adipec 2013 technical conference manuscript - marszal
DESCRIPTION
The placement of gas detectors has traditionally been an imprecise field of engineering. With no detailed prescriptive rules on when and where to place gas detection equipment, designs have been left to experts who use their judgment along with rules of thumb to set designs. These ad hoc methods have left industry in a position where different process units within the same refinery have vastly different gas detection designs for equipment in similar operating profiles. Furthermore, often no documented basis for the selection exists making it difficult to justify the differences in designs between units to stakeholders and regulators. In 2011, ISA released a technical report describing performance based methods for fire and gas system (FGS) design. This technical report laid out a safety lifecycle and introduced the new metric of “coverage” to define FGS designs. The approach presented in ISA TR84.00.07 was applied to the problem of H2S gas detection on a refinery Sulfur Recovery unit. All of the process equipment was assessed using calibrated semi-quantitative techniques, resulting in graded areas with associated coverage targets. Fire and gas mapping software was then utilized to confirm that the assigned coverage targets were achieved. This paper describes how that project was executed, presents an overview of the results, and compares the resulting design against other process units and expectations.TRANSCRIPT
1
ADIPEC 2013 Technical Conference Manuscript
Name: Edward Marszal
Company: Kenexis Consulting Corporation
Job title: President
Address: 3366 Riverside Dr, Columbus, Ohio, 43221, USA
Phone number: 614‐451‐7031
Email: [email protected]
Category: Case Studies
Abstract ID: 555
Title: Case Study: Implementing Performance Based Gas Detector Placement per ISA TR 84.00.07 on a Sulfur Recovery
Unit
Author(s): Edward Marszal, Elizabeth Smith
This manuscript was prepared for presentation at the ADIPEC 2013 Technical Conference, Abu Dhabi, UAE, 10‐13 November 2013.
This manuscript was selected for presentation by the ADIPEC 2013 Technical Committee Review and Voting Panel upon online submission of an abstract by the named author(s).
Abstract:
The placement of gas detectors has traditionally been an imprecise field of engineering. With no detailed prescriptive rules on when and where to place gas detection equipment, designs have been left to experts who use their judgment along with rules of thumb to set designs. These ad hoc methods have left industry in a position where different process units within the same refinery have vastly different gas detection designs for equipment in similar operating profiles. Furthermore, often no documented basis for the selection exists making it difficult to justify the differences in designs between units to stakeholders and regulators. In 2011, ISA released a technical report describing performance based methods for fire and gas system (FGS) design. This technical report laid out a safety lifecycle and introduced the new metric of “coverage” to define FGS designs. The approach presented in ISA TR84.00.07 was applied to the problem of H2S gas detection on a refinery Sulfur Recovery unit. All of the process equipment was assessed using calibrated semi‐quantitative techniques, resulting in graded areas with associated coverage targets. Fire and gas mapping software was then utilized to confirm that the assigned coverage targets were achieved. This paper describes how that project was executed, presents an overview of the results, and compares the resulting design against other process units and expectations.
2
Introduction A Fire and Gas System (FGS) performance‐based design basis was developed for the Sulfur Recovery Units (SRUs) operated by a US Gulf Coast Refinery. The study was performed to determine the hazard posed to personnel from process material releases in the vicinity of the SRUs. The Sulfur Recovery Units are responsible for treating gases that have elevated levels of Hydrogen Sulfide (H2S) present, e.g. acid gas and sour water stripper gas produced as a byproduct of other refining processes. As a result, the SRU has a significant inherent risk if a release were to occur. The study included analysis of the H2S hazards, dispersion modeling, an assessment of detector coverage, and recommendations for detector placement. The case study was performed based on guidelines from ANSI/ISA‐TR84.00.07‐2010 Technical Report Guidance on the Evaluation of Fire, Combustible Gas and Toxic Gas System Effectiveness. The Performance‐Based FGS Lifecycle Process, as defined by this technical report, is shown in Figure 1. Figure 1 FGS Safety Lifecycle (from ANSI/ISA‐TR84.00.07‐2010)
Step 1 can be determined by either the facility if the project is limited in scope, e. g. SRU included in the Case Study, or can be defined during execution of a project if an entire facility is being analyzed, e.g. offshore platforms. Steps 2 through 6 are discussed in the section of this paper that addresses the Hydrogen Sulfide Hazard Analysis, whereas Steps 7 through 11 are discussed in the Coverage Assessment section, although each step is not explicitly identified. Note that Step 7, in this particular case study was provided as an existing H2S Detector Array that will remain in place, and Step 9 was not within the scope of this case study.
Hydrogen Sulfide Hazard Analysis Hydrogen sulfide‐containing process streams are treated in the Sulfur Recovery Unit to convert gaseous hydrogen sulfide to elemental sulfur. The Claus reaction was used in this particular application.
H2S + O2 S2 + 2 H2O
As the H2S is being converted and recycled, there are various concentrations of H2S present in the SRU, ranging from less than 1% (v/v) to greater than 75% (v/v), or 1,000 ppm to 75,000 ppm. These concentrations are hazardous to personnel. Hydrogen sulfide is a broad spectrum toxin, meaning that it damages several different body systems simultaneously. The most pronounced effects are as a pulmanotoxin resulting in pulmonary edema in concentrations in the low 300 ppm range and a neurotoxin resulting in rapid and then sudden loss of breathing at concentrations as low as the 500 ppm range.
3
Design Basis Scenarios and Dispersion Modeling
The design was based on analysis of location and magnitude of hazard scenarios. In this study, gas dispersion modeling was conducted for the identified credible hazard scenarios to determine the potential size of gas clouds and to determine which equipment had the potential to release hazardous concentrations of H2S. Equipment of interest was identified based on this analysis, and existing detectors were modeled to determine coverage of identified hazard scenarios. Since every possible hazard scenario cannot feasibly be analyzed, a subset of those scenarios that have been chosen as the basis of the design, those scenarios are herein designated as “Design‐Basis Scenarios”. Dispersion modeling was conducted to provide incident outcomes of design basis hazards. In the case of H2S gas, the design basis hazard has been defined as an initial incipient stage gas cloud which can be detected at a point which allows for personnel egress from the affected area as well as preventing personnel entry into the area if unoccupied at the time of release. Dispersion modeling was performed using commercially available consequence analysis software and general parameters reflecting the condition at the facility. In selecting appropriate design‐basis scenarios, consideration was given to the predominant H2S‐containing streams present in the SRU. In sulfur recovery units, release scenarios associated with 5 mm equivalent hole diameter were selected as credible design‐basis scenarios. The released streams that were considered included: Acid Gas, Sour Water Stripper Gas, Tail Gas, and Rich Amine used for both Tail Gas Treatment and Acid Gas Treatment. Tail Gas was modeled as a 25mm hole size due to the less concentrated process conditions and H2S concentrations. Credible hazard scenarios were identified that have the potential to result in personnel injury due to H2S exposure. A small leak (5 mm equivalent hole diameter for excepting Tail Gas) was identified as having the capability to create a hazardous situation while simultaneously being more difficult to detect than larger leaks. For H2S exposure, the level‐of‐concern for injuries is 100 ppm (parts per million v/v) airborne concentration (IDLH as per the United States National Institute of Occupational Safety Hygienists [NIOSH]) and life‐threatening effects at 700 ppm airborne concentration or higher (immediate exposure). In addition, lower concentrations associated with occupational exposure of 10 ppm were modeled, this concentration is the limit of detection / alarm capability for many common H2S detectors. The Short‐Term Exposure Limit (STEL) as established by the American Conference of Government Industrial Hygienists (ACGIH) is 5 ppm as of 2010 publication (the STEL was previously 15 ppm). The STEL is the exposure concentration not to be exceeded for up to 15 minutes, not more than 4 times per 8 hr day. The justification for lowering the STEL was to protect against minor irritation of the respiratory tract and a “brief change” in rate of oxygen uptake. For each representative release scenario, H2S dispersion was modeled to four endpoints, as shown in Table 1. Table 1 – H2S Dispersion Endpoints
Endpoint Concentration
(ppm)
Notes
700 Potentially Fatal – Consistent w/ Course H2S Dispersion Study
100 IDLH – Consistent w/ Course H2S Dispersion Study
10 H2S Detection High Alarm Concentration
4
Commercially available computerized consequences analysis software was used to characterize the extent of toxic dispersion and the potential impact on personnel. Table 2 shows the release rate of H2S, the duration of the release modeled, and the anticipated distance the H2S will travel downwind of the release point. The reported distances are the IDLH concentrations and detectable concentrations, respectively. Table 2 Design Basis Consequence Modeling
Model Description Release Rate (lb/hr)Duration
(min)
Distance to Endpoint (ft)
Distance to
100 ppm
Distance to 10
ppm
S‐01 Acid Gas ‐ 5mm vapor leak 13 60 100 540
S‐02 SWS Gas ‐ 5mm vapor leak 9 60 60 240
S‐03 Tail Gas ‐ 25 mm vapor leak 170 60 50 80
S‐04 Tail Gas Clean‐Up Rich Amine 27 (1) 60 35 160
S‐05 Acid Gas Clean‐Up Rich Amine 144 (1) 60 90 350
(1) Release Rate of H2S only
Figure 2 shows the results of the Acid Gas design basis scenario dispersion modeling when taking a vertical cross‐section of the dispersion profile. This view represents the height of the cloud as well as the downwind distance reached by various concentrations of H2S. Figure 2 Acid Gas ‐ 5mm leak ‐ Vapor Dispersion ‐ Sideview
Figure 3 shows the results of the Acid Gas design basis scenario dispersion modeling when taking a horizontal cross‐section, or “footprint” of the dispersion profile. This view represents the width of the cloud as well as the downwind distance reached by various concentrations of H2S at or near ground level. This model type is used during the detector coverage assessment.
5
Figure 3 Acid Gas ‐ 5mm leak ‐ Vapor Dispersion ‐ Footprint
The gas dispersion analysis is also sensitive to meteorological conditions. A wind rose was obtained to determine the relative probability of the release being blown in any specific direction. Also, the following atmospheric stability and wind speed condition was modeled after considering the range of atmospheric conditions that are likely to occur at the facility.
Adverse Nighttime: F class atmospheric stability with 1.5 m/s wind speed Toxic gas cloud sizes resulting from releases are very sensitive to the selected wind speed. In addition, the wind direction distribution (wind rose) for the case study area was used to determine the probability of gas dispersing in any given direction around the point of release. This is shown in Figure 4.
6
Figure 4 Case Study Wind Rose
Design Basis Consequence Assessment In order to assess the coverage provided by toxic gas detectors, the location, size and frequency of the design basis releases must be determined. This is performed on an equipment‐by‐equipment basis, by first determining which of the representative release scenarios most accurately reflects the process material within each piece of equipment. The equipment must then be assigned a location in relation to other equipment and also gas detectors in the vicinity, which is typically done using plot plans. The frequency of release is based on the type of equipment being analyzed. Generally rotating equipment, such as pumps and compressors, is assigned a release frequency that is greater than that assigned to pressure vessels or shell and tube heat exchangers. These frequencies can be obtained from publicly available databases of leak rates as well as proprietary data from specific process plants and operating companies. Table 3 shows some of the major pieces of SRU equipment considered for this study, the representative scenario applicable to that equipment, and the chosen release frequency (assuming a 5mm leak. Please note, Table 3 does not include all leak sources included for this study. Table 3 Sample of SRU Equipment and Associated Release Scenarios / Frequencies
Equipment Description Representative
Scenario
Release Frequency
(/yr)
Amine Stripper Tower S01 1E‐03
Stripper Reflux Pumps S01 1E‐02
Acid Gas Scrubber S01 3E‐03
SRU Scrubber Pumps S01 1E‐02
Reactor S01 3E‐03
SWS Gas Scrubber S02 3E‐03
7
Equipment Description Representative
Scenario
Release Frequency
(/yr)
SWS Gas Scrubber Pumps S02 1E‐02
SWS Gas Knock‐Out S02 1E‐03
Sour Water Pumps S02 1E‐02
Tail Gas Treatment Reactor Feed Cooler S03 1E‐03
Tail Gas Treatment Reactor S03 3E‐03
Tail Gas Treatment Stripper S03 3E‐03
Tail Gas Rich Solvent Pump S04 1E‐02
Tail Gas Rich Solvent Flash Drum S05 1E‐03
Acid Gas Rich Solvent Pumps S05 1E‐02
Acid Gas Lean / Rich Solvent Exchanger S05 1E‐03
The magnitude and frequency of releases has been determined based on the design scenarios and the types of equipment that process the hazardous material. This is as inputs to the hazard assessment and the gas coverage assessment discussed in the following section.
Gas Coverage Assessment The Gas Coverage Assessment consisted of the following tasks for the case study discussed in this paper:
Define Performance Targets
Mapping of Gas Releases
Assess Detector Coverage
o Unmitigated Risk Assessment (without benefit of detectors)
o Mitigated Risk (first‐pass or existing detector design / layout)
Modify FGS Design
A Gas Detection System performance assessment typically also includes an assessment of the FGS Safety Availability. The Case Study on which this paper is based did not include this in the scope of the H2S hazard assessment, only an assessment of coverage was carried out.
Defining Performance Targets The performance grade is a specification that defines the ability of an FGS function to detect, alarm, and if necessary, mitigate the consequence of a fire or gas release upon a demand condition. In concept, a higher hazard installation should require higher levels of performance; while a lower hazard installation should allow for lower levels of performance, so that FGS resources can be more effectively allocated. Specification of target performance grade for fire and gas systems is an exercise in risk assessment. For the Case Study, a semi‐quantitative risk assessment was performed to specify risk reduction requirements for each area using a scoring procedure. The goal of this risk assessment was to assign a performance grade for each leak source within the Sulfur Recovery Unit zone. Each process area containing toxic gas hazards was characterized by one of four performance grades. These grades are listed in Table 4.
8
Table 4 Gas Performance Grades
Grade Exposure Definition Required
Detector Coverage
A Hydrocarbon processing, with high exposure. 90%
B Hydrocarbon processing, with moderate exposure. 80%
C Hydrocarbon processing, with low or very‐low exposure. 60%
No Grading Risk is tolerable w/o benefit of FGS ~
Each of the grades serves to define a relative level of gas risk, with grade A being the highest risk areas and grade C being the lowest risk areas necessitating detection. The grades correspond to required levels of FGS performance (i.e. detector coverage) as shown in Table 4. The Required Detector Coverage metric was chosen for this project. Alternatively, the release frequency, detector coverage, FGS availability, and other factors (e.g. occupancy) can be assessed and compared verified against allowable risk targets (e.g. Target Mitigated Event Likelihood) as defined by the organization or facility. Gas performance grades were specified for gas functions which protect H2S processing areas. Performance targets were selected toxic gas detection for each major equipment item. A summary of the results are presented in Table 5, which provides the selected performance target for gas functions in the SRU for the equipment containing the defined design basis scenarios. The grade selections are based on equipment type, occupancy, process conditions, and toxic concentration. Table 5 SRU Grade Selection Summary Plant: 1. Sulfur Recovery Unit
Zone ID Zone Description Zone Type Area Grading Area Grading
Equip Tag Equipment Description Hazard Type Grade
SRU Sulfur Recovery Unit X - Toxic Acid Gas Acid Gas 79% H2S <50 psig
H2S Gas A
Sour Water Stripper Gas
Sour Water Stripper Gas 40% H2S <50 psig
H2S Gas A
Tail Gas Tail Gas 0.8% H2S <50 psig
H2S Gas B
TGU Rich Amine Tail Gas Unit Rich Amine 1% H2S 50 - 100 psig
H2S Gas C
ARU Rich Amine Amine Regeneration Unit Rich Amine 2% H2S 100 - 150 psig
H2S Gas B
Note: the grades reported here are only indicative of design basis scenarios within the zone. FGS detection for all major equipment within each zone is defined independently based on the process stream conditions and equipment type. As the detector coverage is relevant to an entire zone and ell equipment contained within the zone, this Case
9
Study chose the highest grade, A, as the criteria that must be met, i.e. the detectable scenarios must be greater than or equal to 90% of the total scenarios.
Mapping Gas Releases The gas releases, as shown in Figures 2 and 3 and Table 3 are represented graphically using a commercially available FGS assessment tool. The mapping analysis involves obtaining an equipment layout or plot plan of the zone being analyzed. Figure 5 presents the Unmitigated Gas Risk. Figure 5 Unmitigated Gas Risk
This is a composite of all of the equipment that were considered leak sources, a sample of this equipment is listed in Table 2. The major equipment is shown here (with all identifying information removed) – the plot plan and overlay can be as detailed as the plot plans provided, resulting in varying degrees of detail in the geographic risk image. The FGS mapping software aggregates the release scenarios, consisting of concentration, frequency, and dispersion size, and plots the “footprint” overlaid on a plot plan of the unit. The location of the release is defined by the user to coincide with the equipment that is the source of the leak. The software output also shows the wind distribution and release frequency. As shown in Figure 5, the highest hazard frequency typically occurs in areas that contain the most equipment of concern. Alternatively, it can occur in the vicinity of higher risk (i.e., higher leak rate) equipment, such as pumps or compressors. Although not utilized in this case study, the release scenarios can also be represented in either the upwind or downwind direction from equipment by an appropriate distance – this distance is a property of the release orientation, release height, process conditions, and the material being released.
10
Detector Coverage Assessment FGS detector coverage was assessed in order to determine the risk posed by an H2S release in the SRU to personnel located in the SRU and ensure that the performance coverage targets are achieved. The existing layout of detectors was assessed to ensure the coverage footprint is sufficient to provide the required hazard alarms and control actions. The effective range of selected detectors considers the expected environment based on test data. The existing design was analyzed to determine the gas scenario coverage. This involves an assessment to determine how effective the proposed array of detectors will be in detecting an incipient hazard at a level that will alert personnel to a toxic gas release. Using the FGS assessment software tool, Kenexis has reviewed the existing design and modified as necessary to deliver acceptable gas coverage. As the detection of H2S within the process area of this Case Study does not initiate any automatic actions, and is instead used only for alarm, the coverage numbers were based on any single detector being in alarm state, or 1ooN. It is also possible to report one or more detector’s ability to detect a hazard, i.e. 2ooN coverage, in instances where appropriate. This is often used for fire detection as well as automated actions (e.g. shutdown, isolation, blowdown, and fire mitigation measures) based on gas detection as 2ooN vote‐to‐trip will reduce spurious activation. The initial design, which accounted for the location of existing detectors, included 11 gas detectors. This resulted in an achieved coverage of 58.1%, which is graphically represented in Figure 6 and reported in Table 6. Table 6 Detector Coverage Summary
Zone Hazard Layout Number of
Detectors Achieved Coverage
(%, 1ooN)
SRU Toxic – H2S Existing 8 58.1%
Recommended 15 90.1%
11
Figure 6 Mitigated Gas Risk – Existing Detector Layout
Figure 7 represents detector placement after inclusion of the recommended detectors. These detectors are in addition to those detectors already existing. The Case Study focused on the inclusion of existing detectors with the addition of detectors to meet the performance targets.
12
Figure 7 Mitigated Gas Risk – Recommended Detector Layout
It is likely that the performance targets could have been achieved using fewer detectors had the Case Study included
provisions for an “optimized design”, i.e. achieving desired coverage with the fewest installed detectors. This is not
meant to imply that the additional detectors do not offer hazard reduction to the facility, as they may be used for earlier
detection, may be placed at locations that are heavily trafficked and therefore represent increased hazard to personnel,
or represent areas in which increased awareness is desired.
13
References ISA‐TR84.00.07 The Application of ANSI/ISA 84.00.01‐2004 Parts 1‐3 (IEC 61511 Parts 1‐3 Modified) for Safety
Instrumented Functions (SIFs) in Fire & Gas Systems
Kenexis Case Study on H2S Hazard Analysis and Assessment performed for a US Refinery’s Sulfur Recovery Unit.