Confidential to EPSC Members Report No 31

Gas Detection
Confidential to EPSC Members

Gas Detection

Gas detectors can provide a valuable contribution to process safety. They can be used to trigger
alarms and provide early warning of loss of containment to ensure personnel safety. Yet, gas
detectors cannot prevent leaks or indicate what action should be taken. This report aims to provide
guidance for both installation and use of Fixed Gas Detectors Systems (FGDS).

This report is the outcome of a series of EPSC workshops run since 2006 under the title of Design for
Emergency Response. EPSC decided to narrow the focus of the group in early 2008 which resulted in
a workshop on gas detection in March 2008.

© EPSC 2008
The information held in this report is given in good faith and belief in its accuracy, but does not imply
the acceptance of any legal liability or responsibility whatsoever by the European Process Safety
Centre or by the authors, for the consequence of its use or misuse in any particular circumstances.

Any enquiries about this report, or other EPSC matters, should be addressed to Mr Lee Allford,
Manager – EPSC Operations
Table of Contents

1. Problem Statement and Main issues………………………………….……3

2. Architecture of a FGDS………………………………………..…………….…...3

3. Position of the FGDS in the safety management system….………4

4. Technical Standpoint………………………………………….…….……………..6

4.1 Periodic risk assessment………………………………………..….…………….6

4.2 Placement of detectors……………………………………………….…………..7

4.3 Alarm and SIL requirements………………………………………….…………8

4.4 Computer Simulations…………………………………………………….……...9

5. Managerial Aspects…………….……………………………..……………………10

5.1 Training………………………………………..…………………………………………12

5.2 Emergency response strategy…………………………………………………12

5.3 Case Study………………………………………..…………………………………...14

Further Guidance when implementing a FGDS………..………………….......17

2
 1. Problem Statement & Main Issues

Implementing and designing a Fixed Gas detection system is a complex task. Making the most of it is
another problem. In most cases, the installation and maintenance work is delegated to third parties.
This creates a situation in which there is little ownership for the system within the owner company.
This lack of awareness from personnel and process safety managers can result in several adverse
situations. First, this creates a false feeling of safety from the staff. Then, if the FGDS is not well
integrated in the safety management system, emergency response strategies may just be
unworkable. If some companies are able to show some organizational awareness of their FGDS can
they still ensure that actions to be taken on hearing alarms will enable risk to be reduced? Eventually
to what extent can it be argued that a safety management system is able to accommodate the
FGDS?

2. Architecture of a FGDS

5. Automatic 6. Central signal 7.Operator

Action visualisation Action

4. Central Signal Processing

Area 1 Area 2 Area 3

3. Signal ...
Transmission

2. Local Signal
Processing

1. Local Detection

Fig. 1 Architecture of a Fixed Gas Detection System

3
In hardware terms a typical fixed gas detection system is made up of a sensor, a control unit, an
alarm and connecting cables.

There are three main types of system:

a. Sensor & control unit both situated within the hazardous area, the two parts may be
combined into one unit or supplied separately.
b. Separate sensor and control unit, usually with the sensor within the hazardous area and the
control unit outside the hazardous area.
c. Sampling apparatus with one or more sample points in the hazardous area, each point is
drawn to the sensor via a sample line.

3. Position of FGDS in the safety management system

Fixed gas detection systems are one risk reduction measure to protect installations within e.g. the
chemical and petrochemical sector from major accidents. There are many other technical and
organizational barriers of defence being applied to prevent damage to life and / or installations. This
chapter provides a broad view of common “barriers of defence” to clarify the position of fixed gas
detection systems within this context.

Fixed gas detection systems come into action when several other layers of protection have already
failed (see illustration below). A FGDS cannot prevent the occurrence of a gas release. It is designed
to warn the personnel of a site as early as possible about a gas release taking place and it can start
automatic actions if designed so. Consequently, according to the author´s view a fixed gas detection
system should be considered as part of the plants´ emergency response.

Gas Detectors and KPIs:

Some Companies include the activation of safety devices as leading or

lagging indicators for Process Safety. Such activations are reported
through their IT system. It seems that considering the triggering of gas
detectors as a Process Safety Indicator is a smart way to raise
awareness of the fixed gas detectors network among personnel.

4
Fig. 2 Position of FGDS in terms of layers of protection

However whether FGDS are viewed as a prevention or mitigation measure depends on the point of
view of an individual company. Some organizations would deem that FGDS are an element of Plant
Emergency Response whilst some may view FGDS as a layer of Critical Alarm and Operator
Intervention. The latter is not easy to justify.

5
The following illustration provides a more specific and detailed overview of the variety of existing
risk reduction measures. They are part of a fire and explosion hazard management (FEHM) as
proposed by the Energy Institute 1.

Training Hazard elimination Hazard Alarm systems

substitution
Emergency shutdown
Process design Ignition Passive
source control protection
Access control
Plant layout
Fire and Gas
detection First-aid fire equipment
Inventory Control

Fire protection systems

Area classification FEHM
Fire brigade
Escape routes
Process sparing
Mobile equipment
Salvage
Spillage Insurance
control
Plant maintenance Contingency planning

Operating practice Mutual aid

Process reinstatement

Fig. 3 Mitigation and Risk Reduction Measures

4. Technical Standpoint

4.1 Periodic risk assessment

A periodic risk assessment may be implemented using Layer of Protection Analysis (LOPA)
considering FGDS. Internal audits are often initiated and supervised from corporate level of the
companies. External audit can also be carried out.

Other possibilities include scenarios focusing specifically on fire emergency response. A multi-
disciplinary group including among others, representatives from the fire department carry out this
exercise of which the evaluation of gas detection systems is one crucial aspect.

Overall, practices on risk assessment differ from company to company. The attention paid to
detailed FGDS issues varies considerably.

6
 4.2 Placement of Detectors

Among the factors considered When installing sensors, the following requirement should be met:

when implementing fixed gas

detection system:  Detectors are preferably located in the vicinity of higher
probability release points such as pumps, compressors, seals,
- Atmospheric conditions vents etc. (hot-spot strategy).
 Understand thhe response time (from leak to detection and
- Nature of Gas (density)
alarm) of the new detectors (response time)
- Turbulences created by  Ensure the Reliability of the detectors and the control
major equipment systems (failure modes)
 Ensure that “enough” detectors are installed to cover the
- Topography area and meet redundancy requirements.
- Diffusion rates  Ensure that the new detectors can measure all flammable
- Electromagnetic gases present at the location (adequacy of detection)
interferences
- Maintenance matters In particular with respect to toxic substances the following criteria
- Functional redundancy should be mentioned:
- SIL redundancy
- Mitigation Systems Fence-line monitoring is crucial to make sure that toxic gas releases
- Vibrations leaving the site are monitored, even on low level. Toxic gas releases
- High pressure leaks can form a threat to the human beings of the community.

- Manufacturers’
Staff have to carry personal monitoring equipment when entering
specifications the plant to make sure that low level exposure can be detected and
legal threshold can be kept. Numerous low-level gas detectors are
... distributed in the plant, much closer to potential leaks.

For the monitoring of high level toxic gas releases selective detectors
are installed. Generally, they are installed close to high risk release
locations. A two out of three voting will be applied for process trips to safe state. That means three
detectors will be clustered at an identified risk location. If two of the three measure toxic gases
above a defined value the plant / unit trips. Normally a single detector positive signal is used for
alarm purposes.

Cross-sensitivity of the detectors is an important selection criterion in the context of toxic

substances. Before installation it has to be investigated thoroughly whether or not the plant’s

7
chemicals can create cross sensitivity or poisoning of the detectors. Cross-sensitivity might for
example lead to false alarms and thus to unsafe situations.

Concerning the number of detectors some companies have developed technical rules for this issue
which partly depend on local judgment. Some software has also been created in this view to map
out risk. Shell has developed its own software ‘Shell Solutions mapping project’ which optimizes the
positioning of sensors and computes their required number.

In some organizations, calculation of the number of detectors is the task of the laboratory specialist
in cooperation with the fire brigade to determine the numbers of gas detectors needed.

Apart from this, incident and accident investigations can be used to evaluate whether further
detectors can be necessary.

When implementing a network of sensors companies have a different approach on who should
design the FGDS:

 Project teams and instrumentation engineers.

 Project teams by approval during the management of change procedure.
 A multidisciplinary team involving staff from plant production, the HSE and Instrumentation
& Analyzer department.
 Analyzer and HSE department.

4.3 Alarms and SIL Requirements

LOPA studies generally indicate that we should rely more on automatic trips to safe state than
manual alarms requiring operator intervention. An equivalent failure on demand to ‘Safety integrity
Level 1’ is applied for alarms responded to by the operator. The underlying reason is human actions
high probability of failure. Human error is the weakest part in the steps necessary for an alarm and
response from an FGDS alarm so that a higher SIL for the components of an FGDS will not result in
higher reliability (Lower Probability of Failure on Demand). The additional cost will not result in
improved reliability.

The practices concerning automatic trip to safe state systems vary between applying SIL 1 and SIL 2.
However, assigning a SIL to a FGDS is not in compliance with the definition of an Independent
Protection Layer.(within LOPA methodology). Specifically ‘A layer of protection which will interrupt a
hazardous scenario irrespective of the performance of another layer of protection or the initiating
failure.’

8
 4.4 Computer simulations

Sophisticated computer programs and simulations are available to predict the effects of fires and
explosions and to assess whether incident escalation is possible.

Types of models available at the moment can for instance simulate pool and jet fires, explosions and
the dispersion of gas. The latter is of importance with respect to fixed gas detection systems. Gas
dispersion modeling can for instance be used for the following tasks:

 Estimation of likely size, composition and flammability characteristics of gas releases.

 Direction and speed of travel of the gas cloud
 Supporting the determination of locations for flammable or toxic gas detection since it can
simulate gas concentrations at specific locations.
 Determination whether or not a flammable gas will accumulate at low points when heavier
as air.

Gas dispersion models can produce 2 or even 3-dimensional results. They can take into account
structures of buildings, wind and turbulences

With respect to gas dispersion modeling we can mention the following benefits of using this
software:

 Dynamic and transient modelling possible

 The influence of impulse and turbulence can be simulated.
 Gas dispersion modeling can show the potential location of a gas cloud to the emergency
response personal.
 It can support the process of determining the location of gas detectors.

Yet, Gas dispersion modeling can have the following limitations:

 3-D modelling is very time consuming, in particular for more complex scenarios. There area
vast number of possible scenarios.
 Gas dispersion modelling software delivers inexact results when applied for short distance
modelling. The results give only a very rough indication.
 At the moment of an emergency situation there is no time to do useful simulations.
 Well trained and skilled staff is necessary to produce reliable and correct results.
 Maintenance of the modelling system requires significant effort.
9
 5. Managerial Aspects
It is difficult to assess the extent to which companies have integrated the different relevant aspects
of fixed gas detection systems within their organizations. For example does a holistic view and vision
exist concerning FGDSs? Starting at top management level and being part of all relevant levels of the
organization? Or is the approach e.g. more a site-specific ad-hoc approach which has to rely on the
qualification of
some technically
qualified “Data available reveals a significant difference in practice
individuals from between the different companies. A lack of management
the analyzer
strategy for FGDS on corporate and site level indicates a lack of
department?
awareness and ownership from management”

Guido Prinz – Total Petrochemicals

Figure 4: Elements of a Management System

10
The process of fire & explosion hazard management as described by the Energy Institute can be a
useful mean to manage fire & explosion hazards in an efficient and holistic manner including all
aspects of fixed gas detection systems. Additionally, it can be based on the management system
approach of ISO. Apart from this, it might be possible to extrapolate FEHM to toxic substances. The
following flow chart describes the main elements of FEHM:

Ignition sources Evaluate alternative Equipment maintenance

Hazardous materials Prevention, protection and Formalisation Preplanning
mitigation measures Exercises
Incident descriptions Legislation
Fire training

Fire Review risk Define Implement

scenario Reduction FEHM FEHM

analysis
options policy policy

periodic update
CONSEQUENCES

Life safety

Business interruption

POSSIBLE INPUT TOOLS POSSIBLE INPUT TOOLS POSSIBLE INPUT TOOLS

HAZOP Fire engineering Publications, Codes of

practice, design standards,
QRA, SWIFT Fire & gas dispersion specifications, guidance etc.
modelling
Incident experience...
Cost benefit analysis

Fig. 5 Main Elements of a FHEM

The corresponding code of safe practice of the Energy Institute contains more background
information and details on FEHM.

11
5.1 Training

Training of operators should be planned so as to enable thorough understanding of the devices and
the linked controls, checking of batteries and functions, implementation of visual and audible
alarming systems, as well as staff to contact in case of fault. It is important that operators receive
instructions to follow in case of triggering of an alarm. Periodic training sessions are recommended.

5.2 Emergency response strategies on alarms

According to a participant, related to gas releases
accident analyses of big
accidents with gas releases as It is agreed that adequate and well trained emergency
main cause have revealed that response strategies are crucial for a quick and successful
50% of the flammable gas response on gas releases.
releases ignited within 3
Many of the “typical” emergency response strategies have
minutes time. 90% of these
been developed a couple of decades ago. The strategies have
flammable gas releases
not kept pace with operational circumstances and
ignited within 10 minutes1.
contemporary sensibilities regarding life safety. In many cases
they are obsolete.

A common “typical” emergency response situation is that an operator receives a gas alarm in the
control room. In best case he immediately gives evacuation alarm manually and the fire brigade is
informed. In a worse case the operator sends a colleague to investigate the situation in the field. A
habit which seems still to be widely spread in the petrochemical sector. Both cases cost precious
time.

In the second case the person entering the zone of the potential gas release does not have sufficient
information to assess the situation before entering the potential danger zone. His life is at risk and
additionally precious minutes of time are spoiled in case it is an ignitable gas release. In the
meantime there is still no evacuation alarm started. Other persons being present on site stay in the
plant and stay at risk in the meantime. The following flow chart depicts the “typical” emergency
response approach:

12
 In the returned questionnaires the
Fig. 6 Classical emergency response strategy in petrochemicals following emergency response
companies strategies are mentioned for toxic
gases:
Consequently, time available for emergency response is the
limiting factor when developing emergency response strategies.
The development of site specific fast and intelligent emergency
 Shelters (safety islands)
response strategies is of utmost importance.
are installed where
The design and underlying philosophy of a fixed gas detection possible and necessary.
system for flammable gases can have a strong influence on the  The mitigation and
time available for emergency response. Consequently, when dispersion of gas releases
designing a new FGDS the whole sequence of events has to be by means such as water
taken into account including emergency response and time curtains.
available before ignition.  Preventing the spread of
In case of emergency situations with toxic substances the main contamination from
focus of the emergency strategies lay on the prevention of persons and equipment.
exposure of human beings and the surrounding areas. In general  Go to safety island if
toxic substances do not bear the risk of ignition. Consequently, in “toxic” alarm sounds and
the case time is not as crucial for the emergency response of toxic contact gatehouse. Gas
gases as it is for flammable gases. masks are available to
allow evacuation to
Poorly-designed alarm systems may hinder rather than help the alternative location if
operator and may result in failure to identify a need to act, or necessary.
failure to select an effective course of action, especially in
emergency conditions. Generally a change in the company policy
on alarm should be adopted to improve emergency responses to

13
alarms and trips in this case.

Case Study: alarm set points and emergency response

One workshop participant shares a case study concerning the deployment of FGDS on a LPPE plant
which has helped clarify thinking on emergency response with respect to detection of release of
combustible gas. The plant has to cope with process pressures up to 2000 to 3000 bar. An early
detection of gas leaks is essential to prevent major accidents.

Historically the plant has made extensive use of gas detection systems with installation at high
probability release points such as pumps, compressor seals and vents etc. Alarm trip points have
normally been in range of 10-60% LEL which have been influenced largely by seeking to avoid
nuisance alarms.

The latter can be caused by fugitive emissions, minor leaks and malfunctioning detectors. Emphasis
has been placed on early detection of a leak often aided by the human operator sense of smell in the
case of ethylene. The emergency response strategy is comparable to the “classical approach”.

For LPPE production in that company a paradigm shift is occurring in deployment of FGDS in the
plant where a zero tolerance policy for gas leaks has been implemented alongside the explicit
requirement that operators should not be put in danger by well meaning but ill considered
intervention to approach potentially present vapour clouds.

With support of the gas dispersion model PHAST several scenarios have been simulated with varying
leak sizes (0,5, 1 and 3 mm). Results revealed that it may be useful to use much lower gas alarm trip
points to detect gas leaks and to safe precious time for emergency response also taking into account
time before ignition.

In practice, semi-conductor sensors with the multiple alarm set points have been used with the
following trip points:

1st trippoint: 500 ppmV (2% LEL)

2nd trippoint: 3000 ppmV (12 % LEL)
3rd trippoint: > 100 %-LEL (magnitude)

With respect to emergency response the company considers an approach which can be summarized
in three scenarios following:

14
 Scenario 1:

Only one or a few low level alarms (2% LEL) emerge.

Information available in the control room indicates a gas leak which is probably of low risk. Based on
this, the operator may send staff in the field to investigate the situation.
● ● ●
Remaining operations such as maintenance works etc. continue.
In the context of
FGDSs a “zero
tolerance of leak”
policy might also
catalyze the
development of Scenario 2:
more intelligent and
better designed One intermediate and a few low level alarms emerge.
FGDSs for example
aiming at the
detection of lower Information available in the control room indicates a more critical situation
concentrations of than in scenario 1. The presence of a potentially hazardous vapor cloud is
hydrocarbons than possible. The risk for personnel is heightened. The operator´s first action is to
practiced today. In evacuate the plant to ensure that all personnel reach safe locations and the
practice, this will emergency response team is informed immediately. Similarly, operators can
require an evaluation analyze the situation to decide whether field inspection is possible. Isolation
of costs versus of sources of leaks by block valves is carried out where possible.
benefits and support
by management
commitment to the Scenario 3:
policy.
Multiple intermediate & high alarms emerge.

Information available in the control room indicates a very critical situation.

● ● ● The presence of an enormous vapor cloud is highly likely. In this case, the
evacuation sequence can be started by automatic action to heighten
reliability and to save time. At the same time, operators and emergency
response team have to analyze the situation to make sure that appropriate steps can be taken
without taking people at risk.

15
Eventually, experience with the multiple and very low alarm set points has been positive. Alarms of
the FGDS formerly often considered as nuisance alarms have turned out to be real leaks and
emissions in many cases and now taken seriously due to the management´s decision for a “zero
leak”-policy.

The triple alarm setting approach helps operators to estimate whether a current alarm is linked to a
small or high risk emission before intervening the danger zone. In the classical approach this was
much more difficult and operators started interventions “blindly”.

The following flow charts describes which steps of the “classical” emergency response approach were
influenced by the strategy followed within this case study.

16
 6. Further Guidance when implementing a FGDS

By the end of the workshop the respondents decided to summarize the lessons learned in the course
of the day by formulating a number of generic principles for fixed gas detection systems.

1) Define clear objectives what you want to achieve with or without fixed gas detection system
based on scenario based risk assessment.
2) If you decide to have a gas detections system, analyze the whole sequence of events
including emergency response strategy at the start of a project, taking into account available
time before ignition.
3) Verify that the proposed solution meets the objectives defined at the beginning of the
project.
4) Zero tolerance of leaks. A “zero tolerance of leaks” policy within a company will help to
reduce the probability of potential flammable and toxic gas releases. Similarly, it will lower
VOC emissions being relevant for the costly issue of emission right trading.
5) Make sure operator gets sufficient information for effective response without putting people
at risk.
6) Analyze first, than intervene.
7) Bring people to safe locations in a critical situation.
8) Design and maintain the fixed gas detection system in a manner that the system is reliable
and trusted by the operators (e.g. elimination of nuisance alarms and trips).
9) Use dedicated alarms if toxic and flammable chemicals on one site.
10) Use multi-disciplinary teams to define locations and technologies of detectors.

11) Establish ownership for fixed gas detection systems on the different levels of your
organization.
12) Anchor deployment, maintenance and MOC for fixed gas detection systems in your
management systems.
13) Avoid the creation of a false feeling of safety by installing fixed gas detection systems for
which effectiveness of detection and response are not sufficiently thought through.

17