SPE 101422

Can the Existing F&G Detection System Provide Safe Guard Against All Possible Gas
Releases? (Case Study)
Ashraf E. Shabaka, ZADCO
Copyright 2006, Society of Petroleum Engineers
This paper was prepared for presentation at the 2006 Abu Dhabi International Petroleum
Exhibition and Conference held in Abu Dhabi, U.A.E., 58 November 2006.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than
300 words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
The Fire and Gas detection systems in many of the
existing facilities were designed in a conventional
method without software modeled design and therefore
their performance are questionable. This is due to the
fact that at the time when designed, the dispersion and
fire were not yet widely implemented.
Since the major accidents are unlikely to occur and it is
therefore uncommon to revisit and verify the adequacy
of the design of these systems.
This paper discusses the importance of revisiting the
design of the existing fire and gas detection systems and
its rule in saving our facilities from major losses.
Introduction:
ZADCO top priority goal is to achieve the required
production target within 25 years plan, while maintaining
their successful record of safety of personnel and
assets.
ZADCO Central Complex, presently is handling process
streams having very low H2S content. In near future, gas
breakthrough in the field will result in H2S contamination
of the process streams at Central Complex. The H2S
concentration is predicted to reach 1.0 mol% in process
streams by year 2020.
The existing facilities of Upper Zakum were designed for
sweet crude/gas, however new installation of sour gas
injection and gas lift facilities together with the likelyhood
of gas breakthrough thus a possibility of sour gas
emission. For mitigation, a Quantative Risk Assessment

Study was carried out resulting in identifying the

necessity for a new toxic gas detection system.
Therefore a complete F&G study was conducted to
identify the requirements of the toxic gas detection
system plus assess the performance of the existing gas
and flame detection system on Central Complex.
A detailed performance target criteria for each
processing area was developed to define specific
performance requirements for the fire and gas detection
(both flammable and toxic) systems in deifferent
processing areas of the Central Complex.
The performance-based selection criteria for the overall
Fire and Gas detection system take the following into
consideration:

Type of detection
Response time
Coverage area
Maintenance requirements
Environmental conditions
Performance of the complete loop

The dispersion results identified the requirements of the

new toxic gas detection system, and highlighted that the
existing flammable gas detectors are inadequate and
recommended additional flammable gas detectors to the
existing gas detection zones.
The study also identified that in some areas of the plant,
despite not containing major hazard source but is
considered a critical location since major gas releases
from adjacent decks can subject those zones to
flammable gases.
Furthermore, fire events modeling was performed for
representative fire events on the platform to define the
basis for fire detector locations in the various areas of
the plant.
The fire events identified that the existing frangible bulbs
are very slow and insensitive to most fire events, this
cannot therefore be relied upon to detect small-medium
fires before they escalate and the present arrangement
is unable to meet the performance target criteria.

www.petroman.ir

SPE 101422

areas handling flammable gases and liquids, both point

source and grid concept have been proposed.

Gas Dispersion Modeling

As undicated above, dispersion modeling was performed
for Small and Medium category of failure events from
main process equipment items on each deck/level of the
plant to accomblish a comprehensive basis for locating
hydrocarbon and toxic gas detectors.
Sensitivity runs for change in relative humidity from 7090% for gas release from gas manifold and flashing
liquid ( oil manifold) was performed and the results
highlighted that there is no significant difference in the
downwind dispersion distances were observed.

A grid concept is utilized for locating flammable gas

detectors in areas where accumulation of flammable gas
cloud can occur. This concept is also utilized in areas
where numerous gas/ liquid release sources are present.
The advantage associated with a grid concept is that it
enhances the chances of a gas release getting detected
irrespective of the source of release and the prevalent
wind direction. The gas detector grid spacing was
defined taking into consideration the dispersion modeling
results and the area layout.

Similarly sensitivity runs for affect of change in ambient

temperature from 20 to 45 deg C (night/day
temperatures) was performed for gas release from gas
manifold and flashing liquid (oil manifold). The results
highlighted that there is no significant difference in the
downwind dispersion distances was observed.

Point source method was also utilized in areas where

potential release sources are distinct and the gas cloud
is expected to disperse to safe toxic concentrations after
traveling large distances. In such cases, the detectors
have been placed near potential release sources and
also along escape ways in the area.

Dispersion was performed for obstructed and

unobstructed/ impinged scenarios. Unobstructed
dispersion indicated that gases released from a
pressurised process section, getting diluted faster due to
the jet momentum associated with the initial release. In
such releases, the gas cloud does not lose its
momentum and the cloud width of such releases is
relatively small. While in case of obstructed/ impinged
dispersion, the initial momentum of the gas release is
lost due to impingement with nearby
structure/equipment; following which the gas cloud
disperses passively i.e. result in longer downwind
distances and wider gas cloud.

The gas detectors were proposed to be located so as to

result in confirmed high level gas detection following a
gas release in an area i.e. at least two detectors should
lie within the 60% LFL for flammable gas cloud or 20
ppm gas concentration for H2S gas cloud.

Dispersion results were utilized to define the basis for

gas detection in different areas of the plant. The
dispersion was performed for gas breakthrough
scenario, when H2S concentration in process stream can
be as high as 1 mol%.
A small impinged gas release has been considered for
determining the toxic gas detection requirements, since
such releases can subject significant areas to harmful
H2S concentrations. Small unimpinged releases were
not considered as they are expected to disperse in short
distances to safe concentrations, thus have low hazard
potential.
As the objective of flammable gas detection system is
explosion prevention, major gas release events were
considered to define the gas detection requirement in an
area. This approach has been utilised for all decks,
except one deck (due to congestion), since likelihood of
gas accumulation on decks following small releases is
considered low due to the fact that the platform decks
are relatively open/ less congested areas.
The above approach was utilized to define specific gas
detection requirement for each area. For gas detection in

Fire Events
Where the needs for fire detection have been identified,
the required performance of the fire detection system
has to be specified considering the predicted fires and
the consequence of those fires. The performance
specification (Grading) defines fire size and response
time thresholds for alarm and action(s).
Fire performed for representative credible failure
scenarios on the plant indicated that the RHO
associated with such small fire ranges from 60 kW to
650kW, thus such fire events can lead to escalatory
events.

PERFORMANCE CRITERIA RULE SETS

The categorizations of the hazards considered for
defining the performance criteria for the plant were
defined as follows:

FLAMMABLE GAS DETECTION

Flammable gas detection is provided in the facility with
the prime objective of minimizing the likelihood of a
vapour cloud explosion and its effect on assets and
personnel. The damaging potential or the intensity of a
gas cloud explosion is dependent on the following major
parameters:

www.petroman.ir

SPE 101422

Confinement / Congestion The greater the

confinement/ congestion, the higher the
likelihood of experiencing high explosion
overpressures. In general, an area having a
blockage ratio of 0.3 and above is considered
confined/congested for example- if in a module
the floor and the roof are plated ( 2 sides), with
the other four sides open, the blockage ratio is
still 0.3.
Gas Cloud Size- A large gas cloud has much
higher likelihood of resulting in an explosion than
a small cloud. However it should be noted that
this may not be true in case of a small gas
release in a highly confined/ congested area. In
such areas due to limited natural ventilation
there is less dilution of the gas cloud and gas
concentration builds up, which on ignition can
explode.
Gas type Methane has a lower explosion
energy compared to Propane and Butane. Thus,
the damage potential associated with propane/
butane is expected to be higher than that
resulting from similar gas cloud of methane.
Location and type of ignition source- Location
of the ignition source within or on the periphery
of the gas cloud can significantly change the
resultant explosion overpressures. Ignition
source energy also influences the resultant
explosion overpressures, though to a smaller
extent.

Based on the above explosion characteristics, there are

only two parameters i.e. gas cloud accumulation in a
confined/ congested area and ignition sources, where
flammable gas detection can effectively control/ minimise
the explosion potential of a release. Therefore, gas
detection should be provided in areas where gas
accumulation is possible and near ignition sources to
minimize the likelihood of a gas cloud explosion.
In addition, for effective gas detection in an area the
dispersion characteristics of a release i.e. type of release
(pressurized or atmospheric release), gas properties/
physical conditions, air movement in the area, should
also be taken into consideration. For example, a
pressurized gas release will disperse quickly to safe
concentrations in short distances due to initial
momentum associated with the release. While in case of
passive dispersion, the gas cloud disperses slowly and
over longer distances along the prevalent wind direction.
Gas properties such as molecular weight and physical
conditions of the release also play a major role in
defining the extent of dispersion of the gas cloud.
Therefore, flammable gas detectors should be located
taking into consideration the following:

Dispersion characteristics of a release

Confinement/congestion in the area
Potential ignition sources present in the area

A grid concept for detection of flammable gas cloud is

preferred as it will result in gas detection irrespective of
the direction of the dispersing gas cloud.
HYDROCARBON GAS DETECTION CRITERIA
To define the specific rule sets for the flammable gas
detection system on the plant, the results of the
dispersion study were critically analysed.
It was observed from dispersion results that small
releases are expected to disperse to safe concentrations
in short downwind distances. Furthermore, most of the
plant deck areas are relatively open i.e. less congested
except for one deck. Thus gas cloud accumulation and
risk of potential gas cloud explosion, in general, is low.
On such open decks, only major gas releases can
potentially result in significant gas cloud formation and
accumulation due to the fact that the gas release rate is
much higher and that such releases are likely to loose
their initial momentum due to impingement with the
nearby structure/ equipment and disperse passively.
Therefore, for relatively open decks and areas, where
there is adequate ventilation/ air flow, the flammable gas
detector grid shall be defined based on major (medium
hole size) gas releases that can occur in that area.
The flammable gas detectors are to be placed in a grid
to ensure that such major releases are detected
irrespective of the direction of the prevailing wind and
appropriate control action initiated i.e. confirmed
detection results.
In case of confined/ congested areas, unobstructed
dispersion of even small release is unlikely. The gas
cloud formed following a release is expected to
accumulate, due to congestion/ limited air flow and
poses an explosion hazard. Therefore flammable gas
detectors grid should be determined based on dispersion
modeling results for small impinged releases.
In addition, gas detectors shall be provided near
potential ignition sources present on the platform to
reduce the likelihood of ignition of a gas cloud.
TOXIC GAS DETECTION
As an example, the majority of the hydrocarbon process
streams on the Central Complex will contain H2S,
following gas lift and gas breakthrough in the field. The
H2S concentration in the streams ranges from 400 8800
ppm, with majority of the gas streams having
concentrations in the range of 1100-1400 ppm.

www.petroman.ir

SPE 101422

The prime objective of toxic gas detection system is to

provide protection to personnel from toxic hazards by
prompt detection and initiation of alarm/appropriate
control actions.
As observed from toxic gas results, small unimpinged
gas releases are expected to dilute to safe
concentrations in short distances and therefore
personnel are not expected to be adversely affected by
such releases. However, in case of small impinged
releases the gas cloud is expected to travel relatively
large distances, covering significant area of the deck, to
dilute to safe concentrations.
As personnel under such situations can be exposed to
harmful H2S concentrations, therefore toxic gas
detectors should be located so as to detect such
releases.
For toxic gas detection in manned areas of the Complex,
both a grid concept and point source method are
proposed to be utilized. A grid concept is to be utilized in
where a number of release sources are present in the
area frequented by personnel. The grid shall take into
consideration the dispersion study results. The location
of the detectors and size of the grid shall be such that at
least 2 detectors detect a release on the platform.
A point source method is utilized in areas where the leak
sources are few and distinct. In areas where the harmful
H2S concentrations following a release can impact large
distances, a point source method is utilized with
additional coverage along the nearby escape routes and
EER (Escape, Evacuation, Rescue) facilities.
In addition, the following specific areas essential for EER
activities and where a large population is expected to be
present, shall be provided with toxic gas detection
coverage:
Along Escape routes including staircases
EER facilities i.e. lifeboat, liferaft embarkation
and muster area
Air intakes and entrances of manned/ normally
unmanned buildings
It should be noted that since process streams are both
flammable and toxic, a toxic gas detection and alarm
should be treated as detection of a leak in the process
area.

FIRE DETECTION
Where the need for fire detection has been identified, the
required performance of the fire detection has to be
specified considering the predicted fires and the
consequence of those fires. The performance
specification (Grading) defines fire size and response
time thresholds for alarm and action(s).

GRADING
The grading of an area defines the sensitivity, response
time and availability requirements of the fire detection
required in that area. For convenience of design,
operation and maintenance, the detection requirements
have to be set using thresholds of minimum sensitivity,
time and probability of detection. These are applied
where appropriate to trigger the required risk reduction
measures in that area.
The grading of a hydrocarbon handling area is primarily
based on the assessment of the damage potential
associated with a fire event should it occur. The damage
potential associated with an event is assessed in terms
of Radiant Heat Output (RHO) of a fire and value of
assets in the area. Accordingly the alarm and control
actions are defined based on the grading.
Radiant Heat Output is that portion of a fires
combustion energy delivered as radiant heat and is
typically between 25 to 35% of the calorific value of the
fuel consumption rate. It is expressed in Kilowatt (kW).
For estimating the RHO from the radiation intensity
(kW/m2) of a flame, the area to be considered is the
apparent area of the flame to a bystander and not the
base area.
Thus RHO for a much smaller size jetfire (high radiation
intensity) can be same as for a much larger poolfire
(lower radiation intensity), which gives a good picture of
the damage potential associated with a fir event.
The following classification/ grading of areas are
considered for defining the performance requirement for
fire detection system:
Grade A is used where the items, or area, being covered
have higher than average escalation potential. Grade A
requires that the system detects, alarms and initiates
control action in response to the minimum size of fire
which it has been judged to have a significant probability
of causing unacceptable loss or damage (this value has
been set at 10 kW Radiant Heat Output).
Grade B is used where;fires of the specified size,
although very undesirable, are not likely to cause
unacceptable loss or damage within the time taken for
detection and effective control actions.
Rapid escalation, from fires at the alarm threshold, is
unlikely and the effect on production due to fire damage
of equipment is small.
Grade B requires that the system detects alarms and
initiates control action in response to the minimum size
that has been judged to have a significant probability of
causing unacceptable loss or damage. This value has
been set at 50 kW RHO.
Grade C is used for areas that are of low risk and where
in the unlikely event of a large fire losses will be of low
significance.

www.petroman.ir

SPE 101422

Grade C requires that the system detect a size of fire,

significantly larger than required for Grade A or B, which
has been judged to have a significant probability of
causing unacceptable loss or damage. This value has
been set at 100 kW RHO. Grade C detection will initiate
control action in response to a 250kW RHO fire.
Hydrocarbon Risk Areas at the plant
Based on the above fire area grading, the processing
areas at Central Complex will be graded as grades A,
B, C.
It should be noted that fire events modeling conducted
for representative credible failure scenarios on the plant
indicate that the RHO associated with such small fire
ranges from 60 kW to 650kW. Thus in terms of alarm
and control actions, Grade A and B will essentially be
the same. Therefore, these grades will be used only to
define the escalation potential associated with a fire
event in an area.

REFERENCE DOCUMENTATION
1.
HYSIS file, WITH GBT IN 2010-ONE
MOLE% H2S.HSC, provided by the
COMPANY on 4th August 2004.
2.
Quantitative Risk Analysis of Toxic Gas lift/
Gas Breakthrough for Upper Zakum Field
Facilities by RMRI, Rev 1.
3.
Assessment of Toxic Gas Risk for Upper
Zakum Field Facilities by RMRI, Rev 0.
4.
Micropack Report, ZADCO Zakum Central
Complex Gas Treatment plant Fire & Gas
Detection Study, MEL ZCC GTP 01 Rev 1.1.
5.
Dispersion Study Report, 9503E-001-RT1900-02 Rev.0.
6.
Central Complex Safety Manual
7.
Central Complex, Fire& Gas System Matrix,
Philosophy & Layout.

The fire hazards in the processing facilities have been

graded as Grade A or B depending upon the escalation
potential associated with a fire event. As the RHO of a
potential fire event involving small release is higher than
60 kW, the control action will be required in case of all
fire events involving processing facilities i.e. confirmed
fire detection by at least 2 detectors.

Results
The gas dispersion proved that the existing flammable
gas detection system was inadequate and the
distribution of the existing detectors can not meet the
requirements of the performance target criteria. The
design of the existing gas detection system was based
on point source concept in which a confirmation of a gas
release may not occur with the current locations of the
existing detectors. Therefore a gas release may not be
detected by two gas detectors located on the same skid
and may be detected by adjacent detectors for other
package. This resulted in redefining the gas detection
zoning on the basis of grid concept. The advantage
associated with the grid concept is that it enhances the
chances of a gas release getting detected irrespective of
the source of release and the prevalent wind direction.
Fire modeling performed for representative small and
medium release events indicated that, such fire events
can lead to escalatory hazards and since all the grading
for most of fire event are either A or B (typicaly, small
and medium releases) which cannot be detected by the
existing frangible bulbs/tubing
It was therefore recommended to install optical flame
(UV/IR) detectors which can provide a rapid and reliable
detection of hydrocarbon fire hazards.

www.petroman.ir