Cross-country Pipeline Safety Assessment

a, * b
H.N. Mathurkar  and Dr. A. Gupta
a b
Scientist, HOD, National Environmental Engineering Research Institute (NEERI), Nagpur, India

Abstract

Pipeline safety assessment studies were carried out at the design stage of cross-country Natural

gas pipeline for implementation of various safety measures. The risk level based decisions for the

pipeline are easy for implementation at this stage, since the pipeline is proposed and various

modifications can be easily carried out. The costs incurred on such modifications are also

minimal. Risk level for the pipeline is computed through the use of Event tree analysis,

Consequence analysis, Vulnerability analysis and Individual risk computation. Pipeline design

parameters like pipeline thickness, its depth cover, routing, locations of sectionalizing valve and

site specific surrounding population density, terrain etc. was considered for arriving at the risk

level for the pipeline. Mitigation measures and design modifications were implemented to bring

the undesired risk to acceptable level.

This paper quantitatively evaluates the risk involved based on the actual population data along

the pipeline route for the entire stretch of the pipeline. This pipeline is one of the longest cross

counter pipeline proposed in India, transporting Natural gas.

Keywords   –  Cross - country pipeline, Event tree, Jetfire, Vapour Cloud Explosion (VCE), Flash
Keywords
Fire (FF), Individual Risk (IR)

* Corresponding author: Phone: (0712) 2249885

Fax: (0712) 2249900
email: hemanthn@rediffmail.com

1
 
 

Introduction

Cross-country pipeline transport for Natural Gas (NG) seems to be the most preferred convenient

economical & reliable method. The intrinsic properties of natural gas and its increasing industrial
industrial /

commercial importance has resulted in greater movement of natural gas through the pipelines.

These pipelines are often referred as highways of natural gas transmission. NG is transported at

high pressure in the pipeline at pressures anywhere between 30 ~ 100 bar. This reduces the

volume of natural gas being transported, as well as providing propellant force to move the NG,

through the pipeline

 A 20-inch cross-country pipeline (820 km) is being proposed based on the

t he estimated demand of 

20 MMSCMD. The pipeline passes through 14 districts from the receiving port (Jamnagar) to the

terminating point (Bhopal) as presented in Fig.1. The pipeline pressure varies all along the line

and intermediate booster stations at Ahmedabad, Dahod and Indore (at 309, 481 & 669 kms) are

required to generate enough pressure for gas movement. The 820 km line has 68 sectionalizing

valves dividing the pipeline in 69 sections. In addition to compressing natural gas, a compressor 

station has condensate removal system for removing moisture and higher hydrocarbons liquefied

in the line.

The pipeline is underground buried in trench 1~1.5 m deep covered with earth. The starting point

is Jamnagar pumping station at an elevation of 4 m above MSL and the terminating point at

Bhopal is at 523 meters above MSL. Natural gas have customers on both ends of pipeline, the

producers and processors that input gas into the pipeline and the consumers and local distribution

companies that take gas out of the pipeline. Hence Supervisory Control & Data Acquisition

(SCADA) system for electronic monitoring of pipeline is done for the entire stretch of pipeline from

the control station.

Pipeline failure

Failure of gas pipeline can occur due to a number of different causes such as external

interference (Referred as thirty party interference by EGIG), corrosion, fatigue, ground movement,

material or construction defects. Based on the causes, the failures can lead to leaks or ruptures

(Pluss, Niederbaumer & Sagesser, 2000). The failure mode is determined by the length, depth

and type of defect, and is dependent on the pipe diameter, wall thickness, material properties and

2
 
 

operating pressure. Estimates of failure probability is made from historical data reported by

European Gas Incident Data Group (EGIG).

Line rupture - Due to the pressure at which transmission pipelines are operated, a failure of a

pipeline leads to a turbulent and complex gas release. Following a rupture, or large puncture,

there will be rapid depressurization in the vicinity of the failure. For buried pipelines, the overlying

soil will be ejected with the formation of a crater of a size and shape, which influences the

behavior of the released gas. Depending on the alignment of the pipe ends in the case of a

rupture, the gas will escape to the atmosphere in the form of a jet, or jets. At the start of the

release, a highly turbulent mushroom shaped cap is formed which increases in height above the

release point due to the source momentum and buoyancy, and is fed by the gas jet and entrained

air from the plume which follows. In addition to entrained air the release can also result in

entrainment of ejected soil into the cap and plume. Eventually, the cap will disperse due to

progressive entrainment & a quasi-steady plume will remain. (Acton, Gosse & McCollum, 2002).

Line leaks:
leaks: In case of leaks from buried pipelines, the transient flow will be channelised through

the voids from the overlaying soil. If the leak is sufficiently large then soil will be ejected above the

pipeline. The jet will entrain air as it moves upward and will get disperse depending on the

prevailing wind direction and other meteorological conditions. The concentration of gas till its

Lower Flammability Limit (LFL) is hazardous, as it can catch fire on availability of Ignition source.

The total duration of release and its impact/consequence will depend on how quickly the release

is identified and the sectionalizing valve isolates the pipeline section.

Ignition: The minimum ignition energy for pure Methane at atmosphere pressure is 0.29 mJ, which

is very low when compared to static discharge of 22 mJ (on walking) or Ordinary spark plug

discharge of 25 mJ, hence ignition can occure at any time during the gas release (Crowl &

Louvar, 1990). If it occurs immediately or shortly after, the line ruptures, a transient fireball will

occur. The fireball, which is the result of combustion of the mushroom-shaped cap lasts, typically,

for up to thirty seconds. The heat radiation from such fireball can result in maximum damage to

surrounding population, such fireball can also result in major psychological fear to surrounding

population.

Leaks from pipeline can result in Jet flow, which on ignition can lead to Jetfire / Torch. In areas

where NG is not odorized these Jetfire incidents can occur all of a sudden surprising the resident

3
 
 

population and authorities, when they are least prepared. The levels of incident thermal radiation

can affect people and property in the vicinity of an ignited pipeline release. Casualties can occur 

as a result of receiving large doses of thermal radiation and buildings can be ignited by thermal

radiation directly from the fire or from secondary fires (e.g. from burning vegetation).

In this present case of Cross-country pipeline, hazardous incident outcomes for Full Bore Rupture

(FBR) and leak are sequentially developed through an Event tree. The database of European

Pipeline Incident Data Group (EGIG) representing almost 2 million kilometer year of pipeline

operations was used to determine a failure frequency for this gas pipeline. The failure rate
-4
reported by EGIG for on-shore gas pipeline with design pressure greater than 15 bar is 4.76 x 10  

per km per year.

-5
Full Bore Rupture (FBR) represents 13% of the cases (6.188 x 10 failure /km/yr.) and 87 % of 
-4
the cases represents Leaks (4.14 x 10 failure /km/yr.). These failure frequencies were used in

Event tree for quantifying the p

probability
robability of haz
hazardous
ardous incident outcomes A to F (Fig. 2).

Case A & B 
B 

Fireball and Jetfire scenarios are because of immediate ignition of released gas. The ignition

probability is assumed to be dependent on the population density. The population density around

the first section of the pipeline is given in Table I and Ignition probability values as a function of 

population density is given in Table II.

Thermal radiation from fireball is intense and is inversely proportional to the square of distance

from the center of fireball. The casualties from the fireball will depend on the thermal dose

experienced by surrounding persons. Jetfire thermal radiation will have less effect since majority

of jets will be vertical with very less thermal radiation load on the ground level. Sequential

development of fireball and Jetfire because of Full Bore Rupture (FBR) and pipeline Leaks are
-5 -4
quantified through Event tree as 1.856 x 10 per km. per yr. and 1.242 x 10 per km per yr.

respectively as presented in Fig. 2.

Delayed ignition: Gas release and fire hazards on immediate ignition are often considered as the

safest scenario for flammable chemicals since the flammable material gets consumed in this

process. Delayed ignition can occur in case the released gas is not immediately ignite but finds on

ignition source after the gas has dispersed and its concentration is still in flammable range.

4
 
 

Methane is lighter than air and hence will disperse more rapidly in open terrain on release and

hence delayed ignition for methane can only result in flash fire with no overpressure wave.

Case C & D 
D 

In case the pipeline is passing through congested area with certain degree of confinement for 

vapours, with relatively stable atmospheric condition, then there are chances of explosion. FBR or 

Leaks from pipeline can release maximum gas inventory approximately equivalent to the volume

between two sectionalizing valves. The released heavier gases; propane, butane & pentane can

settle down on ground leading to heavy gas dispersion for these hydrocarbons after initial buoyant

plume release. These hydrocarbons can lead to unconfined or confined vapour cloud explosion

on availability of ignition source. Inventory of only these hydrocarbons are considered for 

computation of overpressure wave generated in case of Vapour Cloud Explosion (VCE). The VCE

chances are considered as only 10% and remaining 90% cases results in flash fire incidents,

hence VCE frequencies VCE1, VCE2 for FBR and Leaks are 3.465 x 10 -6 per km per
per yr. and
-5
2.318 x 10 per km per yr. respectively.

Case E & F 
F 

The released gas (after FBR / Leaks) can disperse in open atmosphere and the dispersion zone/

area will depend on quantity of release, the atmospheric conditions (wind speed, atmospheric

stability, temperature, pressure etc.) and terrain conditions. The gas cloud is hazardous till the

concentration is between its flammability limits. In most conditions the Lower Flammability Limit

(LFL) is considered as the limiting range for dispersion zone / impact area computation. Any

delayed ignition within this zone will invariable result in Flash Fire (FF) incident. The thermal

radiation load within this zone can be fatal, hence this entire LFL zone is considered as 100%

fatality zone for FF incident. FBR and Leaks both can lead to FF. It has been assumed that 80%

of the times delayed ignition takes place. Out of these only 10% cases leads to VCE because of 

confinement of vapous and remaining 90% cases leads to Flash Fire (FF), hence FF frequencies
-5 -4
FF1, FF2 for FBR and Leaks are 3.119 x 10 per km per yr. and 2.087 x 10 per km per yr.

respectively.

5
 
 

Consequence and Impact analysis 

analysis 

Determining the impact of each incident requires two steps. First a model estimates a physical

concentration of material or energy at each location surrounding the facility for example, radiant

heat from fire, overpressure from an explosion and concentration of flammable material till LFL

limit in atmosphere. A second set of model estimates the impact that this physical concentration

of material or energy has on people. Accordingly a threshold for fatality levels for 100 % and 1 %

has been considered as:

 100 % fatalities  Fireball radius / FF zone or 

or distance
distance for 0.3 bar overpressure wave
 1 % fatality  Distance for 12.5 kW/sq.m. heat load or 0.1
or 0.1 bar overpressure wave

The impact evaluated in terms of damage distances for these two fatality levels are presented in

Table III for the first section of the pipeline.

Risk Estimation

 All the 69 pipeline sections were considered for evaluating the damage distances
d istances for each of the

six cases A to F after considering the holdup volume in each section. The population density

adjacent to the pipeline was used for computing the risk level. The actual fatalities from each of 

these cases A to F were estimated as Individual risk (F-X) curve. Individual risk (IR) at any

specific point is given by the summation of all risk (Hendershot, D.C. 1997) at that point and is

mathematically expressed as

IR =  Fi x Pi
where:

n = the total number of incident outcome cases considered in the analysis

Fi 
Fi  = frequency of incident outcome case i , (per year)

Pi = probability that incident outcome case i will result in a fatality at that location
location  

Individual risk has been computed considering all the six cases A to F. IR values for the first

section of the pipeline is presented in Table IV. Accordingly IR values were computed for each of 

the 69 sections of the pipeline. The variation of the IR values for the entire pipeline along the axis

and along the perpendicular distance is presented in Fig. 3 as IR surface.

Conclusion

6
 
 

The planning & construction of the cross-country pipeline has to take into account the various

population densities all along the length of pipeline.

 According Risk surface curve Fig. 3 was used for presenting the risk level from a cross-country
cross- country
 –4
 –4
pipeline to the surrounding population. Areas having risk level more than 10 fatalities per year 

(Fat./yr.) was considered unacceptable and was further investigated by weighing its cost affective

risk reduction and public opinion in that area. The various risk mitigation measures considered

(Thomson, 1987) at specific sections of the pipeline as per increasing cost are:

 Increasing depth of cover 

 Concrete slab covering
 Sleeving
 Increasing pipe wall thickness
 Reducing the distance (inventory) between sectionalizing valves

Risk surface can be estimated once again after the decision for risk mitigation implemented and

design for the various pipeline sections frozen. The exercise can be repeated till all the pipeline

sections are brought in acceptable zone. The authorities can also establish an objective for those
non-acceptable areas were balance of interest from public, private and statutory organizations

can be considered before laying the pipeline.

Acknowledgement

The authors express their sincere gratitude to the Director, NEERI, Nagpur for according

permission to publish this paper. The cooperation and assistance provided by Mr. Prashant

Lokhande and Ms. Saroj Hadge is also gratefully acknowledged.

Reference

1. Pluss, C., Niederbaumer, G., & Sagesser, R. (2000). Risk assessment

assessment of the Transitgas  – 
 – 
pipeline. Switzerland: SKS Ingenieure AG, Consulting Engineers, Zurich.

2. Acton, M.R., Gosse, A.J. and M McCollum,

cCollum, D (2002). Pipeline risk assess
assessment
ment new
development for natural gas and hydrocarbon pipelines. Advantica, Loughborough, UK

3. Crowl, D.A. & Louvar, J.F (1990). Chemical process safety fundamentals with applications. 
applications. 
New Jersey: Prentice Hall

4. Hendershot, D.C. (1997) A simple problem to explain and clarify the pri principles
nciples of risk
calculation”, Bristol, P.A : Rohm and Hass Company Engineering division. 
division.  

5. Thomson, J.R. (1987). Engineering safety assessment an introduction. England: Longman

Scientific and Technical

7
 
 

Fig. 1 : Proposed route for the 820 km. natural gas pipeline

8
 
 

Immediate Delayed Confinement

Ignition Ignition of vapours
-5

A fire ball 1.856 x 10

-
30% B Jetfire 1.242 x 10

-
1. FBR - 6.188 x 10
- 10% C VCE1 - 3.465 x 10

-4 80% D VCE2 - 2.318 x 10

-
2. Leak - 4.14 x 10

-
90% E FF1 - 3.119 x 10

70% F FF2 - 2.087 x 10

-4

20%
 
NO HAZARD

Fig. 2 : Event tree

tree for hazard
hazardous
ous incident ou
outcomes
tcomes for NG pipeline
pipeline 

9
 
 

10-4 

Fig. 3 : Individual risk surface for the Natural Gas pipeline 

10
 
 

Table I : Population density around the first section of the cross-country pipeline

Distance from pipeline Population density

0 - 200 meters 47.44 persons / sq.km.
200 - 400 meters 51.09 persons / sq.km.
400 - 800 meters 56.57 persons / sq.km.

Table II : Ignition Probability as function of Population density

Ignition Probability Population density

0.85 Less than 200 persons / sq. km
0.90 200 to 400 persons / sq. km
0.95 More than 400 persons / sq. km.

Table III : Damage distance for various failure cases for the first section of the pipeline
Damage distance in meters for 
Cases Consequence Effect
100 % fatalities 1% fatality
Case A –
A – Pipeline rupture Fire ball Thermal radiation 200 450
Case B –
B – Pipeline leaks Jetfire Thermal radiation - 20
Case C –
C – Pipeline rupture VCE
VCE –
 – 1 Overpressure wave 70 210
Case D –
D – Pipeline leaks VCE –
VCE – 2 Overpressure wave 40 115
Case E –
E – Pipeline rupture Flash fire –
fire –1
1 Dispersion & fire 681 -
Case F –
F – Pipeline leaks Flash fire –
fire –2
2 Dispersion & fire 330 -

Table IV : Individual risk at various distance from the pipeline for the first section

Individual Risk Distance from the pipeline

-4
7.17 x 10 Fat./ yr. 100 meters
-4
3.19 x 10 Fat./ yr. 300 meters
-4
1.94 x 10 Fat./yr. 600 meters

11