of Contents v

1. INTRODUCTION................................................................................................................1

1.1 Scope and Objective 1

1.2 Technical Background 1
1.3 Report Organization 2

2. HAZARD MODEL ..............................................................................................................3

2.1 Overview 3
2.2 Fire Model 3
2.3 Effective Release Rate Model 5
2.4 Heat Intensity Threshold 7
2.5 Hazard Area Equation 10

3. MODEL VALIDATION......................................................................................................12

4. REFERENCES.................................................................................................................15

iv
1. INTRODUCTION

1.1 Scope and Objective

This report summarizes the findings of a study conducted by C-FER Technologies (C-FER),
under contract to the Gas Research Institute (GRI), to develop a simple and defendable approach
to sizing the ground area potentially affected by the failure of a high-pressure natural gas
pipeline. This work was carried out at the request of the Integrity Management and Systems
Operations Technical Advisory Group (IM&SO TAG), a committee of GRI.

1.2 Technical Background

The failure of a high-pressure natural gas pipeline can lead to various outcomes, some of which
can pose a significant threat to people and property in the immediate vicinity of the failure
location. For a given pipeline, the type of hazard that develops, and the damage or injury
potential associated with the hazard, will depend on the mode of line failure (i.e., leak vs.
rupture), the nature of gas discharge (i.e., vertical vs. inclined jet, obstructed vs. unobstructed jet)
and the time to ignition (i.e., immediate vs. delayed). The various possible outcomes are
summarized in Figure 1.1.

product immediate release delayed delayed

release ignition unobstructed local ignition remote ignition

yes
Fireball ⇒ Jet/trench fire

no yes yes
Jet/trench fire

no
No significant hazard*

no yes
Jet/trench fire

no yes
Flashfire ⇒ Jet/trench fire
No significant hazard*
no

* ignoring hazard potential of overpressure and flying debris

Figure 1.1 Event tree for high pressure gas pipeline failure
(adapted from Bilo and Kinsman 1997).

For gas pipelines, the possibility of a significant flash fire resulting from delayed remote ignition
is extremely low due to the buoyant nature of the vapor, which generally precludes the formation
of a persistent flammable vapor cloud at ground level. The dominant hazard is, therefore,
thermal radiation from a sustained jet or trench fire, which may be preceded by a short-lived
fireball.

In the event of line rupture, a mushroom-shaped gas cloud will form and then grow in size and
rise due to discharge momentum and buoyancy. This cloud will, however, disperse rapidly and a
quasi-steady gas jet or plume will establish itself. If ignition occurs before the initial cloud

1
disperses, the flammable vapor will burn as a rising and expanding fireball before it decays into a
sustained jet or trench fire. If ignition is slightly delayed, only a jet or trench fire will develop.
Note that the added effect on people and property of an initial transient fireball can be accounted
for by overestimating the intensity of the sustained jet or trench fire that remains following the
dissipation of the fireball.

A trench fire is essentially a jet fire in which the discharging gas jet impinges upon an opposing
jet and/or the side of the crater formed in the ground. Impingement dissipates some of the
momentum in the escaping gas and redirects the jet upward, thereby producing a fire with a
horizontal profile that is generally wider, shorter and more vertical in orientation, than would be
the case for a randomly directed and unobstructed jet. The total ground area affected can,
therefore, be greater for a trench fire than an unobstructed jet fire because more of the heat-
radiating flame surface will typically be concentrated near the ground surface.

An estimate of the ground area affected by a credible worst-case failure event can, therefore, be
obtained from a model that characterizes the heat intensity associated with rupture failure of the
pipe, where the escaping gas is assumed to feed a sustained trench fire that ignites very soon
after line failure.

Because the size of the fire will depend on the rate at which fuel is fed to the fire, it follows that
the fire intensity and the corresponding size of the affected area will depend on the effective rate
of gas release. The release rate can be shown to depend on the pressure differential and the hole
size. For guillotine-type failures, where the effective hole size is equal to the line diameter, the
governing parameters are, therefore, the line diameter and the pressure at the time of failure.
Given the wide range of actual pipeline sizes and operating pressures, a meaningful fire hazard
model should explicitly acknowledge the impact of these parameters on the area affected.

1.3 Report Organization

The hazard model developed to relate the area potentially affected by a failure to the diameter
and pressure of the pipeline is described in Section 2.0. Validation of the proposed hazard area
model, based on historical data from high-pressure gas pipeline failure incidents in the United
States and Canada, is presented in Section 3.0.

2
2. HAZARD MODEL

2.1 Overview

An equation has been developed that relates the diameter and operating pressure of a pipeline to
the size of the area likely to experience high consequences in the event of a credible worst-case
failure event. The hazardous event considered is a guillotine-type line rupture resulting in
double-ended gas release feeding a trench fire that is assumed to ignite soon after failure.

The hazard model upon which the hazard area equation is based consists of three parts: 1) a fire
model that relates the rate of gas release to the heat intensity of the fire as a function of distance
from the fire source; 2) an effective release rate model that provides a representative steady-state
approximation to the actual transient release rate; and 3) a heat intensity threshold that
establishes the sustained heat intensity level above which the effects on people and property are
consistent with the definition of a high consequence area. Note that in the context of this study,
an HCA is defined as the area within which the extent of property damage and the chance of
serious or fatal injury would be expected to be significant in the event of a rupture failure.

The basis for each model, and any underlying assumptions, are described in Sections 2.2
through 2.4. The hazard area equation obtained by combining the model components is
described in Section 2.5.

2.2 Fire Model

A jet flame can be idealized as a series of point source heat emitters spread along the length of
the flame (see Figure 2.1). Each point source can be is assumed to radiate an equal fraction of
the total heat with the heat flux I i at a given location resulting from point source i being given
by (Technica 1988):

η X g Qeff H c
Ii = [2.1]
4 n p π xi2

where H c = heat of combustion (constant for given product) ≅ 50,000 kJ/kg for methane;
η = combustion efficiency factor = 0.35;
Xg = emissivity factor = 0.2;
np = number of point sources;
Qeff = effective gas release rate; and
xi = radial distance from heat source i to the location of interest.

The total heat flux reaching a given point is obtained by summing the radiation received from
each point source emitter.

3
 Fire
Plume Thermal
Radiation
Damage
Receptor

Figure 2.1 Conceptual fire hazard model.

A simplifying assumption, that generally yields a conservative estimate of the total heat flux
received by ground level damage receptors, involves collapsing the set of heat emitters into a
single point source emitter located at ground level (see Figure 2.2).

Figure 2.2 Simplified fire hazard model.

The resulting equation for the total heat flux I at a horizontal distance of r from the fire center is
given by:

η X g Qeff H c
I= [2.2]
4π r 2

4
This simplification is, in some respects, more consistent with the geometry of a trench fire
which, due to the jet momentum dissipation (see Section 1.2), concentrates more of the heat-
radiating flame surface near ground level. Note, however, that while a ground-level point source
model represents a conservative approximation to a vertically-oriented jet flame or trench fire,
this conservatism is partially offset by the fact that the model does not explicitly account for the
possibility of laterally-oriented jets and/or the effects of wind on the actual position of the fire
center relative to the center of the pipeline.

Note, also, that for a single point source emitter located at ground level directly above the
pipeline, the locus of points receiving a heat flux of I defines a circular area of radius r centered
on the pipeline. Thermal radiation hazard zones of increasing impact severity are, therefore,
described by concentric circles centered on the pipeline having radii that correspond to
progressively higher heat fluxes.

The adopted heat flux versus distance relationship given by Equation [2.2] represents an
extension of the widely recognized flare radiation model given in API RP 521 (API 1990). It can
be shown to be less conservative than the API flare model (i.e., it gives lower heat intensity
estimates at a given distance) but this should not be considered surprising since the API model is
widely recognized to be conservative (Lees 1996).

The adopted model is also preferred over some of the more generic, multi-purpose models
available for industrial fire hazard analysis because it acknowledges factors, ignored by other
models, that play a significant role in mitigating the intensity of real-world jet fire events. In
particular, it accounts for the incomplete combustion of the escaping gas stream (through the
combustion efficiency factor η ), and it acknowledges (through the emissivity factor X g ) that a
significant portion of the radiant heat energy will be absorbed by the atmosphere before it can
reach targets at any significant distance from the flame surface.

2.3 Effective Release Rate Model

The rate of gas release from a full-bore line rupture varies with time. Within seconds of failure,
the rate of release will have dropped to a fraction of the peak initial value and over time the
release rate will decay even further. This tendency for rapid release rate decay is illustrated in
Figure 2.3, which shows how the rate would be expected to vary with time for two representative
line diameter and operating pressure combinations. The relative release rate estimates shown in
the figure were calculated using a non-dimensional rate decay model presented in a study by the
Netherlands Organization of Applied Scientific Research, Division of Technology for Society
(TNO 1982) which is based on realistic gas flow and decompression characteristics and which
acknowledges both the compressibility of the gas and the effects of pipe wall friction.

5
 1

Release Rate (fraction of initial)

0.8 36 inch line at 870 psig

0.6 8 inch line at 580 psig

0.4 λ = 0.33

0.2

0
0 10 20 30 40 50 60
Time (seconds)

Figure 2.3 Release rate decay.

The peak initial release rate from the single end of a full-bore line rupture can be estimated using
the widely recognized gas discharge equation given by the Crane Co. (1981) for sonic or choked
flow through an orifice:

π d2 ϕ
Qin = Cd p [2.3a]
4 a0

γ +1
 2  2(γ −1)
where ϕ = flow factor = γ   ; [2.3b]
 γ + 1
γ RT
a0 = sonic velocity of gas = ; [2.3c]
m
Cd = discharge coefficient ≅ 0.62;
γ = specific heat ratio of gas ≅ 1.306 for methane;
R = gas constant = 8,310 J/(kg mol)/K;
T = gas temperature ≅ 288 K or 15 C;
m = gas molecular weight ≅ 16 kg/mol for methane;
d = effective hole diameter ≅ line diameter; and
p = pressure differential ≅ line pressure.

Given that the release rate is highly variable, it follows that the size and intensity of the
associated fire will also vary with time and the peak intensity of the fire will depend on exactly

6
when ignition occurs. The hazard model developed herein accounts for the above by
approximating the transient jet or trench fire as a steady state fire that is fed by an effective
release rate. The effective release rate is a fractional multiple of the peak initial release rate that
can be used to obtain estimates of sustained heat flux that are comparable to those obtained from
a more realistic transient fire model that assumes a slight delay in ignition time.

For a guillotine-type failure of a pipeline resulting in double-ended release, the effective release
rate that is assumed to feed a steady-state fire is given by:

π d2 ϕ
Qeff = 2 λ Qin = 2 λ Cd p [2.4]
4 a0

where λ is the release rate decay factor and the factor of 2 acknowledges that gas will be
escaping from both failed ends of the pipeline.

In general, the most appropriate value for the release rate decay factor will depend on the size of
pipeline being considered, the pressure in the line at the time of failure, the assumed time to
ignition, and the time period required to do damage to property or cause harm to people. Given
that even immediate ignition will require several seconds for the establishment of the assumed
radiation conditions and given further that a fatal dose of thermal radiation can be received from
a pipeline fire in well under 1 minute (see Section 2.4), it follows from Figure 2.3 that a rate
decay factor in the range of 0.2 to 0.5 will likely yield a representative steady state
approximation to the release rate for typical pipelines.

In a study of the risks from hazardous pipelines in the United Kingdom conducted by A. D. Little
Ltd. (Hill and Catmur 1995), the authors report using a release rate decay factor of 0.25.
A slightly more conservative value for λ of 0.33 has been adopted herein to ensure that the
sustained fire intensity associated with nearly immediate ignition of fires associated with large
diameter pipelines will not be underestimated (see Figure 2.3). Given that anecdotal information
on natural gas pipeline failures suggests that the time to ignition may typically be in the range of
1 to 2 minutes (as in the Edison, New Jersey incident of 1994), the adopted release rate decay
factor will likely yield an effective release rate estimate that overestimates the actual rate for the
full duration of a typical gas pipeline rupture fire.

2.4 Heat Intensity Threshold

For people, the degree of harm caused by thermal radiation is usually estimated using a model
that relates the chance of burn injury or fatality to the thermal load received where the thermal
load Lp is given by an equation of the form (Lees 1996):

Lp = t I n [2.5]

where t is the exposure duration, I is the heat flux and n is an index.

Various recognized thermal load vs. effect models based on Equation [2.5] are summarized in
Table 2.1 together with calculated estimates of the exposure times required to reach various

7
conditions of injury and mortality for persons exposed to specified heat intensity levels. If it is
assumed that within a 30 second time period an exposed person would remain in their original
position for between 1 and 5 seconds (to evaluate the situation) and then run at 5 mph (2.5 m/s)
in the direction of shelter, it is estimated that within this period of time they would travel a
distance of about 200 ft (60 m). On the further assumption that, under typical conditions, a
person can reasonably be expected to find a sheltered location within 200 ft of their initial
position, a 30 second exposure time is considered credible and is, therefore, adopted as the
reference exposure time for people outdoors at the time of failure.

Radiation Radiation Time to Time to Blister Time to Blister Time to Time to Time to
1 1 3
Intensity Intensity Burn Threshold Threshold - lower Threshold - upper 1% Mortality 50% Mortality 100% Mortality
or Heat Flux or Heat Flux (Eisenberg et al. 1975) (Hymes 1983)2 (Hymes 1983)2 (Hymes 1983)2 (Hymes 1983)2 (Bilo & Kinsman 1997)
2 2 1.15 1.33 1.33 1.33 1.33 1.33
(Btu/hr ft ) (kW/m ) t*I = 195 t*I = 210 t*I = 700 t*I = 1060 t*I = 2300 t*I = 3500
1600 5.05 30.3 24.4 81.3 123.1 267.1 406.4
2000 6.31 23.5 18.1 60.4 91.5 198.5 302.1
3000 9.46 14.7 10.6 35.2 53.4 115.8 176.2
4000 12.62 10.6 7.2 24.0 36.4 79.0 120.2
5000 15.77 8.2 5.4 17.9 27.0 58.7 89.3
8000 25.24 4.8 2.9 9.6 14.5 31.4 47.8
10000 31.55 3.7 2.1 7.1 10.8 23.3 35.5
12000 37.85 3.0 1.7 5.6 8.4 18.3 27.9
Note: 1) Hymes gives a thermal load range (210 to 700) rather than a single value for blister formation
2) the thermal load values given by Hymes are based on a revised interpretation of the results obtained by Eisenberg et al.
3) Bilo and Kinsman assume that 100% mortality corresponds to a lower bound estimate of the thermal load associated with the spontaneous ignition of clothing

Table 2.1 Effects of thermal radiation on people.

The exposure time estimates closest to this reference time are highlighted in Table 2.1 for each
different thermal load effect. Note that the onset of burn injury within the reference exposure
time is associated with a heat flux in the range of 1,600 to 2,000 Btu/hr ft2 (5 to 6.3 kW/m2),
depending on the burn injury criterion. The chance of fatal injury within the reference exposure
time becomes significant at a heat flux of about 5,000 Btu/hr ft2 (15.8 kW/m2), if the significance
threshold is taken to be a 1% chance of mortality (i.e., 1 in 100 people directly exposed to this
thermal load would not be expected to survive).

For property, as represented by a wooden structure, the time to both piloted ignition (i.e., with a
flame source present) and spontaneous ignition (i.e., without a flame source present) can also be
estimated as a function of the thermal load received. For buildings, the thermal load Lb is given
by an equation of the form (Lees 1996):

Lb = (I − I x )t n [2.6]

where Ix is the heat flux threshold below which ignition will not occur.

Models based on Equation [2.6], developed from widely cited tests as re-interpreted by the UK
Health and Safety Executive (Bilo and Kinsman 1997), are summarized in Table 2.2 together
with calculated estimates of the exposure times required for both piloted and spontaneous
ignition at selected heat intensity levels.

8
 Radiation Radiation Time to Time to
Intensity Intensity Piloted Ignition1 Spontaneous Ign.1
or Heat Flux or Heat Flux (Bilo & Kinsman 1997) (Bilo & Kinsman 1997)
(Btu/hr ft2) (kW/m2) (I-14.7)*t0.667=118.6 (I-25.6)*t0.8=167.6
4000 12.62 no ignition no ignition
5000 15.77 1162.3 no ignition
8000 25.24 37.8 no ignition
10000 31.55 18.7 65.0
12000 37.85 11.6 26.3
Note: 1) based on experiments on American whitewood

Table 2.2 Effects of thermal radiation on wooden structures.

From Table 2.2 it can be seen that 5,000 Btu/hr ft2 (15.8 kW/m2), corresponds to piloted ignition
after about 20 minutes (1,200 seconds) of sustained exposure. The table further shows that
spontaneous ignition is not possible at this heat intensity level. It is therefore assumed that this
heat intensity represents a reasonable estimate of the heat flux below which wooden structures
would not be destroyed, and below which wooden structures should afford indefinite protection
to occupants.

Note that the model employed for estimating the effects of thermal radiation on property
explicitly considers the duration of exposure required to cause ignition. Some earlier wood
ignition models, which appear to be the basis for the often cited 4,000 Btu/hr ft2 (12.6 kW/m2)
threshold for piloted wood ignition, are in fact associated with an almost indefinite time to
ignition and are, therefore, considered to be overly conservative given the transient (decaying)
nature of real pipeline rupture fires.

In light of the above, if a high consequence area is defined as the area within which both the
extent of property damage and the chance of serious or fatal injury would be expected to be
significant, it follows that this area can reasonably be defined by a heat intensity contour
corresponding to a threshold value below which:

• property, as represented by a typical wooden structure, would not be expected to ignite

and burn;
• people located indoors at the time of failure would likely be afforded indefinite
protection; and
• people located outdoors at the time of failure would be exposed to a finite but low chance
of fatality.

The information presented on thermal load effects suggests that below 5,000 Btu/hr ft2, a wooden
structure would not be expected to burn and it, thereby, affords indefinite protection to sheltered
persons. Also, this heat intensity level corresponds to approximately a 1 percent chance of
fatality for persons exposed for a credible period of time before reaching shelter. A heat flux of
5,000 Btu/hr ft2 has, therefore, been adopted as the threshold heat intensity for the purpose of
sizing a high consequence area.

9
2.5 Hazard Area Equation

Substituting the expression developed for the effective release rate (Equation [2.4]) into the heat
intensity versus distance formula (Equation [2.2]), replacing all constants and rearranging gives
the following expression for the radial distance to locations where the heat flux is equal to the
threshold value:

2348 p d 2
r= (ft) [2.7]
I th

where I th = threshold heat intensity (Btu/hr/ft2);

p = line pressure (psi); and
d = line diameter (in).
For a threshold heat intensity of 5,000 Btu/hr ft2, the above expression reduces to:

r = 0.685 p d 2 [2.8]

Equation [2.8] can, therefore, be used to estimate the radius of a circular area surrounding the
assumed point of line failure within which the impact on people and property would be expected
to be consistent with the adopted definition of a high consequence area.

Hazard area radii, as calculated using Equation [2.8] are plotted in Figure 2.4 as a function of
line diameter and operating pressure. The figure shows that, for pipelines operating at pressure
levels in the range of 600 to 1,200 psi, the calculated hazard area radius ranges from under 100 ft
for small diameter lines to over 1,100 ft for large diameter lines.

Note that the concept of relating the potential hazard area to the line diameter and operating
pressure is not new. An approach similar to that described herein has been an integral part of the
high pressure gas transmission pipeline code in the United Kingdom since 1977 (Knowles et
al. 1978 and IGE 1993). The standard as developed in the United Kingdom incorporates the
concept of a Building Proximity Distance (BPD), multiples of which serve to define
development exclusion zones and establish the pipeline corridor width for the purpose of
determining Location Class. The BPD is calculated directly from the line diameter and the
maximum operating pressure.

10
 1320
Nominal Diameter

1155 42 in

990 36 in
Hazard area radius (ft)

825 30 in

24 in
660

18 in
495
12 in
330
6 in
165

0
500 600 700 800 900 1000 1100 1200 1300 1400 1500
Maximum operating pressure (psi)

Figure 2.4 Proposed hazard area radius as a function of line diameter and pressure.

11
3. MODEL VALIDATION

Pipeline incident reports, located in the public domain, were reviewed to provide a basis for
evaluating the validity the proposed hazard area model given by Equation [2.8]. The data
sources reviewed included reports on pipeline incidents in the United States prepared by the
National Transportation Safety Board (NTSB) going back to 1970, and similar reports on
incidents in Canada prepared by the Transportation Safety Board (TSB) going back to 1994.
Note that the information extracted from these reports required some interpretation due to
differences in the way the information was reported. The processed data together with hazard
area estimates obtained using Equation [2.8] are summarized in Figure 3.1. A summary of the
information that forms the basis for Figure 3.1 is given in Table 3.1.

TSB-P95H0036 (42@880)

TSB-P94H0036 (36@1000)

TSB-P94H0003 (42@1207)

NTSB-PAR-95-1 (36@970)

NTSB-PAR-87-1 (30@987)

NTSB-PAR-87-1 (30@987)

NTSB-PAR-86-1 (30@1016)

NTSB-PAR-83-2 (20@820)
Proposed HCA radius
Maximum offset to burn extent
NTSB-PAR-77-1 (20@785)
Equivalent radius of burn area
Maximum offset to injury
NTSB-PAR-75-3 (12@497) Maximum offset to fatality

NTSB-PAR-75-2 (30@718)
TSB - Transportation Safety Board (Canada)
NTSB - National Transportation Safety Board (US)
NTSB-PAR-71-1 (14@785)

0 200 400 600 800 1000 1200

Distance (ft)

Figure 3.1 Comparison between actual incident outcomes and the proposed hazard area model.

In interpreting the incident outcomes summarized in Figure 3.1 note the following:

• the equivalent radius of burn area is the radius of a circle having an area equal to the
reported area of burnt ground;
• the maximum offset to burn extent is the maximum reported of inferred lateral extent of burnt
ground measured perpendicular to a line tracing the alignment of the pipeline prior to failure;
and
• the maximum offset to injury/fatality is the maximum reported or inferred distance to an
injury/fatality again measured perpendicular to a line tracing the alignment of the pipeline
prior to failure.

12
Figure 3.1 shows that in every case the hazard area calculated using the proposed equation is
greater than the actual reported area of burnt ground. In addition, with the sole exception of one
of the incidents reported in NTSB-PAR-87-1, the radius obtained from the hazard area equation
conservatively approximates the maximum lateral extent of the burn zone. Finally, in all cases
the calculated hazard zone radius significantly exceeds the maximum reported offset distance to
injury or fatality.

Note, however, that whereas the interpretation of reported burn areas and burn distances is
obvious, caution should be exercised in interpreting maximum offset distances to injury and
fatality. Given that most of the incidents occurred in sparsely populated areas, the reported
injury and fatality offsets are more indicative of where people happened to be at the time of
failure rather than being representative of the maximum possible distances to injury or fatality
for the incident in question.

Acknowledging the uncertainty associated with interpreting reported offsets to injury and
fatality, the balance of information still overwhelmingly indicates that the proposed hazard area
radius equation provides a reasonable, if somewhat conservative, estimate of the zone of high
consequence.

It is thought that one of the main reasons for the apparent conservatism in the proposed hazard
area model is that it is based on an effective sustained release rate that is consistent with the
assumption of almost immediate ignition. The actual time to ignition for many of the reported
incidents is probably longer (see incident notes in Table 3.1) making the effective release rate
approximation conservative.

13
Date Report Location Incident Damage Maximum Burn Diameter Pressure
Distance (in) (psi)
1969 NTSB-PAR-71-1 near Houston, Texas Rupture at 3:40 p.m. on Burned area 370 ft long by 300 ft 300 ft 14 789
September 9th, wide (all to one side). Houses
explosive ignition 8 to 10 destroyed by blast to 250 ft, heat
minutes after failure. damage to 300 ft, 106 homes
damaged, 9 injuries, and 0
fatalities.
1974 NTSB-PAR-75-2 near Bealeton, Virginia Burned area 700 ft by 400 ft. 30 718

1974 NTSB-PAR-75-3 near Farmington, New Rupture at 3:45 a.m. on Earth charred within a 300 ft 12.75 497
Mexico March 15th, ignition soon diameter circle, 3 fatal injuries
after failure. (within 60 ft offset)
1976 NTSB-PAR-77-1 Cartwright, Louisiana Rupture at 1:05 p.m. on Burn area 3 acres (implies a 200 ft 20 770
August 9th, ignited within radius circle), 6 fatalities (within
seconds about 100 ft offset) and 1 injury.
1982 NTSB-PAR-83-2 Hudson, Iowa 5 fatalities (within 150 ft, less than 20 820
50 ft offset).
1984 NTSB-PAR-86-1 near Jackson, Rupture at 1:00 p.m. on Burned area 1450 ft long by 360 ft Offset 180 ft. 30 1016
Louisiana November 25th, ignition wide (furthest fire extent 950 ft), 5 Distance 950 ft.
soon after failure. fatalities (within 65 ft, 0 ft offset),
and 23 injuries (within 800 ft,180 ft
offset).
1985 NTSB-PAR-87-1 near Beaumont, Rupture at 9:10 p.m. on Burned area 500 ft wide by 700 ft Offset 350 ft. 30 990
Kentucky April 27th, ignition soon long. 2 houses, 3 house trailers Distance 500 ft.
after failure. and numerous other structures and
equipment destroyed. 5 fatalities
due to smoke inhalation in house
318 ft from rupture (150 ft offset), 3
people burned running from house
320 ft from rupture (200 ft offset)
one hospitalized with 2nd degree
burns.
1986 NTSB-PAR-87-1 near Lancaster Rupture at 2:05 a.m. on Burned area 900 ft by 1000 ft. 2 Offset 700 ft. 30 987
Kentucky February 21st, ignition houses, 1 house trailer and Distance 800 ft.
soon after failure. numerous other structures and
equipment destroyed. 3 people
burned running from house 280 ft
from rupture (requiring
hospitalization), 5 others received
minor burn injuries running from
dwellings between 200 and 525 ft
from rupture (250 ft offset).
1994 NTSB-PAR-95-1 Edison, New Jersey Rupture at night on Burned area 1400 ft long by 900 ft Offset 720 ft. 36 970
March 23rd, ignition wide. Fire damage to dwelling units Distance 960 ft.
within 1 to 2 minutes up to 900 ft from rupture, dwelling
after failure. units at 500 ft and beyond caught
fire between 7 to 10 minutes after
failure, no fatalities but 58 injuries.
1994 TSB Report No. Maple Creek, Rupture at 7:40 p.m. on Fire burn area 21.0 acres (8.5 42 1207
P94H0003 Saskatchewan February 14th, ignition hectares).
soon after failure.
1994 TSB Report No. Latchford, Ontario Rupture at 7:13 a.m. on Fire burn area 11.8 acres (4.77 36 1000
P94H0036 July 23rd, ignition soon hectares), heat-affected area 18.6
after failure. acres (7.52 hectares).
1995 TSB Report No. Rapid City, Manitoba Rupture of 42 inch line at Fire burn area 48.5 acres (19.6 42 880
P95H0036 5:42 a.m. on July 29th, hectares), heat-affected area 198
ignition soon after failure acres (80 hectares).
leading to rupture and
fire on adjacent 36 inch
line at 6:34 a.m.

Table 3.1 Summary of relevant North American pipeline failure incident reports.

14
4. REFERENCES

API 1990. Guide for Pressure-Relieving and Depressuring Systems. American Petroleum
Institute, Recommended Practice 521, Third Edition, November.

Bilo, M. and Kinsman, P.R. 1997. Thermal Radiation Criteria Used in Pipeline Risk
Assessment. Pipes & Pipelines International, November-December, pp. 17-25.

Crane Co. 1981. Flow of Fluids through Valves, Fittings and Pipe, Metric Edition - SI Units.
Technical Paper No. 410M, Crane, NY, USA.

Eisenberg, N.A., Lynch, C.J. and Breeding, R.J. 1975. Vulnerability Model: A Simulation
System for Assessing Damage Resulting from Marine Spills. Environmental Control,
Report CG-D-136-75, Rockville, MD, USA.

Hill, R.T. and Catmur, J.R. 1995. Risks from Hazardous Pipelines in the United Kingdom.
Health and Safety Executive Contract Research Report No. 82/1994.

Hymes, I. 1983. The Physiological and Pathological Effects of Thermal Radiation. Systems
Reliability Directorate, Report SRD, R275, Culcheth, Warrington, UK.

IGE 1993. Steel Pipelines for High Pressure Gas Transmission. Recommendations on
Transmission and Distribution Practice IGE/TD/1 Edition 3 (Communication 1530), The
Institution of Gas Engineers, London, UK.

Knowles, A.E., Tweedle, F. and van per Post, J.L. 1978. The Background and Implications of
IGE/TD/1 Edition 2. Gas Engineering and Management, July, p. 247.

Lees, F.P. 1996. Loss Prevention in the Process Industries: Hazard Identification, Assessment
and Control. Second Edition, Vol. 2, Butterworth-Heinemann, A division of Reed
Educational and Professional Publishing Ltd., Oxford, UK.

Technica 1988. Techniques for Assessing Industrial Hazards: A Manual. World Bank
Technical Paper Number 55, The International Bank for Reconstruction and
Development, The World Bank, Washington, DC, USA.

TNO 1982. Safety Study on the Transportation of Natural Gas and LPG by Underground
Pipeline in the Netherlands. Netherlands Organization for Applied Scientific Research,
Ref. No. 82-04180, File No. 8727-50960, translation of a report by the Division of
Technology for Society, commissioned by The Minister of Public Health and
Environmental Hygiene, The Netherlands.

15