Datasets:

hndc

/

AI4Hazard-small

like 0

5.4 Worker-Related Element Gr ouping |197 processes as well as to transient situations such as startups, shutdowns, and batch processes. Good OPs also describe the process, identify process hazards, and describe the measures required to safeguard against those hazards, not only process safety but also environmental, health, and occupational safety. Good OPs also describe safe operating limits, a troubleshooting guide, and emergency actions. Consequences of deviating beyond safe operating limits must be included to m aintain the sense of vulnerability . Operators m ust be trained on a procedure before they follow it in the field, and they m ust understand everything contained in the OPs. Following training, performance assurance should be done to m ake sure the operator follows and continues to understand the procedure. Ideally, training on procedures should be based on the procedure docum ent itself, rather than on separate training m aterials. When it is necessary to have separate training m aterials, and especially when the separate training m aterials are being used in lieu of the procedure, this is a warning sign that the procedure is not adequate. The necessity of following procedures was discussed in section 5.1 (Conduct of operations). This means also that procedures should be written so that they match what operators actually do. Moreover, procedures should be written in plain language, written to a com prehension level of no higher than 8th grade. Text should be well spaced with a line length no longer than the text on this page. Tables, figures, and illustrations should be provided as needed to enhance comm unication. All these m easures will help operators follow the procedures and resist their tem ptation to stray from the procedure, leading to normalization of deviance . Operating procedures should be controlled documents that are kept up to date whenever there is a change. Operators must use the current OP, and no old versions should exist except in the document management system. Most changes com e through the

Hazard Identification Learning Objectives The learning objectives of this chapter are as follows. Having completed the chapter, the reader should be able to: Understand hazard identification methods. Participate in hazard identification studies, and Create a potential process safety incident scenario. Incident: Esso Longford Gas Plant Explosion, Victoria, Australia, 1998 Incident Summary A major explosion and fire occurred at Esso’s Longford gas processing site in Victoria, Australia in 1998. Two employees were fatally injured, and eight others injured. The incident caused the destruction of Plant 1 and shutdown of Plants 2 and 3 at the site. This shutdown resulted in total loss of gas supply to Victoria and cons equential business interruption and economic impact. A process upset in a set of absorbers eventu ally caused temperature decreases and loss of flow of a “lean oil” stream. This allowed a metal heat exchanger to become very cold and brittle. When operators restarted flow of the lean oil to the heat exchanger, it ruptured, releasing a cloud of gas and oil. When the cloud reached an ignition source, the fire flashed back to the release point resulting in addition al equipment ruptures and an escalating fire. Figure 12.1. Photograph of th e failed end of GP905 reboiler (LRC)

38 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Figure 3.2. DOT-117 train car (DOT 2015) Lessons Compliance with Standards . The MMA SOP required a prescrib ed number of hand brakes to be set depending on the number of railcars and the grade of the parking location. The MMA- 002 train was not in compliance with this requ irement. Additionally, the brake effectiveness check was not performed correctly in that the check was conducted with the air brakes set. Standards, whether governmental or company, should be followed. When standards are not followed and work is completed based solely on one’s experience or judgment, then the benefit of other person’s experiences, hard learnings, and even expert calculations are a resource and opportunity wasted. Asset Integrity and Reliability . The locomotive that failed had engine problems in October 2012 and a repair was made. Two days before the Lac-Megantic incident the locomotive engineer reported problems with th e same engine surging. When the locomotive was parked at Nantes, the smoke and oil spra y was noticed by the taxi driver, but the locomotive engineer and the rail traffic contro ller felt it could wait until morning to be addressed. Nonetheless, this sa me engine was the only one left running and the sole source of air pressure for the parked train. After the incident, tests showed that the cam bearing had fractured when the mounting bolt was over-tightened after the non-standard repair in October. Repairs should be made following expert direction. Making do with materials on hand and over-tightening bolts are frequently noted in accident reports. Additionally, operational issues with equipment that has just been repaired should be reported and investigated to ensure that it is fit for continued service. Management of Change . This incident is an example of creeping change in an industry over multiple years. The industry was, for the most part, satisfied with the performance of the DOT-111 cars. However, the number of cars in a single train, the volume of crude oil being transported, and the properties of the crude oil we re changing significantly. The impact of this

292 | Appendix E Process Safety Culture Case Histories A facility included a KPI based on the num ber of overdue ITPM tasks in the AI/MI element, which m any facilities do. The facility defined the KPI as any ITPM task that was overdue in 2 m ain asset- tracking software packages. One software was used to m anage rotating equipment, instruments, and electrical equipment, while the other was used to m anage fixed equipment pressure vessels, tanks, piping, and relief devices. Upon im plementation, this KPI revealed a few items overdue month-to-month, but the value was low, as was the aging of the overdue ITPM tasks. Two years later, during a PSMS audit, auditors found that there were other ITPM tasks that were important to process safety that were overdue but were not tracked in either of the two tracking software packages and therefore were excluded from the KPI. And those results were m uch less favorable. The Fire Chief tracked fire system ITPM in his electronic calendar. The annual fire pum p flow tests had not been conducted for two years and the ITPM tasks required by NFPA-25 were not included in the calendar. The Instrument shop supervisor tracked the annual calibration of testing equipment in a spreadsheet and there were ten pieces of test equipment that were overdue for annual calibrations.

Piping and Instrumentation Diagram Development 202 11.3 Different Types of Heat Exchangers and  Their Selection Different types of heat exchangers are used in the process industry. The reason for the different available types is the differences in indirectly contacting two fluids to each other. For example in shell and tube heat exchangers cold and hot streams transfer the heat through the peripheral area of several tubes. The most common type of heat exchanger is the “shell and tube” (S&T HX) type. In this type of heat exchanger, a bundle of tubes is secured on one or both sides in tube sheets and is placed inside a cylindrical body, or shell. A tube sheet is a perforated sheet that secures the tubes in its holes. One fluid is flowing through the tubes and the other fluid is flowing in the space between the outside of the tubes and the shell. If a tube bundle comprises only one tube, this could be considered as a variation of a S&T HX and it is named a “double pipe heat exchanger. ” After shell and tube heat exchangers, possibly the most common heat exchangers are “plate and frame (P&F HX) types. P&F HXs provide channels for each stream. In this type of heat exchanger, several plates are put together in the form of a sandwich and are secured between two “jaws. ” Plates are generally in rectangular shape. The plates are separated from each other by peripheral gaskets, which provide a gap between every two adjacent plates. Hot streams and cold streams flow through these narrow gaps, or channels. The last type of heat exchanger is the spiral type. To visualize this type of heat exchanger, consider a P&F HX with a larger‐than‐usual plate size. If someone rotates and wraps this “sandwich” around a core, then a spiral heat exchanger is obtained. Spiral heat exchangers are the least common and the most expensive type of heat exchangers. There are plenty of criteria that should be considered for decision on a specific type of heat exchanger. Table  11.1 summarizes some rules of thumb for the selection of heat exchangers. Another rule of thumb uses the required heat transfer area for the selection of heat exchangers. This rule of thumb is presented in Figure 11.2. At the end of this section one important piece of te rminology needs to be discussed. Each heat exchanger has two enclosures in contact with each other and with a common wall. Process heat ex changer Utility heat ex changer Furnace Figure 11.1 Usage of hea t transfer units in process plants.Table 11.1 Rule of thumbs for selec ting heat exchanger types. Heat exchanger type When? Shell and tube (S&T) heat exchangerFixed headDefault choice but applicable if desired temperature is less than 50–60 °C (or le ss than 80–90 °C wit h expansion ring) Floating headWhere the desired temperature is more than 50–60 °C ●Where the shell side fluid is fouling Double pipe heat exchangerWhen the decision is S&T but the required heat duty is low Plate and frame (P&F) heat exchanger ●Wherever is not enough room for an S&T type but enough pressure is available ●Pressure and temperature cannot be severe otherwise the tolerance of the gaskets of the P&F HX will be exceeded ●Where the required heat duty is not certain and the modular structure of P&F helps in the future addition of plates (and increase in the heat transfer area) Spiral heat exchanger For very fouling services Aerial cooler When cooling down to approximately 65 °C is ade quate Helical coil For small heat duties, for example sample coolers and pump seal water heat exchangers Helical coilD ouble pipe 5 m220 m21,000 m25,000 m2Shell & Tube-floating headHeat transfer area Shell & Tube-fixed head Figure 11.2 Selec tion of heat exchangers based on the required heat transfer area.

Appendix B. Inherent Safety Analysis Approaches Inherent safety (IS) can be analyzed in a number of ways, but in all cases, the intent is to formalize the consid eration of inherent safety, rather than to include it by circumstance. By formally including inherent safety, in either a direct or indirect way, facilities can fully realize the potential benefits of inherent safety. In additi on, all IS considerations will be fully documented. Three analysis methods can be used to evaluate implementation of IS: 1.Inherent Safety Analysis: Guided Checklist Process Hazard Analysis (PHA) 2.Inherent Safety Analysis: Indepe ndent Process Hazard Analysis (PHA) 3.Inherent Safety Analysis: Integr al to Process Hazard Analysis (PHA) Method 1 employs a specialized ch ecklist containing a number of practical inherent safety consider ations organized around the four strategies of minimization, substitu tion, moderation, and simplification. The advantage of this approach is that it is very direct and asks pointed questions that have proven to be va luable in reducing hazards at past locations. The disadvantage is that, as with any checklist, it may be limiting in that other ideas may surf ace if the team was asked to more creatively determine applications for the inherent safety strategies given a safety objective. (Note that the checklist appearing here is only a representative subset to illustrate its use. See Appendix A for the complete checklist.) For the second method, the team is asked to avoid a particular hazard at a designated part of the pr ocess. In this case , the team reviews a problem, determines which of th e inherently safer strategies may apply, and then brainstorms possible ways the hazard can be reduced or eliminated. 455(VJEFMJOFT
GPS
*OIFSFOUMZ
4BGFS
$IFNJDBM
1SPDFTTFT
"
-JGF
$ZDMF
"QQSPBDI #Z
$$14 ¥

5IF
"NFSJDBO
*OTUJUVUF
PG
$IFNJDBM
&OHJOFFST

DETERM INING ROOT CAUSES 217 At this point, the logic tree structur e is examined to en sure that the tree is logically consistent and compatible with the known facts. In some instances, there may be inconsistencies, and application of the fact/hypothesis matrix will be appropriat e. Inconsistencies found at this point require further tree development or rearrangement. Once the logic tree structure appear s to be consistent, the first of three quality assurance tests is applied by examining the overall logic tree structure for completeness. The logic in each branch of the tree should be tested to determine if it is necessary and sufficient. (Details and tips for testing the logic are discussed in Section 10. 5.2.) If the tree appears to be complete, the next quality assurance test is initiated. If the tree is incomplete, then the fact or logic problem is identified and the entire process is repeated. This is called an iterative loop. If the logic tree appears to be complete, then the second quality control test is applied by asking the questi on, “Are the causes that have been identified actually relat ed to management systems?” If the answer is yes, then the investigation proceeds to the third quality control test—the final overall review. If management system causes have not been found, then the iterative loop process is used. It is important to note that not all management system causes may be located at the extreme bottom points on the logic tree. Some of the management systems-related causes can be - and often are - located in the upper or middle portions of the logic tree diagram. Some causes can also be identified by the logic tree structure it self. For example, an overview of the entire tree structure may indicate significant gaps or overlaps in responsibilities, or it may disclose conflicting activities or procedures. These insights may be overlooked if the inve stigators limit their cause search to only the bottom level of the structure and fail to review the entire tree and the interrelationships between branches. If the test for systems-re lated causes is satisfac tory, then the third and final quality assurance test is applied. This is an overall review of the logic tree as a whole for both facts and logi c. A conscientious review of each branch should be made to look for possible conflicts or inconsistencies. It is a pause to focus on the logic tree fr om an overall perspective, not just each branch. The final logic diagram should be thoroughly checked against the final timeline to ensure that these two are in complete agreement. The team should also verify that none of the facts is in conflict with the tree. If the incident investigation team is satisfied with the causes i dentified, then the

142 | 4 Applying the Core Pr inciples of Process Safety Culture steadily increasing share price. Therefore, boards are heavily influenced to focus their attention on matters that relate directly to the share price. As discussed relative to the financial com munity, this can unintentionally motivate m anagement to normalize deviance . Some boards are more independent or more supervisory than others. Com panies with weaker process safety cultures often have boards that acquiesce in management decisions and take little interest in process safety. Com panies with stronger process safety cultures have boards that recognizes that a strong PSM S and culture will help reduce risk and protect the company im age, two things that help increase share price. B oards should oversee auditing program s to ensure that they have a true measure of the health of the company’s process safety effort. Towards this, some com panies have found it helpful to have one or more board m em bers who understand process safety concepts. This approach is expected to gain in practice in the com ing years. 4.5 Process Safety Culture Metrics Like other aspects of a PSMS, process safety culture should be m easured periodically to monitor progress, guide improvem ent, and detect regression. Culture changes, especially positive changes are usually slow and hard to discern. Nonetheless, historic experience dem onstrates that regular monitoring can reveal cultural changes over time. Practical experience as well as formal study by the International Atom ic Energy Agency (IAEA) (Ref 4.16) has shown that a single quantitative measurem ent of culture m ay be impossible. Instead, culture can be sensed from qualitative indicators. Therefore, facilities and com panies should select a range of indicators that reflect the individual culture core principles. These indicators m ay be based on observable behavior, conscious attitudes, perceptions, or beliefs

16 INVESTIGATING PROCESS SAFETY INCIDENTS This example illustrates that event trees can be useful models of an incident sequence because they provide a graphical, logic-based depiction of the various potential consequences that could occur, depending on the p a t h w a y of a n e v e n t . T h i s i s a mor e structured sequen ce model than the three-phase model, but it does not fu lly address the weaknesses in barriers and the management systems behind them. 2.1.3 Swiss Cheese Model Another way to represent the staged even ts and conditions that result in an incident is by using the Swiss Cheese model (Reason, 1990). This model takes one of the failure paths d efined in the event tree that leads to a consequence of concern. The protective barriers (safety systems) are represented by parallel slices of Swiss cheese. Thes e barriers represent the equipment, procedures/practices, and people that comprise elements of the management system for the facility. Ideally each barrier shou ld be robust, but like th e holes in Swiss cheese, all barriers have weaknesses (Figure 2.2) resulting from: Active failures (e.g., equipment fa ilures, unsafe acts, human errors, procedural violations, etc.). Latent failures (e.g., design/equipment deficiencies, inadequate/ impractical procedures, time pressure , unsafe conditions, fatigue, etc.) – see Section 2.1.4 below. These weaknesses can lead to manage ment system failures resulting in a process safety incident (see Section 2.2.2 below). Figure 2.2 Swiss Cheese M odel

22 Human Factors in emergencies 22.1 Learning objectives of this Chapter This Chapter provides an overview of the Human Factors of emergency response. By the end of this chapter, the reader should be able to: • Understand how Human Factors affe ct performance and management of emergency situations. • Recognize the importance on non-technical skills in emergency response. 22.2 An example accident 22.2.1 Milford Haven refinery explosion, Wales, 1994 On July 24th, 1994, a large explosion occurred at Texaco Refinery, Milford Haven in Wales, which caused injury to 26 peop le [87]. The blast from the explosion damaged properties in a 10 mile (16 kilo meter) radius and was heard 40 miles (64 kilometers) away. The site suffered severe damage to the process plant, the building, and storage tanks. A summary of the event is given in B.4 (page 389). During the sever electrical storm that proceeded the explosion, operators and operations management failed to identify the underlying causes of the problem or to recognize that they had the potentia l to lead to hazardous consequences, despite these data being available to them. They continued to operate in a disturbed environment for five hours prior to the explosion. All the information, including alarms, was available to the operators via six distributed control systems (DCS) screens, which were used to contro l the process and to diagnose faults. Many alarms, in the plant were sounding simultaneously, all with the same - high priority. In the 15 minutes before the explosion, operators were receiving alarms at a rate of one every two seconds. Thirty minutes before the accident, a critical alarm went off. Had the operator s recognized the criticality of the final alarm and taken appropriate action, the explosion may not have happened. The accident was caused by a combination of factors, including: • A control valve shut when the control system indicated it was open. • A modification that was carried ou t without proper assessment of consequences. • Control panel graphics that did not provide the necessary process overview. • Attempts to keep the unit running when it was supposed to be shut down. • Inadequate emergency management. Human Factors Handbook For Process Plant Operations: Improving Process Safety and System Performance CCPS. © 2022 CCPS. Published 2022 The American Institute of Chemical Engineers.

152 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Table 9.1 continued Severity Points Consequence Categories Safety/Human Health c Direct Cost from Fire or Explosion Material Release Within Any 1- Hr Period a, d, e Community Impact Off-Site Environmental Impact b, c 9 points — A fatality of an employee, contractor, or subcontractor, or — A hospital admission of a third party. Resulting in $10,000,000 ≤ Direct Cost Damage <$100,000,000 Release volume 9x ≤ Tier 1 TQ < 27xoutside of secondary containment. Officially declared evacuation > 24 hours < 48 hours. — Resulting in $10,000,000 ≤ Acute Environmental Cost <$100,000,000, or — Medium- 27 points — Multiple fatalities of employees, contractors, or subcontractors, or — Multiple hospital admission of third parties, or Resulting in ≥$100,000,000 of direct cost damages. Release volume ≥ 27x Tier 1 TQ outside of secondary containment. Officially declared evacuation > 48 hours. — Resulting in ≥ $100,000,000 of Acute Environmental Costs, or — Large-scale injury or death of aquatic or land-based a Where there is no secondary containment, the quanti ty of material released from primary containment is used. Where secondary containment is designed to only contain liquid, the quantity of the gas or vapor being released and any gas or vapor evolving from a liquid must be calculated to determine the amount released outside of secondary containment. b Judging small, medium or large-scale injury or deat h of aquatic or land-based wildlife should be based on local regulations or Company guidelines. c The severity weighting calculation includes a category for “Off-Site Environmental Impact” and injury beyond first aid (i.e. OSHA “recordable injury”) level of Safety/Human Health impact that are not included in the Tier 1 PSE threshold criteria. However, the purpose of including both of these values is to achieve greater differentiation of severity points for events that result in any form of injury or environmental impact. d For the purpose of Severity Weighting, general paving or concrete under process equipment, even when sloped to a collection system, is not credited as secondary containment. e Material release is not tabulated for fires or explosions. These events severity will be determined by the other consequence categories in this table.

Piping and Instrumentation Diagram Development 234 If the relieving gas/vapor is not innocent, other, more expensive, options should be considered. The options depend on the nature of the released gas/ vapor; if it is flammable it can be burnt in a flare; if it is absorbable in water the stream can be sent to a catch vessel. Figure 12.31 shows the different available options for gas/vapor relieving. The last choice for gas/vapor relieving is “system relieving. ” As was mentioned before, system relieving is not common for gases and vapors. However there are some cases that no other option is available and system relieving is the only technically doable option, for example in the oil extraction industry. Well pads are not necessarily close to the central plant, which has a flare system. There are, however, some units on well pads that have PRDs. Sometimes a large “pop tank” is located to release gases from the PRDs on  the  well pad. Pop tanks can be used for liquid relieving too.12.16.3 Two‐Phase Flow Handling Here two‐phase flow refers to gas–liquid two‐phase flow. We generally don’t provide a “two‐phase flow disposal system. ” What we try to do is to separate the two‐phase flow to its components, gas and liquid, and then deal with each of them separately. The reason for the separation of two‐phase flow is two‐ fold. On the one hand the design and fabrication of a two‐phase flow collection network is more complicated and expensive. On the other hand it is not easy to find a disposal system suitable for both liquid and gas at the same time. However, if the liquid fraction or gas fraction of two‐ phase flow is very small, the two‐phase flow can be con-sidered to be a single‐phase flow. The two phase separators could as simple as Tee‐ se parators, to the more complicated options of cyclone separators and knock‐out drums (blow‐down drums) (Figure 12.32). Sometimes even a combination of them is used. To Atm. To Safe Location Flare DisposalQuench/Catch Tank Figure 12.31 Differ ent gas/vapor disposal systems. Min. 3 m 15mGround or platformMin. 2 mFigure 12.30 Gener al meaning of “safe location” for releasing to atmosphere.

261 is not forgotten or overlooked as personnel and organizational changes occur (Ref 10.8 CSChE). Hendershot (Ref 10.15 Hendershot) argues that this is especially critical when dealin g with inherent safety and inherently safer design (ISD) features. He gives several examples where ISD features were essentially put at risk because the reasons they were implemented were not clearly and ad equately documented. Potentially, this could compromise facility safety when future modifications are made by people who do not understand the original designer’s intent, or what is involved in a particular sa fety feature. Examples include why a certain size feed line for limiting ma ximum reagent flow is part of the safety design basis, or how the routing of a pipe is intended to minimize consequences of a spill. ISD features are particularly susc eptible to lapses in corporate memory given that, unlike an add-on device such as a high-pressure alarm, they are such a fundamental pa rt of the design that their purpose may not be obvious (Ref 10.15 Hend ershot; Ref 10.1 Amyotte). All these inherently safer design features must be documented in original design manuals, and appropriate process safe ty information, including P&IDs and SOPs, must be readily available. This issue is also pertinent to M a n a g e m e n t o f C h a n g e i n t h a t a proposed modification should be reviewed carefully to ensure no ISD features are being compromised. 10.8.1 IS Review Documentation Whichever type of IS analysis is cond ucted, a report of the review should be generated to document the study. This report should include, at minimum, the following information: A summary of the approach used for the IS review (i.e., methodology, checklist used, etc.). Names and qualifications of the team facilitator/leader and team makeup, including positions, names, and any relevant experience or training. IS alternatives considered, as well as those already implemented or included in the design. If an independent inherently safer systems analysis was conducted, documentation should include the method used for the analysis, what inherently sa fer systems were considered, and the results of each consideration. If an IS checklist was used,

170 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS called a pitot tube. The difference between these two pressures, corrected for air density/height, is used by the flight computer to calculate the velocity of the aircraft relative to the air. The A330 features three sets of pitot tubes (“probes” – see Figure 7.1) (BEA 2012) and six static pressure sensors. The probes are fitted with drain holes to prevent water from accumulating and are heated to prevent icing. In December 2001, Air France had its first delivery of the A330 airc raft, which were all originally fitted with pitot probes manufactured by Thales, model number C16195AA. The A330 on flight AF447, registration F-GZCP, fi rst went into service in April 2005. Figure 7.1. The Three Pitot Tubes on the A330 Aircraft

374 Human Factors Handbook This model uses the terms active and latent failures. • Active failures are the unsafe acts committed by people who are in direct contact with the system. These include slips, lapses, or mistakes, such as omitting an operational task or performing a task incorrectly. • Latent failures are “resident pathogens” within a system caused by decisions made by engineers, proc edure authors, and management for example. These can create “error pr ovoking conditions” such as time pressures and understaffing and poor procedures. They may lay dormant for many years until a comb ination of events reveals them. Latent failures are also referred to as “psychological precursors” as they also create the conditions for error. This model has been used to help understand accidents and the role that the systems of management created the (hidde n) conditions for human error. This includes the notion that latent failures can cause multiple defenses to fail, and thereby undermine “defense in depth” sa fety management systems. The model is also used to prompt the identification and resolution of latent failures before they contribute to an accident. The concept being that resolving one latent failure would avoid many active failures. A.2 Compliance concepts A.2.1 “Violations” The term “violations” is not typically used in current human performance discussions. However, Professor James Re ason in his 1997 book “Managing the risks of organizational accidents” [119] st ated that “Violations are deviations from safe operating procedures, standards or rules. Such deviations can be either deliberate or erroneous…” (p72). Reason listed three major catego ries of safety violations: • Routine: these tend to be habitual “corner-cutting” in skilled performance. They may be associat ed with “clumsy” procedures and rare sanctions. • Optimizing: Professor Reason describe s these as “violating for the thrill of it”. • Necessary: These involve non-compliance being “essential” “to get the job done”. These tend to be related to organizational failings such as tools. Reason defined these as non-malevolent acts. The actions are intended but the harmful consequences are unintended.

120 Human Factors Handbook Figure 10-1: Competency Management

DETERM INING ROOT CAUSES 225 Consider the incident scenario discussed previously: A worker was walking on a co ncrete walkway in the process unit. There was some lube oil on the pad. He stepped into the oil, slipped, and fell. It was a sunny day; the worker was not carrying anything, was not dist racted, and was not doing any urgent task . The top portion of the logic tree may look something like the tree in Figure 10.11. Figure 10.11 Example Top of the Logic Tree, Employee Slip Each of the succeeding lower level events is further developed by repeatedly asking the question, “Why did this event occur?” Pursuing just one branch, for example the Oil Spilled on Pad branch would lead to at least two possible sources: Leak from Pipe and/or Hand Carried Containers , as shown in Figure 10.12 and Figure 10.13 . Figure 10.12 Example Logic Tree Branch Level, Oil Spill

92 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS operations team situational awareness is one way for operating companies to enable a proactive operating posture. (Bullemer & Reising, Effective Console Operator HMI Design (2nd Editio n), 2015). This is also the case for the outside (field) operators who interface/interact with the valves, gauges, and other pieces of equipment. Design engineers normally play an important role in recommending the design and layout of the console displays, alarm systems, field equipment and almost all other features of the HMI. Therefore, design engineers should be made aware of the management of abnormal situations, which should be included in their training sessions. This is especially important when al arm structure and display are being considered. Some published RAGAGEP standards such as those by ISA, EEMUA, and IEC (ANS I/ISA 2009; EEMUA 1999/Revised 2013; IEC 2014) should be applied when appropriate. In addition to applying these RAGA GEP standards, several techniques that can be employed in the design of the HMI include: Alarm configuration such as “First out alarm reporting” is key to understanding the initial alarm that directly indicates the presence of a developing abnormal situation. It is, by definition, what alarm cascade management is intended to solve. Alarm management strategy such as “alarm flood suppression” is the ability to support alarm cascade management by limiting initial alarms to those that are meaningful and actionable. For example, if a high-high level trip activates on a compressor inlet knockout drum, shutting down the compressor, the op erator only requires an alarm indicating the high-high level and the trip. The operator does not need subsequent alarms for high suction pressure, low discharge pressure, low lube oil pressure, and all of the other alarms that would accompany a compressor shutdown, which can make it harder for the operator to determine what happened and what to do next. This strategy is typically applied to complex process equipment systems such as refrigeration systems, burner management systems, and self-c ontained packaged units. The design and layout of plant and control equipment in the field is also highly relevant to the effective management of abnormal situations. The logical positioning of controls and valves relative to the items they affect will h e l p t o e n s u r e t h a t t h e c o r r e c t d e v i c e i s o p e r a t e d i n a n e m e r g e n c y . The clear and correct labelling of equipm ent, locally visible valve position

353 been home to chemical/processing industry facilities, including oil refineries and chemical plants. Following a series of serious indust rial accidents, the County enacted the Industrial Safety Ordinance (ISO ) in 1999 and has revised it several times since then, most recently in 2 014 (Ref 14.4 CCC ISO). Designed to be “the most stringent in the United States, if not the world,” (Ref 14.6 CCHS 2017) the ordinance expands on the California Accidental Release Prevention (CalARP) Program (Ref 14.1 CalARP) for petroleum refineries and chemical plants in the County. The City of Richmond, CA, located within Contra Costa County, adopted its own ordinance in 2001 (Ref 14.3 RISO). The Richmond Industrial Safety Ordinance (RISO) mirrors the county ordinance. Inherent Safety Requirement of the CCC ISO/RISO . A facility covered by the CCC ISO or RISO is required to conduct an inherently safer systems analysis (ISSA) for each covered proc ess as follows (Ref 14.4 CCC ISO): The facility conducts an ISSA on existing covered processes every five years. The facility conducts an ISSA in the development and analysis of recommended action items identified in a PHA. Whenever a major change is prop osed at a facility that could reasonably result in a major chem ical accident or release, the facility conducts an ISSA as part of the required management of change review. If an incident occurs and the inci dent investigation report or its associated root cause analysis recommends a major change that could reasonably result in a major chemical accident or release, the facility commences and completes an ISSA of the recommended major change as soon as administratively practicable after completion of the incident investigation or root cause analysis report. The facility conducts an ISSA duri ng the design of new processes, process units and facilities. Imme diately upon completion of the ISSA report advise CCHS of the av ailability of the ISSA report. The facility prepares a written report documenting each ISSA including identification and description of the inherently safer system(s) analyzed in the ISSA, description of the methodology used to analyze the inherently safer systems(s), the conclusions

254 INVESTIGATING PROCESS SAFETY INCIDENTS Figure 10.29 Analysis of the Human Engineering Branch When the first causal factor is analyzed using the remaining applicable branches (i.e., Work Direction, Pr ocedures, and Management System), the following root causes are identified: 1. Monitoring alertness needs improvement. 2. Shift scheduling needs improvement. 3. Selection of fatigued worker. 4. The “no sleeping on the job” policy needs to have a practical way to make it so that people can comply with it. The investigation team then repeats the process by considering the remaining causal factors one at a time: • Fire hose ruptures • Automatic shut -off jumpered • Contract operator cannot hear alarm due to noise Finally, the investigation team considers generic causes that pertain to the overall management system for th e process plant by considering the operating history and any other incidents that may have related causes.

1. Introduction 5 • Provides an explanation of how people think and behave, why people make mistakes, and how to help people perform process operational tasks successfully. This includes how to support human performance through procedures and job aids, training and learning, effective task planning, high reliability communicat ions, fatigue risk management, development of error management skills, and preparing people to perform emergency response tasks. • Briefly covers the Human Factors of change management and managing contractors. It also offers help on how to learn from errors, and how to use indicators of human performance to improve support to people. 1.2.2 Other guidance How does this handbook fit with other guidance documents? Safety culture, leadership, and process safety management are covered in other CCPS publications, as shown by the book front covers. Most chemical process businesses have a set of process safety management systems in place already. The advice in this handbook can be integrated into these process safety management systems. Human Factors methods, such as error analysis and Human Reliability Assessment, typically applied during a “Hazard Identification and Risk Analysis”, are not covered in this handbook. CCPS books on “Bow Ties in Risk Management” and “Guidelines for Integrating Process Safety into Engineering Projects” are available if further information is needed. This handbook does outline forms of error assessment that can be used by everyone involved in task planning and task management.

OVERVIEW OF IN CIDEN T CAUSATION 19 causes for the loss of containment or energy and implementation of appropriate remedial actions can pr event a more serious outcome in the future. Appropriate remedial actions are likel y to follow Inherently Safer Design (ISD) principles such as the measures lis ted in Figure 2.4 to prevent, control, and mitigate incidents (based on Haddon, 1980). Figure 2.4 Incident Prevention Strategies Haddon recognized that not all hazard s (chemical/energy sources) can be eliminated, and to protect vulnerable receptors (e.g., people), most remedial actions will likely reduce risk by applying additional safeguards or improving the management of existing sa feguards. This is consistent with CCPS guidance on Inherently Safer Design (ISD), which recommends a hierarchical and iterative approach covering first order (hazard elimination) and second order (reduction of severity or likelihood) ISD approaches (CCPS, 2007b). It is also important to consider why the magnitude of the consequence of an incident was, or under slightly different circumstances could have been, as severe. The potential consequence of an incident is often a function of the following five factors: 1. Inventory of hazardous material : type and amount 2. Energy factor : energy of a chemical re action or material state 3. Time factor: the rate of release, its du ration, and the warning time 4. Intensity-distance relation : the distance over which the hazard may cause injury or damage

Piping and Instrumentation Diagram Development 194 mentioned that the maximum discharge pressure of these pumps is limited by the pressure of the air source. The other option is used in “solenoid driven pumps. ” In “solenoid driven pumps” the drive is a stroking shaft driven by a repeatedly magnetized solenoid. Companies may or may not decide to show the pump drives. Figure  10.36 shows the P&ID representation of different drives. Even if it is decided to not show pump drives, it is very difficult to not show air‐operated drives as they are a type of process elements.10.7.6 Sealing S ystems for PD Pumps PD pumps similar to centrifugal pumps may need a sealing system. However, their system is not standardized and is designed by the vendor. If the rpm of the shaft is low enough, the designer may decide to use packing rather than a mechanical seal and then the need for a sealing system is eliminated. In a reciprocating pump the shaft is reciprocating rather than rotating and the sealing concept could be totally different. 10.7.7 Met ering Pumps (Dosing Pumps) Metering pumps or dosing pumps are pumps that are able to generate flow with adequate accuracy within a reasona-ble range of upstream and downstream pressures. Because obviously changing upstream and downstream pressures will change the flow rate of centrifugal pumps they cannot be categorized as metering pumps. Metering pumps are generally referred to as PD pumps. The majority of metering pumps are reciprocating type PD pumps because they have more flexibility in controlling them. Dosing pumps may be controlled through VSD and also change in stroke length (refer to Chapter 15). As the amount of injected chemical is generally important, the injecting flow rate should be checked occasionally by operators. In the majority of cases this is done by providing a drawdown (calibration) column on the suction side of the pump. The rounding operator closes the blocking valve upstream of the calibration column and allows the pump to get suction from the liquid inside of the column (instead of the upstream container) and measures the time for a specific drop in the liquid of the calibration column to calculate real flow rate (Figure 10.37). In some cases it is very important to install a back‐ pressure regulator. This is to mitigate one condition that may cause uncontrolled injection of chemical to the host stream. If the pressure of the host stream (which could be fluctuating) goes below the suction pressure of the Reciprocating pump sR otary pumps Electro motor Solenoid Air operated(in diaphragm pumps) (in diaphragm pumps)Electro moto rMM Figure 10.36 PD pump driv es. PG 123 Figure 10.37 P&ID repr esentation of a dosing pumps. MT MT MT MT Figure 10.35 Differ ent arrangements of dissimilar pumps in series.

Piping and Instrumentation Diagram Development 106 A valve is either throttling or blocking based on the structure and shape of the valve’s plug. These plugs are discussed later in this chapter. When placing a non‐special valve in a system, the first question is if a throttling or blocking valve should be used. Since each calls for different types of valves, a wrong decision may cause internal valve leakage (passing‐by) or premature failure of the valve. The design process engi ­ neer can decide whether the suitable duty of valve is throttling or blocking for a specific situation. Furthermore valves can also be classified according to the number of ports: two‐port or multi‐port valves. The most common valves are the two‐port valves, which have only two ports (i.e. one inlet and one outlet). However, multi‐port valves have more than two ports, possibly three or four. They also have more than one inlet or outlet ports. Also they can be called by various names depending on the function of the valve. Table 7.1 shows the four types of valves and their spe cific function. Based on the table a two‐port stopping valve is called an isolation valve. A two‐port adjusting valve is called an adjusting valve. A multi‐port stopping valve is called a diverting valve. A multi‐port adjusting valve is called an adjusting‐diverting valve. The two‐port valves are by‐default valves, and although the term “two‐port” is not stated, it is generally assumed to be two‐port valve. Diverting valves are valves that divert the flow from one destination to another. Each diverting valve can be replaced with two or more blocking valves in a specific arrangement. Adjusting‐diverting valves are valves that divert a portion of flow from one destination to another. Each adjusting‐diverting valve can be replaced with two or more throttling valves in a specific arrangement. Multi‐port valves have no advantage over the two‐port valves except saving space and money. They are not as robust as two‐port valves. Through the use of multi‐port  valves, a huge cost saving is gained specially if using remotely operated valves (ROT). The remotely operated valves will be discussed later, but in a nutshell, they are actually valve actuators. When it is supposed to use remotely operated valves, merging few of them together and using multi‐port valve generates a big saving as the saving in valve actuators are big. 7.4.1 Valv e Plug: Throttling vs. Blocking Valves Throttling valves can adjust the flow anywhere from 0% (fully closed valve and no flow) to 100% (fully open valve and full flow). However, the performance of the majority of these valves is best when they operate between 20 and 80% of flow. In contrast, blocking valves allow for full flow or no flow at all. Although using a blocking valve in a throttling application is possible, it is detrimental to the valve inter ­ nals in the long term. Throttling valves can also be used for isolation purposes, and although this does not damage the valve internals, they do not generally provide a tight shutoff (TSO). Blocking valves can be purchased as a TSO type. The concept is outlined in Table 7.2. Table 7.1 Valve action vs. the number of ports. Conventional (two‐port) Multi‐port Stopping flow Blocking valve Diverting valve Adjusting flow Adjusting valve Adjusting‐diverting valve Table 7.2 Fea tures of throttling vs. blocking valves. Throttling Blocking Application Throttling Stopping (blocking, on/off) Stem travel: appropriate positions0–100% wide (20–80%) Zero OR 100% wide Passing by? Generally no tight shutoff (Could be) tight shutoff Example A fully open valve may allow, for example, 45 m3 h−1 of flow; then a fully closed valve creates zero flowA fully open valve may allow, for example, 45 m3 h−1 of flow; a partially closed valve may allow, for example, 30 m3 h−1 of flow Interchangeability? Can also be used in blocking applications Cannot be used in throttling applications Valve operator Valve plug Figure 7.1 Struc ture of a typical valve.

CHEMICAL HAZARDS DATA SOURCES 125 NFPA 704 Standard System for the Identification of the Hazards of Materials for Emergency Response . This system is used by emergency resp onders to quickly identify the risks of hazardous materials involved in an emergency. This helps inform the response method, the materials used in the response, and the personal protective equipment that may be required. The NFPA 704 graphic uses a diamond as shown in Figure 7.3. The four corners of the diamond represent he alth (blue), flammability (red), reactivity (yellow) and special hazards (white) with numbers or symbols in each corner indicating the severity or type of special hazard. A summary is provided in Table 7.2. Full details are contained in the NFPA 704 standard. Figure 7.3. Example NFPA 704 (OSHA f)

40 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS The investigation into the Texas City incident resulted in 81 recommendations and led to majo r changes in process safety management that went far beyond the immediate causes of the incident. 3.2.1 Process Control Systems –the First Line of Defense The first line of automated defense is usually the process control system. These systems vary from simple analog or digital control instrument loops and alarms, through complex Di stributed Control Systems (DCS), and even to Advanced Process Control Systems (APC). APC refers to advanced control techniques, such as feedforward, decoupling, multivariable, and inferential control that are often used to automatically optimize and monitor the operating variable set points to achieve production targets with lower costs, while adhering to the safe operating limits. CCPS has published two book s that provide clear and concise information on guidance for control systems: Guidelines for Safe Automation of Chemical Processes (CCPS 2017c) and Guidelines for Safe and Reliable Instrumented Protective Systems (CCPS 2017b). When properly designed, tested, and maintained, th ese systems can provide a highly reliable approach to managing many abnormal process situations that were generally anticipated and had th erefore been identi fied during the design process. Some abnormal situations, however, may not be anticipated at the process design stage, when the protective safeguards are being developed, or when they are subseq uently modified. These may include, but are not limited to, a liner that co llapses in a vessel, agitators that decouple, a packing collapse in a column, a chemical composition change, a plugged instrument sensing line or nozzle, and in some cases, processes being impacted by certain weather extremes such as freezing of instrument lines or process vent s. The ability to troubleshoot and quickly recognize such issues as ab normal situation-conditions is the goal of this book. Chapter 5 of th is book discusses several tools and methods that fall under Hazard Identi fication and Risk Analysis (HIRA), including: Hazard and Operability An alysis (HAZOP), Fault Trees, and other methodologies that can be he lpful in pre-emptively identifying abnormal situations at the design stag e to ensure that adequate controls are provided.

141 researcher is asked to assign a “lev el of severity” to each factor, and then total all the figures. A si mplified example is shown below: Index Sheet for the Proposed Di-lithium Crystal Process: Toxicity/Flammability/Reactivity: 5 Quantity: 4 Reaction Conditions (T, P, corr/ero, dust, operability): 2 TOTAL: 11 Comments: Need additional proposals From an IS standpoint, the higher the resulting number, the higher the hazard of the chemistr y or process. Several alternative chemistries and/or processes must be proposed , and an index sheet completed for them as well. If competing chemis tries and/or processes exhibit a lower total “level of severity,” th en the researcher is obligated to defend his choice. No chemistry that exhibits "severe" factors in all categories is accepted. 8.3.1 Inherently Safer Synthesis Personnel engaged in research activi ties, e.g., chemists, engineers, and other scientists have many possi ble opportunities to incorporate inherent safety in the developmen t of process technology, including: synthesis routes catalysis leading to less severe operating conditions the use of a less reactive reagen t, or enzyme-based chemistry and bio-synthesis reduction or elimination of hazardous solvents reduction of solvent hazards by tempering a reaction by using a more volatile solvent that will boil and more reliably remove the heat of reaction immobilization of hazardous reagen ts and catalysts by attaching active groups to polymeri c or immobile backbones dilution of reagents reactions in water as opposed to those that proceed in a hazardous solvent elimination of hazardous unit operations

ACRONYMS xix PSSR Pre-Startup Safety Review (Ch. 5) PSV Pressure Safety Valve (Ch. 6) RAGAGEP Recognized And Generally Accepted Good Engineering Practice (Ch. 3) RBI Risk Based Inspection (Ch. 3) RBPS Risk Based Process Safety (Ch. 1) RCM Reliability Centered Maintenance (Ch. 3) SA Situational Awareness (Ch. 3) SIS Safety Instrumented System (Ch. 3) SME Subject Matter Expert (Ch. 4) SMS Safety Management Systems (Ch. 7) SOP Standard Operating Procedure (Ch. 4) TOH Transient Operation HAZOP (Ch. 5) UCDS User Centered Design Services (Ch. 1) VCE Vapor Cloud Explosion (Ch. 2) VDU Vacuum Distillation Unit (Ch. 7)

CONTINUOUS IMPROVEMENT 163 Management review can also be applied to improve how abnormal situations are addressed by the facilit y. However, the review should be part of a larger review process that addresses any weakness in RBPS elements, including those that are re lated to abnormal situations. The depth and frequency of each manage ment review should be governed by past incidents and abnormal si tuations, in addition to results obtained through auditing, metric s, and previous reviews, and management’s view of perceived risk posed by abnormal situations. 6.6 SUMMARY This chapter has highlighted the importance of integrating the management of abnormal situations within a facility’s existing safety management system, as many of the RBPS elements can be applied to reduce abnormal situations. Th e aim should be continuous improvement in how a facility addre sses abnormal situations. To this end, four elements of RBPS (metrics , incident investigation, auditing, management review & continuous improvement) are the primary elements that can deliver continuous improvement. Chapter 7 provides an in-depth revi ew of three case studies detailing serious events that could have been avoided by the implementation of a continuous improvement process for managing abnormal situations.

358 In the 2007 ISO Annual Report, CCHS reported that the number and severity of Major Chemical Accident s or Releases (MCARs) have been decreasing since the implementation of the ISO. However, the small number of MCARs (fewer than a dozen total incidents in any given year since 1999) makes it difficult to demonstrate a linear trend, or to establish a direct causal relati onship between the ISO and/or implementation of ISS and the numbe r of incidences. Figures 14.1-14.3, taken from the 2017 ISO Annual Report (Ref 14.6 CCHS 2017), display, respectively, the number and severity of MCARs that have occurred in the county since 1999–incidents at the eight ISO/RISO and CalARP covered facilities, Figure 14.1: Major Chemical Accidents and Releases (MCAR)

HUM AN FACTORS 263 Figure 11.1 Common Human Factors M odel (CCPS, 2007) Workers interact with facilities an d equipment and management systems every day. Human performance problems are typically the result of these complex interactions. Facility designers should strive to design equipment to meet workers’ expectations, which may vary throughout the world. For example, to turn on a light switch in the US, the switch is pushed up, but to turn it on in Europe it is pushed down. Color- coding schemes may vary from plant to plant. The best approach is to ask the end users abou t any local practices for equipment operation. A good human factors design is import ant. For most normal operating conditions, the human operator can cope with the incremental additional mental load of inconsistencies. Duri ng emergencies or other high-stress periods, however, each ad ditional mental task is an opportunity for error. Examples: 1. Conforming to certain expected conventions and meeting normal patterns of actions and habits can enhance human performance. The incident investigation team should be alert for built- in design deviations from normal conventions. In some countries, people expect the hot water tap to be on the left side and the cold water on the right side. W hen this is not the case, they can become confused and make mistakes. Rising- stem gate valves are expected to clos e if the handle is turned in a clockwise

DETERM INING ROOT CAUSES 237 At this point, the investigation team reaches a stage where they have more than one hypothesis for the reason the isop entane line ruptured. The pressure could have exceeded the desi gn pressure for the pipe or the pipe could have failed at a point below the design pressure. The team could use a simple fact-h ypothesis matrix to decide which branch to pursue. An example matrix is shown as Figure 10.20. Figure 10.20 Fact/ Hypothesis M atrix for the Kettle Exit Piping Failure In this example, assume the team obtained pipe samples of some of the remaining pipe and finds evidence of external corrosion. The team concluded that the feed line failed du e to higher than normal pressure combined with corrosion of the pi ping system (an AND-gate). These relationships are shown in Figure 10.21.

APPENDIX D – REACTIVE CHEMICALS CHECKLIST 485 heat is generally only removed through an ex ternal surface of the reactor. Heat removal capability increases with the square of the lin ear dimension. A large reactor is effectively adiabatic (zero heat removal) over the short time scale (a few minutes) in which a runaway reaction can occur. Heat removal in a small la boratory reactor is very efficient, even heat leakage to the surroundings can be significan t. If the reaction temperature is easily controlled in the laboratory, this does not me an that the temperature can be controlled in a plant scale reactor. You need to obtain the heat of reaction data discussed previously to confirm that the plant reactor is capable of maintaining the desired temperature. 6. Use multiple temperature sensors, in di fferent locations in the reactor for rapid exothermic reactions. This is particularly important if the reaction mixture contains solids, is very viscous, or if the reactor has coils or other internal elements which might inhibit good mixing. 7. Avoid feeding a material to a reactor at a higher temperature than the boiling point of the reactor contents. This can cause rapid boiling of the re actor contents and vapor generation.

172 Vessels with passive design can fully withstand any achievable overpressure without exce eding the yield stress of the materials. If the overpressure in a vessel remains in the elastic range, the metal returns to its normal crystalline state after st retching. Systems designed to “bend but not break” slightly exceed the plastic region of the metal and are deformed (hardened). The vessel is th en actually made stronger by this process; however, the new hazard that is introduced is that the vessel will not stretch and will usually burst if the scenario is repeated. Thus, vessels subjected to plastic range stresses require more frequent inspections for deformation and integr ity. A truly passive design is not only safer, it is more cost effectiv e when the lifetime te st and inspection requirements are considered. A robust design must be extended to include all system hardware elements. Little is gained if contai nment is lost when connected piping, joints, or instruments are designed for lower pressure than the vessel and then fail due to overpressure . In designing the process and equipment, the same engineering prin ciples described in Section 8.3 should be used to minimize the accu mulation of, and to contain, energy or materials: Specify design pressures high enough to contain pressures generated during exothermic re actions and avoid opening the relief valve and/or exceeding the maximum allowable working pressure of the vessel (the sa fe upper pressure of the safe operating envelope). Use physical limits of pipe size , restrictive orifices, and pump sizing to limit excessive flow rates. Use gravity flow in the equipment layout where feasible to minimize the need for pumps or solids handling equipment for hazardous materials. Review injection points and piping runs for erosion concerns. Design injection points, elbows, turns, and other erosion-prone areas for lower velocities. Use materials of construction with low corrosion rates. Use materials of construction that are applicable over the full range of operating conditions, su ch as normal startup, normal shutdown, emergency shutdown, and system draining. For

68 Human Factors Handbook Table 6-5: Pros and cons of electronic job aids Pros Cons Can hold a large amount of information Can provide interactive functions Can be centrally updated Can provide enhanced functionality, such as entry of data and completion of records Can enable automated calculations to be performed Can have look up functions to search databases, such as lists of parts when maintaining equipment Can enable remote verification of a task step, a condition, or location when a second person is not available at the location Limited use on sites with zones in which an explosive or flammable gas atmosphere could exist. Potential problems in navigating through screens and sub-screens Screen size may create problems in legibility of information Glare and bright light conditions may reduce visibility. Complex imagery may be limited on a small screen Power (of the device) and internet / data transmission mode access may not be available all of the time Wireless devices may be subject to cyber intercept or attack. 6.5 Key learning points from this Chapter Key learning points include: • The need for a job aid depends on task safety criticality, frequency, complexity and time available to complete the task. • Many types of job aids are available. • The best type of job aid depends on the type of task performance.

54 Human Factors Handbook • Less frequent, more complex tasks Less frequent, more complex and crit ical tasks may benefit more from step-by-step instructions and checklists. An example is process start-up. Such tasks may be prone to errors (s lips and lapses), especially if they have many steps or take a long time . If the task is complicated and involves judgment and decision-mak ing, then SOPs and job aids can support “rule-based” performance. If the task is infrequent and complex, then it may be helpful to use decision-making aids, such as diagnostic flow charts. These can give operators the knowledge they need to decide what actions to take, especially in abnormal or unique oper ational situations, such as process upsets. • Time critical emergency response tasks Time critical emergency response tasks may be best supported by shorter and easy-to-read job aids. A very detailed and long SOP may not be practical if it cannot be applied in the time available to perform the task. The “Miracle on the Hudson” (see Chapter 2) involved the use of an emergency response procedure. The pilots started to work their way through the procedure but stopped when they realized that they would crash before being able to complete the procedure. A short “grab card” is likely to be mo re practical if an operator has to decide what actions to take within a few minutes, as when responding to an emergency. In this instance a single sheet of laminated paper could be the best format, stored in the control room for example. The grab card should be technically correct, up to date and specific to the process. 6.2.2 A flow chart for determining need for a job aid 6.2.2.1 Overview Figure 6-1 provides a flow chart to help judge whether a step-by-step guide or a job aid may be more useful for a task. The best type of procedure or job aid depends on: 1. The complexity of the task. 2. The frequency that the task is performed. 3. The importance or criticality of the task. 4. The time available to use the job aid and complete the task.

Evaluating the Prior PHA 61 3.3 PRIOR PHA TOPICS FOR ADDITIONAL EVALUATION In addition to the factors listed in Sectio ns 3.1 and 3.2, as well as Chapters 2 and 4, a few other items to consider when evaluating the prior PHA are addressed in this section. 3.3.1 Status of Prior PHA Recommendations Recommendations from the prior PHA should be resolved before the revalidation, but occasionally, due to their complexity or difficulty of implementation, some recommendations may still be unresolved. The recommendations that were resolved were either accepted or rejected. Regardless of their status, all recommendations should be considered during preparation for the revalidation meetings: Accepted Recommendations. An accepted recommendation usually results in some change to the process equipment or procedures. Thus, in most cases, there will be a corresponding MOC evaluating the hazards of the change, but not always. For example, there may be no explicit MOC for resolution of recommendations for “training.” Th e reviewer should look for the implementation of those changes when evaluating the operating experience (as discussed in Chapter 4). If no MOC can be found, or if it appears that the actions taken may not have adequately accomplishe d the intent of the prior team, those recommendations should be highlighted for reconsideration by the revalidation team. Rejected or Declined Recommendations. If a prior PHA recommendation was rejected or declined, is there any evidence that the prior PHA team was involved in that decision? Perhaps the prior team worded the recommendation poorly and it was rejected because management simply did not understand its value. Another possibility is that the prior team made a technical error of which the upcoming revalidation team should be aware. Perhaps previous management was willing to accept an elevated risk of that particular hazard and documented this decision. Regardless, rejected recommendations should be given special attention by the current revalidation team to verify their concurrence that no further action is warranted. Unresolved, Open, or In-Progress Recommendations. Any unresolved, open, or in-progress recommendations from th e prior PHA should be specifically reviewed during the revalidation. If the te am concurs that the risk is elevated, it should reaffirm the recommendation in the revalidation report. If the team

CON TIN UOUS IM PROVEM EN T 331 15.3 CAUSAL CATEGORY ANALYSIS Each company’s management style an d safety management systems have strengths and weaknesses. These strengths and weakness tend to influence the types and severity of incidents that might occur. An analysis of incident investigation findings, in terms of causal factors and root causes, may identify broad areas or management systems that co ntribute to a higher proportion of incidents. Causes of incidents that repeat over t i me may al so b e i n di c at i v e of a weakness in the investigation system (e.g., lessons are not being learned). The determination of these management system failures allows a broader, more effective approach to the reduction of common cause weaknesses and prevention activities than addressing individual causes might. Table 15.3 is an example of one way to accumulate this data for analysis by using causal categories.

180 Guidelines for Revalidating a Process Hazard Analysis Q T E Were all pertinent hazards (fire, explosion, boiling liquid expanding vapor explosion [BLEVE], toxicity, chemical burn, asphyxiation, etc.) associated with releases addressed in the PHA? Was all equipment containing HHCs, or that could contain HHCs, addressed in the PHA? (Compare th e analysis nodes to process flow diagrams and/or P&IDs) Was contamination of process chem icals addressed in the PHA? Was loss of utilities addressed in the PHA? Was the unit flare header (and knock out drum if one is provided) addressed in the PHA or in a separate PHA? Were the documented hazards consistent with hazards listed on safety data sheets (SDSs) for HHCs? Have incidents since the prior PHA revealed new hazards that were not previously identified? Was there evidence that the PHA team evaluated the range of effects for loss scenarios postulated (e.g., impact area, references to SDSs, health/safety effects possibly identified as part of the hazards of the process, use of a risk matrix)? Was there evidence the organization’s risk tolerance criteria were consistently and correctly applied? Were there recommendations for additional risk controls where consequences of interest were lis ted in the worksheets with no (or weak) existing safeguards? If the prior PHA included recommendations for operability improvements, were their resolutions documented? Were any safety implications of the econ omic loss scenarios overlooked? Did the documentation for closure of previous safety recommendations meet all requirements?

36 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS Example Incident 3.2 – Union Carbide, Bhopal, India 1984 – (cont.) Lessons learned in relation to abnormal situation management Abnormal situation recognition: No Management of Change (MOC) was in place to approve the plant operating conditions during the accident. Abnormal situation status was either not recognized or more likely was a ccepted without a risk analysis. Organizational rol es and work processes: Personnel response to the accident was inadequate and lacked direction from the supervisor. Plant site culture appeared to accept operating without critical safeguards functioning and appeared to allow “Normalization of Deviance” in the plant’s process control limits and alarms. Knowledge and skills: Training and education for responding to upset situations and offsite response was inadequate. Process Monitoring and Control: Some of the MIC equipment was not functioning, was offline, or was inadequate to monitor the process and safety equipm ent properly. Compensating measures should have been considered to address the absence of key process safeguards. The long-term consequences of “Bhopa l” are still being felt, since the area has not been fully decontamin ated. The Institution of Chemical Engineer’s director of policy, Andy Furlong, stated in 2014: “… even though three decades have passed since Bhopal, we must never stop reminding ourselves that the lessons from the past are there to be learned, and crucially, acted upon.” (Furlong Press Release 2014) The BP Texas City Example Incide nt 3.3 resulted in tremendous losses, in both the human casualties and injuries and financial losses.

Appendix E – Classifying Process Safety Events Using API RP 754 3nd Edition E.1 Criterion for PSE This appendix covers how incidents can be classi fied using the guidance provided in API RP 754, 3rd Edition as is discussed in Section 9.3. The CCPS Process Safety Incident evaluation app can assist in classification of a PSE. This app is available at app stores. Classification of a PSE as Tier 1 or 2 requir es understanding of several criteria. The first criterion is that there must be a Loss of Primary Containment (LOPC). E.2 Criterion for Classification Tier 1 and 2 PSEs are LOPCs that occur within any 60-minute window with certain consequences, the severity of which determine th e classification tier. These consequences, as shown in Table E.1, include injury; direct co st from resulting fire and explosion damage; community impact; unsafe release from engineered pressure relief or upset emission from a permitted regulated source; and or acute release above a defined threshold quantity. Loss of Primary Containment (LOPC) - An unplanned or uncontrolled release of material from primary co ntainment, including non-toxic and non-flammable materials (e.g., st eam, hot condensate, nitrogen, compressed CO 2 or compressed air). (CCPS Glossary) Note: Steam, hot condensate, and compressed or liquefied air are only included in this definition if their release results in one of the consequences other than a threshold quantity release. However, other nontoxic, nonflammable gases with defined UN Dangerous Goods (UNDG) Division 2.2 thresholds (such as nitrogen, argon, compressed CO 2) are included in all consequences including, threshold release. (API RP 754). Primary Containment - A tank, vessel, pipe, transport vessel or equipment intended to serve as the primary container for, or used for the transfer of, a material. Primary containers may be designed with secondary containment systems to cont ain or control a release from the primary containment. Secondary cont ainment systems include, but are not limited to, tank dikes, curbing around process equipment, drainage collection systems into segregated o ily drain systems, the outer wall of double-walled tanks, etc. (CCPS Glossary) Process Safety Event – An event that is potentially catastrophic, i.e., an event involving the release/loss of containment of hazardous materials that can result in large-scale heal th and environmental consequences. (CCPS 2019)

| 343 APPEN DIX F: PROCESS SAFETY CULTURE ASSESSM EN T PROTOCOL F.1 Introduction The following questions that can be used to assess the status of the process safety culture in an organization. Some questions are intended to highlight evidence of a positive, while others help diagnose negative process safety culture. Like any checklist or protocol, it is inherently incomplete. Answers to questions m ay prom pt deeper investigation, and facilities may have cultural aspects that this protocol does not address. Any symptom identified through this protocol whose impacts are severe or have resisted correction should be subjected to a separate form al analysis. The questions in this protocol were derived from other sections of this book, particularly Chapter 4, which describes the relationship of process safety culture to each PSMS element, as well as from the references cited at the end of this section. F.2 Culture Assessm ent Protocol Establish an Im perative for Safety1. Has the organization adopted a minimalist approach to PSMS applicability? A minimalist approach refers to a conscious effort to limit the PSMS boundaries only to the strict limits defined by any applicable process safety regulations affecting the facility and nothing else. This is sometimes referred to as a compliance-only approach. The use of a minimalist approach is usually an overt decision but may also occur because of all of the actual process safety risks have not been fully evaluated. For example, the inclusion of utility, support, and other systems that do not contain any process safety related chemicals but are critical to process safety. The failure of some Essential Practices for Creating, Strengthening , and Sustaining Process Safety Culture , First Edition. CCPS . © 2018 AIChE . Published 2018 by John Wiley & Sons, Inc.

28 | 3 Obstacles to Learning In this section we examine some of the most common obstacles to learning from past incidents (Table 3.1). A more detailed discussion of obstacles to institutional learning was described by Schilling and Kluge (Schilling 2009). Keep in mind that not all companies experience every obstacle, and companies may overcome obstacles in the future; meanwhile new ones may arise. It is important to continuously assess the corporate learning process and to address obstacles as they appear. Table 3.1 Common Obstacles to Individual and Company Learning Individual Company Organizational changes Cost and business pressures Retirements and job changes Reverse incentives Natural memory loss and normalization of deviance Leaner organizations Insufficient or incomplete evaluation of hazards Risk misperception Lack of understanding about hazards Compliance-only mindset and over-anticipation of litigation Difficulty to see beyond own experience or type of industry Too many high priorities or rapidly changing priorities Both It can’t happen here attitude—loss of sense of vulnerability Ivory tower syndrome or lack of communication Assessing blame rather than correcting root causes Misplaced conservatism 3.1 The Impact of Individuals In any organization, the personnel roster is in constant flux. People change roles, get promoted, or leave the company for other opportunities. Organizational change can impact corporate memory in several ways. Almost by definition, each organizational change results in an incumbent being replaced by someone with less organizational memory related to that position. Unless the organizational memory has been captured in the company’s PSMS, standards, policies, and culture, memory will be continuously lost.

that they are committed to reducing the potential hazard zones in the area of a plant. •Develop methods to measure various inherent safety process options, an essential first step to widespread implementation. •Develop a method to measure inhere nt safety using “fuzzy logic” mathematics (a concept in which “ranges of truth” rather than discrete true or false values) - something that is now being researched. •Government programs now suppo rt concepts research and development, such as green chem istry, solvent substitution, waste reduction, and sustainable growth, all of which are related to inherent safety. A similar approach involving industry, government and academia ca n enhance inherently safer chemical processes discovery, development and implementation. 16.6 REFERENCES 16.1 Center for Chemical Process Safety (CCPS 1993). Guidelines for Engineering Design for Process Safety . New York: American Institute of Chemical Engineers, 1993. 16.2 Center for Chemical Process Safety (CCPS 1998). Guidelines for Design Solutions to Process Equipment Failures. New York: American Institute of Chemical Engineers, 1998. 16.3 Center for Chemical Process Safety (CCPS 2007). Guidelines for Safe and Reliable Instrumented Protective Systems. New York: American Institute of Chemical Engineers, 2007. 16.4 Dow Chemical Company (1994a). Dow's Chemical Exposure Index Guide , 1st Edition . New York: American Institute of Chemical Engineers, 1994. 16.5 Dow Chemical Company (1994b). Dow's Fire and Explosion Index Hazard Classification Guide , 7th Edition. New York: American Institute of Chemical Engineers, 1994. 439

216 Inherent Safety can affect thre at. For some adversaries, the reduction or absence of the hazard s may lessen or eliminate their interest in a target, or their opportunity to commit an attack. However, inherent safety strategies do not necessarily remove the threat, as a determined adversary may not be persua ded, especially if the inherently safer elements result in a safer operation but doesn’t completely eliminate the opportunity for them to inflict damage. Vulnerability is any weakness that can be exploited by an adversary. Vulnerabilities may include, but are not limited to: 1.structural characteristics 2.equipment properties 3.personnel behavior 4.locations of people, equipment and buildings 5.operational, cyber, an d personnel practices Vulnerabilities are estimates of an as set’s ability to withstand specific attack scenarios, considering existing security elements. The scenarios are usually derived from a combin ation of facility brainstorming (“knowing what I know abou t this asset, this is how I’d bring it down”) and intelligence estimates or other info rmation about the motivations and tactics of specific adversaries. The Department of Homeland Security provides specific scenarios to fac ilities required to conduct Security Vulnerability Analyses (SVAs) under CFATS. Vulnerability may be less affected by inherent safety than other factors unless sources are consolidated, quantities of dangerous chemicals reduced, or otherwise chan ged. It is more likely that the vulnerability is the same following the inherent safety application to the hazard itself (1st or 2nd order) , but, as mentioned before, the consequences may be different. Inhere nt safety applied to layers of security may be effective in making the layers robust, but this isn’t necessarily reducing the hazard. Attractiveness is the perceived value of attacking a given asset. The perceived attractiveness of a specific asset depends upon the goals of the attacker, and how well the potential consequences of a successful attack align with those goals. Other factors that may increase an asset’s attractiveness include:

350 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Safety Instrumented System (SIS) - A separate and independent combination of sensors, logic solvers, final elements, and support systems that are designed and managed to achieve a specified safety integrity level. A SIS may implement one or more Safety Instrumented Functions (SIFs). (CCPS Glossary) Safety Integrity Level (SIL) Discrete level (one out of four) allocated to the SIF for specifying the safety integr ity requirements to be achieved by the SIS. (CCPS Glossary) IEC 61508 is applicable to all industries and describes methods to apply, design, deploy and maintain instrumented safety-related sy stems. IEC 61511 supports IEC 61508 with a focus on the process industry. ANSI/ISA 84.00.01 is essentially the same as IEC 61511. IEC 61508 defines an engineering process called the safety life cycle intended to maintain the integrity of the safety instrumented system ov er its life cycle. This includes direction on assessment, design, and verification of SIS, amongst other topics. Safety integrity levels are defined as shown in Table 15.2. SIL can be thought of as a performance level for the safety instrumented f unction. A typically refinery process unit likely has several SIL 2 SIFs protecting the highest ri sk process units. SIL 4 systems are seen in the protection of nuclear power plants. Table 15.2. Safety integrity levels (IEC 61508) SIL Low-Demand Mode Probability of Failure on Demand (PFD avg) High-Demand (Continuous) Mode Probability of Failure on Demand (PFD avg ) per hour 4 >=10-5 to <10-4 >=10-9 to <10-8 3 >=10-4 to <10-3 >=10-8 to <10-7 2 >=10-3 to <10-2 >=10-7 to <10-6 1 >=10-2 to <10-1 >=10-6 to <10-5 What a New Engineer Might Do A new engineer will likely be engaged in hazard id entification and risk analysis studies, MOCs, or design activities where safeguards, barrier s, and IPLs are identified and evaluated. Understanding the difference between these terms and making sure that only the appropriate measures are credited is important to the inte grity of LOPA studies and SIF specification. New engineers may be involved in explaining process safety concepts or initiatives to operators, maintenance technicians and others. Models such as the Swiss Cheese Model and Bow Tie Model are simple to create and easy to understand and can be very helpful in process safety communication. New engineers may be tasked with engineering calculations related to the specification of common risk reduction measures such as the sizing of pressure relief valves and the

172 Guidelines for Revalidating a Process Hazard Analysis CRITERIA Yes/No Prior PHA Methodology (Section 3.1.1) Was the PHA method used from this approved list or was an appropriate equivalent methodology used? • What-If • Checklist • What-If/Checklist • Hazard and Operability (HAZOP) Study • Failure Mode and Effects Analysis (FMEA) • Fault Tree Analysis (FTA) Was the prescribed PHA method appropriate for the complexity of the process (i.e., was it suitably rigorous for identifying the process hazards present and evaluating their risk)? Was the prescribed PHA method applied in accordance with company guidelines? Prior PHA Inputs (Section 3.1.2) Did the qualifications of the PHA leader/facilitator meet all company and regulatory requirements (i.e., was the leader/facilitator trained, experienced, and competent in the PHA method used)? Did the PHA team make-up/qualifications meet all company and regulatory requirements include, as a minimum: • Operations team member(s) with adequate process and equipment knowledge, including recent hands-on operating experience? • An engineer with industry and specific process knowledge and experience? • An employee who has experience and knowledge specific to the process being evaluated? Was the PHA based on current, complete, and accurate process safety information (PSI)? Prior PHA Scope (Section 3.1.3) Did the physical scope include all equipment “covered” by regulation and company policy? Did the PHA address all the relevant operational configurations (e.g., did the team consider diffe rent batch operations, seasonal

48 Human Factors Handbook accidents. This is especially import ant for emergency or unplanned events and novel situations where no set rules or no previously agreed course of action exist. These references can also help people write operating and maintenance procedures and instructions. It is important to note that the provision of job aids is not an alternative to training. It is not feasible for people to develop knowledge and skills just by reading procedures. Training, job aids and having a defined safe way of performing tasks work together to support successful task performance. Be aware, that the dissemination of too many job aids and written procedures, can create risks including: • Operators may be unable to navigate th rough or be able to identify the relevant aids for the work at hand, • Operators may develop a perception that the volume of procedures is excessive and unrealistic, and • Operators may develop an excessive reliance on procedures instead of thinking and applying their knowledge and judgement. There may be situations, such as em ergency response and process upsets, where actions cannot be defined in detail due to the number of potential events and complexity of responses. 5.4 Approach to developing effective job aids 5.4.1 Attributes of effective job aids The attributes of effective job aids include: 1. Fit for purpose 2. Valid specification of a safe operating procedure 3. Practical and easy to use 4. Intuitive 5. Unambiguous and succinct 6. Accepted by users 7. Up to date More information can be found in the CCPS Guidelines for Writing Effective Operating and Maintenance Procedure [25]. Fit for purpose The type of job aid selected should be the most appropriate way to communicate the information and advice it provides. It should match task demands. For example, a manual can provid e an understanding of a process, while a “grab card” may provide brief reminders of emergency actions in times of limited response. A flow chart may communicate how to decide on an emergency

332 Human Factors Handbook 25.5 Sharing and acting on human performance indicators Lessons learned about human performa nce should be fully understood and shared. This sharing can take place during toolbox talks and team briefings, or during any other meetings between operators and supervisors. Lessons should focus on successful events as well as failure events. Successful events provide opportunities to learn from activities that helped improve human performance. Lessons learned should be fed back into the wider organization, in order to show others how to benefit from the experiences. In particular, where other business units may experience similar issues or be exposed to similar error traps. The effectiveness of lesson sharing should be evaluated, and if actions arising from the lessons learned are appropriate to another unit or team, they should be implemented. The action implementation cy cle shown in Figure 25-5 should result in improved company-wide performance. Figure 25-5: Lessons learned – knowledge sharing Reflect on performance Identify lessons learned Explore lesson sharing Monitor indicators of performance Share lessons learned Check effectiveness of lesson sharing Implement actions

Heat Transfer Units 215 Generally speaking there is more than one burner in each fired heater, therefore a burner header may be needed. It is very important to provide the fuel to several burners of a fired heater with the same flow and pres - sure. This required distribution is specific to the fuel pipe of burners. It discussed in Chapter 17, there are two main types of flow distribution, tree type and loop type. As  there is more chance of pressure and flow swinging in fuel oil  –  rather than fuel gas  –  the fuel oil is distributed amongst the several burners of a fired heater through a loop distribution. A ring header around the furnace pro-vides a fairly even fuel oil for the burners.Figure 11.19 shows the P&ID of a fuel gas burner with pilot gas. As part of SIS, if for whatever reason the flame put off, the unburnt fuel‐air mixture inside of furnace should be displaced as soon as possible. The steam can be used for this purpose. This steam is known as “snuffing steam. ” 11.13 Fire Heater Arrangement Fired heaters are rarely placed in series or parallel. Fired heaters are such expensive pieces of equipment that it is preferred to only have one of them in service. Burner To Burn erTo Burn erTo Burn erTo Burn er To Burn er To Burn er To Burn er To Burn er Duplex FiltersDuplex Filters Natural Gas headerANO THER Natural Gas headerTo Burn er PT 123 PC 123SD SDSD SDMajn gas ring Pilot gas ringFuel Gas Fuel gas ring for one fi red heaterFuel ga s ring for one fi red heaterPilot Gas Figure 11.18 P&ID of a fuel gas burner with pilot gas . Control Valve Station Fuel PeparednessBurner BurnerBurner Header (for each heater) Safety Shutdown Valving SystemSafety Shutdown Valving System Safety Shutdown Valving System Figure 11.19 Fundamen tals of the fuel route to burner.

16 substituting less hazardous materials, using less hazardous process conditions, and designing processes to reduce the potential for, or consequences of, human error, equipment failure, or intentional harm. Overall safe design and operation options cover a spectrum from inherent through passive, active and procedural risk control measures. There is no clear boundary between inherent safety and other strategies.” This is consistent with the definition of inherent safety and its related terms found in the current CCPS Process Safety Glossary (Ref 2.12 CCPS Glossary). In North America, the State of New Jersey, the State of California, Contra Costa County, California, and the federal Environmental Protection Agency are currently the only regulators with mandatory IS-related requirements. In Europe the Belgian Chemical Risks Directorate has published guidan ce for the implementation of the Seveso III Directive (Ref 2.14 Seveso) in Belgium which includes guidance on inherent safety, although this guidance is not mandatory (Ref 2.5 Belgian). Similarly, the UK HSE ha s published extensive non-mandatory guidance on the subject of inherent safety (Ref 2.26 UK HSE HSG-143) (Ref 2.27 UK HSE HSL 2005). The defini tion provided above is consistent with the definitions used by these tw o regulators. Both jurisdictions have used the definitions provided in the previous edition of this book to formulate their Inherent Safety (I S) requirements, and both require mandatory IS reviews/analysis but not mandatory process changes. See Chapter 14 for a detailed discussion of inherent safety regulatory initiatives implemented in the United States. 2.3 SHARED CHARACTERISTICS While the above definition succinctly captures what inherent safety includes, its application encompa sses a broader set of shared characteristics: It is a concept - Inherent safety (and related terms such as inherently safer design, inherently safer technologies, etc.) represents a set of concepts rather than a distinct set of rules. In this book, the terms and acro nyms Inherent Safety (IS), Inherently Safer Technology (IST ), and Inherently Safer Design (ISD) will be used interchangeably. All three terms are used in the literature and in the field, but actually have different

302 INVESTIGATING PROCESS SAFETY INCIDENTS conditions during the time period from 48 hours up to 1 hour before the occurrence, and a third section may address th e background immediately (1-60 minutes) before the occurrence. The background sections may also include information on past incidents in the process unit, including past incide nts that are identica l or nearly so to the actual incident (a “rep eat incent”). Near-misse s and minor incidents are of interest to determine if th ere were any precursor events. 13.4.4 Sequence of Events and Description of the Incident In this section of the report, the occurrence is descr ibed (usually in chronological order) and the outcomes are identifi ed. This is the WHO– WHAT–WHEN–WHERE portion of the report. It includes the actions taken to deal with the situation throughout the timeline of the event. It may give precise and specific information, such as identification numbers and location of process equipment involved in some fa cet of the incident. The extent of injury, details of the da mage, and an estimated ou t-of-production time can be included in this section. Diagrams are often more useful than long paragraphs. If a timeline has been dev eloped, it may be included in the report. The observations can be backed up with statements from those involved. Supporting documentation in the form of drawin gs, photographs, flow diagrams, and calcul ations can be included. 13.4.5 Findings Factual findings are presented in this section. The findings flow from all investigation activities including wi tness interviews, scene and equipment inspection, process data, laboratory tests, equipment testing protocols, engineering analyses, mode ling, etc. The findings provide the foundation for subsequent causal factor identi fication and root cause analyses. It may be helpful to mention the vari ous types of evidence that support the causal factors and root cause conclusions: People (interviews) Physical (for example, equipment, machinery, parts, analytical analysis, metallurgical analysis, testing ) Electronic (for example, operating data recorded by a control system, both current and histor ical, and controller set points) Positions (people and equipment) Paper (for example, procedures, ch ecklists, process data, permits, etc.)

244 | 7 Sustaining Process Safety Culture are not quickly brought up to speed on the culture and their PSMS duties, conflicts can arise that impact com munication and trust. Changes in process safety related policies and procedures: Changes of all kinds should be expected as the culture and the PSMS improves. These inevitably lead to new and potentially unfam iliar responsibilities and activities. These can cause stress and m ay require adjustm ents to com pensation and authority, along with the needed training. As culture is im proved, these personnel issues should be m anaged carefully. New responsibilities also need to be codified in job descriptions, so they can be sustained through future personnel changes. This will help build mutual trust and empowerment . Neglecting these issues m ay lead to resentment, making the culture improvement effort less likely to succeed. Lapses in leadership and failing to learn and advance the culture: As stated at the beginning of this chapter, any lapse of leadership can lead to norm alization of deviance and overall decline of process safety culture. Therefore, leadership of the process safety culture must com e from the top, be encouraged by the B oard of Directors, and cascade through the organization. To com bat this, good checks and balances need to be in place to review adherence on a regular basis. Central to success sustaining culture in the above examples, and in the overall life of the com pany, is making a firm and full com mitment to continuously improving process safety culture. This commitment can be m aintained through six critical success factors that are summarized in figure 7.1. Take Cultural Snapshots Leaders should remain alert to changes in the culture. Periodic reassessm ents are important, and included as one of the success

46 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS Example Incident 3.5 – Buncefield Explosion, 2005 (cont. ) The lighter fractions of the winter blend of gasoline had volatilized, forming a flammable cloud that was estimated to cover an area of some 80,000 m2. The immediate cause was the failure of the level gauge, which was stuck at a fixed reading, preventing the subsequent activation of a high-level alarm. An independent hi gh-high level alarm also failed to activate. The operators had failed to observe or act on the stuck level gauge. They also did not estimate the time to fill the tank, instead relying on the instrumentation. The level gauge was known to be unreliable and the high-high level al arm, a float switch, incorporated a test level that had not been corr ectly repositioned or fitted with a padlock after a high-level trip test had been conducted. Locking of the level trip instrument was not fully understood by site personnel and the UK HSE issued an alert notice for users of similar equipment at other sites. The Investigation Board report contained some 25 recommendations, several of which are relevant to a discussion on abnormal situations. Many of the recommendations were associated with improvements in the safety management systems, leadership, and culture on the site. Lessons learned in relation to abnormal situation management: Procedures: Provide effective st andardized procedures for key activities in maintenance, testing, and operations. Knowledge and Skills: Improve training, experience, and competence assurance of staff for safety-critical tasks and environmental protection activities. Work Environment: Define appropriate workload, staffing levels, and working conditions for frontline personnel. Communications: Ensure robust communications management between sites and contractors and with operators of distribution systems and transmitting sites.

53 4 Application of Process Safety to Wells 4.1 BACKGROUND This chapter addresses the primary equipment, risks, and key process safety measures involved in drilling, completions, workovers and interventions, referred to collectively as well construction. It show s how the concepts contained in RBPS can be applied to further enhance current design, operational practices, and process safety performance. Special attention is given to well control and the safety barriers for preventing and mitigating a loss of well control event. The final section reviews selected process safety tools a pplicable to well construction. Well construction terminology makes use of many unique terms and acronyms that may be unfamiliar to those new to the upstream industry. Basic definitions are included here to enable the reader to understand the topic, but for full definitions one may refer to industry reference books (see list below). In this book, the terms drilling, completion, workover and intervention are derived from the Schlumberger online glossary (www.glossary.oilfield.slb.com). All the terms are associated with a potentially pressurized well that must be managed to prevent a loss of containment event. Drilling: The process of creating a wellb ore down to the reservoir production zone, including casing and cementing to achieve all needed isolations. Completion: The collection of downhole tubulars and tools necessary for the safe and efficient production of hydrocarbons. It usually includes perforating the casing and isolating the well from the reservoir with explosives to provide access to the reservoir fluids. Workover or Intervention: The repair or stimulation of a well to enhance its production rate. It usually involves inserting tools into the wellbore. A workover requires the presence of a rig, whereas an intervention does not. Some process safety topics are common with production activities, covered in Chapters 5 and 6, and in some cases the read er is directed to those chapters rather than repeating the material in this chapter. This section provides an introduction to the technology necessary to understand the application of RBPS to well construction activities. Section 4.2 then addresses risks and key process safety measures, while Section 4.3 covers process safety methods and tools. Details on well construction technology may be obtained from industry references, including the following. ●IADC Drilling Manual , Vols 1 and 2 (IADC, 2015a )Process Safety in Upstream Oil and Gas © 2021 the American Institute of Chemical Engineers

82 | 3 Leadership for Process Safety Culture Within the Organizational Structure In other words, process safety leadership starts with the Board of Directors and senior leadership, and involves the entire organization. Everyone from the Board to the plant floor m ust have the necessary com petence in process safety. The PSLG recognized that some Directors, especially those com ing from other sectors, would not have competence in process safety. To com pensate, they recom mended that one Board member be highly experienced. See sections 5.3, 5.12, and Appendix D for m ore about com petence. After considering the leadership literature, Stricoff, PSLG, the expertise of the CCPS Culture Comm ittee (See page xix), and other sources, some broad themes of process safety leadership can be seen: Achieve a balance between m anagem ent and leadership: o Establish clear roles and responsibilities for managers and others who function as leaders. o Use management to define clear process safety work processes and manage them , o Manage organizational change; and o Use process safety metrics for decision making and balanced scorecards. Inspire subordinates and peers: o Display visible support through felt leadership, leading by passionate example, o Provide adequate, competent resources and annual budget for process safety; and o Follow through on verbal support with personal actions. These characteristics are expanded and described in m uch m ore detail in Sections 3.2 and 5.1, as well as in Appendix D. • •

APPLICATION OF PROCESS SAFETY TO OFFSHORE PRODUCTION 127 Fire protection Firefighting capability (water and foam) is available from hydrants, sprinkler and deluge systems on modern offshore installations and from monitors on offshore support vessels. One advantage offshore is there is an abundance of firewater supply – so long as power is available. However, there is limited line of sight to attack fires inside modules or deep within a congested layout other than using sprinklers or deluge systems installed above vulnerable equipment. Fire protection is a mix of active and passive systems. Active systems include fire and gas detection, alarms, fire water pumps, fire water ring main(s), hydrants/monitors, water sprays, foam sy stems, and deluge systems. Emergency shutdown and isolation of hydrocarbon flows is also an important element of active fire protection. Passive systems include fireproof coatings, fire walls, blast walls and drainage. Escape, Evacuation and Rescue An aspect for offshore operation is the greater difficulty of personnel to escape compared with onshore facilities. Jumping into the water is typically not a good option due to height, ocean conditions, and oftentimes remoteness from shore or other facilities. Offshore facilities typically conduct an EER (Escape, Evacuation and Rescue) study to plan for significant process safety events and to make sure escape pathways and safe refuges are defined, and adequate equipment (e.g., lifeboats) and plans are in place. Temporary safe refuges (TSRs), or temporary refuges, are designed to protect personnel against the possible process safety event consequences for some defined period of time, especially fire or smoke, until an evacuation is ordered, or the situation is resolved. Evacuation offshore is usually achieve d using lifeboats, although there are supplementary systems using escape chutes and life rafts. The authority having jurisdiction may set minimum requirements for lifeboats / life rafts. Helicopter evacuation is an option but it may not be poss ible to land on the helideck if there are flammable vapors present, fire and smoke, or the offshore installation is listing. An inherently safer evacuation exists for bridge linked platforms, where evacuation is by walking over the bridge. Rescue is typically achieved using an offshore support vessel, most of which have firefighting capabilities. These type of support vessels are mandatory in many offshore locations. Oil spill response is another impor tant consideration for offshore Emergency Management . Drain systems are designed to in tercept hydrocarbon leaks preventing them from flowing overboard. Booms and other mitigation equipment are usually present (on the facility or on a nearby support vessel) to contain spills. After the Deepwater Horizon incident, operators in the US and many around the globe joined to form response consortia. These maintain dispersants, oil spill containment, and recovery equipment suitable for rapid deployment to any significant spill incident.

20 INVESTIGATING PROCESS SAFETY INCIDENTS 5. Exposure factor : a factor that mitigates the potential effects of an incident Therefore, incident invest igators should consider not only what went wrong, but also corrective actions bas ed on second order ISD principles that could be taken to minimize the impact of future incidents. 2.2.2 Management System Failure Most active failures and latent failures, whether they are equipment deficiencies, human errors, or unsafe acts/conditions, are the result of weaknesses, defects or breakdowns in the management system(s). Consequently, there is a strong link between root caus es and management systems. Causal factors are unplanned contributors (negative events or undesirable conditions) to an incident, that if e liminated would have either prevented the incident, or reduced its severity or frequency. Therefore, a strong link also exists with causal fa ctors, as these negative events and undesirable conditions involve some of the active and latent failures that contributed to the incident. On rare occasions, an individual may deliberately damage a chemical process to cause an incident, but even then a management system weakness (such as facility security or employee fitness for duty) may be involved. Risk is a measure of human injury, environmental damage, or economic loss in terms of both the incident likelihood and its severity. One reason the management system concept has receiv ed broad recognition relative to chemical incident investigation is that it builds directly on fundamental process safety principles. To manage risk, appropriate management systems need to be in place to ensu re that the barriers agains t incidents remain intact. These preventive, error detection, and mitigation management systems make up the bulk of process safety e fforts. Examples of these include the 20 elements of CCPS’s process safety management system (CCPS, 2007a), such as operating and maintenance procedures, effective training, control of up- to-date process safety information, management of change, performance measurement, auditing, etc. As most root causes are associat ed with weaknesses, defects, or breakdowns in the management system(s), investigators should look for weak barriers. These weak barriers coul d be associated with various aspects of the management system, including, but not limited to, the attributes in Table 2.1.

74 | 6 Implementing the REAL Model during the investigation phase to aid in identifying root causes and causal factors; and then while developing recommendations. The review of incidents should go beyond considering similar industries and processes to include those with similar root causes and causal factors, regardless of industry or process type. Risk Register Many companies maintain a risk register. This is a document or system in which process safety and other risks are determined (for process safety via PHA and Layer of Protection Analysis) and sorted. Leadership uses this information as a guide in managing operations and looking for improvement opportunities. It is natural to focus process safety learning on the highest process safety risks. However, many companies recognize the benefits of also addressing potential scenarios with the most severe consequences, regardless of risk. As process safety pioneer Trevor Kletz said on many occasions, “What you don’t have, can’t leak.” Gut Feel and Warning Sign Analysis CCPS found that people interviewed during incident investigations frequently used the phrase, “I knew we were going to have that incident” (CCPS 2011b). Deeper inquiry showed that the interviewees came to that realization by noticing things that often don’t show up in metrics, audits, or other ways of measuring process safety performance. During a conference commemorating the 25th anniversary of the North Sea Piper Alpha disaster, Steve Rae, an engineer who escaped the platform, talked about the warning signs he had seen (Rae 2013): I want to share with you my initial thoughts on my arrival on Piper Alpha and the three months after: 1. I looked at Piper Alpha on arrival and thought wow, it looks old and tired … when in fact it wasn’t old at all. 2. Obvious that additional structures and modules had been added. 3. Somewhat confusing to navigate around, somewhat of a rabbit’s warren. 4. During nights there were few people around, in particular around the control room where I had to go permits to work. 5. There were times when Piper vibrated significantly, it almost felt like it would be shaken to pieces.

166 | 12 REAL Model Scenario: Overfilling have resulted in our country shoring up its sea defenses to meet the safety level of a flood chance once every 10,000 years for the west and once every 4,000 years for less densely populated areas. The primary flood defenses are tested against this norm every five years.” ”That’s all very interesting, but how does that apply to us?” Frederik said. Alexandre responded, “I wanted to show the great lengths and expense our country is going through to protect us from flooding. Then, I wanted to follow it up with how we are aligned, with our processes set up to handle a flood chance once every thousand years. We’re in good shape, but we have to make sure that when we revalidate our PHAs every five years, we check that our assumptions about flood protection remain valid.” They presented their findings to Jan the following week. Jan said, “Good work team. It’s a comprehensive solution that addresses our problems today, while looking toward the future. I’m willing to invest in upgrades, as long as there are solid justifications for them.” 12.7 Implement Frederik developed a communications plan for the new procedure on tank level readings after a severe storm. The plan focused on how this added verification step was meant to make the plant safer, which meant keeping everyone safer. He took some notes during Alexandre’s presentation on overflow incidents and planned to share a finding from the CSB, that “overfilling was cited as the most frequent cause of an accident during operation; among the 15 overfill incidents found, 87% led to a fire and explosion.” He also planned to show a brief video on Buncefield and Bayamon to drive the point home. With Reed’s help, Frederik rolled out the new procedure to the three shifts, and although there was some grumbling, the operators all understood that this extra work was for the greater good of the organization, and perhaps more importantly, for their own safety. Showing the videos really helped get the message across. Sometimes, pictures, or in this case videos, are worth more than a thousand words. Pamela, in the meantime, developed a list of skills that would be required of the new hires. She wanted to make sure that when they transitioned over to the radar gauges, there would be no issues around accepting the new technology.

138 PROCESS SAFETY IN UPSTREAM OIL & GAS Projects (CCPS, 2019b) provides advice as to which process safety studies occur when in the design activity. 7.4 PROCUREMENT AND CONSTRUCTION Once the detailed design is complete, with all associated process safety assessments, then the project goes out for procurement and construction bidding. Important RBPS Incident: Piper Alpha Design Issue The Piper Alpha incident has been described previously in Chapter 6. While primarily due to miscommunication related to work permit status that allowed a startup with parts of the pipework still open, an important design issue also existed. The facility had been very succes sful and highly productive. The operator sought an increase in production, and this was granted, subject to a condition that gas also be piped to shore and not flared. A gas treatment plant with compression was retrofitted, but the facility design was not changed to add blast walls to acc ount for the greater potential of an explosion event. Also, the compression module was close to occupied spaces (refer to Figure 6-2 in Chapter 6). Broadribb (2014) discusses important changes to design that occurred due to this incident and the Cullen Inquiry recommendations. Among other things, the Inquiry recommended that, forthwith, all facilities in the UK carry out certain studies to enhance inherently safer desi gn. These are known as the ‘forthwith studies’. ●Systematic analysis of fire and explosion hazards ●Analysis of smoke and gas ingress into living quarters, and the requirement for a temporary (safe) refuge capable of surviving the initial fire/explosion and any escalation for a r easonable duration to perm it evacuation and escape ●Analysis of the vulnerability of safety critical equipment or elements, such as emergency shutdown valves (ESDVs) ●Analysis of escape, evacuation and rescue in the event of major incidents One effect of the forthwith studies was th at the design of North Sea installations changed from essentially square cross sections to rectangular. This allowed for increased spacing between major hazard modules and vulnerable areas. Where feasible, bridge linked platforms were used to further increase separation to the accommodation module. RBPS Application Management of Change and Hazard Identification and Risk Analysis : The changes recommended by the Cullen Inquiry are now commonly applied to offshore installation design process safety studies and are embedded in UK regulations.

E.39 Playing the Odds |329 drill on the launch pad. The investigation determined that the oxygen atmosphere in the capsule caused a m inor electrical short to accelerate into a significant fire. The crew and launch attendants outside the capsule tried to open the hatch, but the com bustion gasses had raised the cabin pressure enough so that the inward-swinging hatch would not budge. B efore the incident, Apollo astronauts had expressed m any concerns about their new spacecraft, including a significant amount flamm able nylon webbing throughout the crew cabin. The investigation board noted that NASA had failed to identify flam mability hazards so that they could have been addressed. During the investigation hearings, an astronaut termed the failure to connect flamm ables plus oxygen to fire was a “Failure of Im agination.” Of course, it was not a failure of im agination because the Apollo 1 crew had im agined it – and have even com plained about it. If the crew com plained about a safety problem was there was an understanding of hazards and risk , but a failure at some level of the organization to act on these hazards and risks ? Were the crew aware of the hazards but other astronauts failed to imagine it? If so, was there a gap in open and frank communication ? Did the others not have the same sense of vulnerability , or did they not trust their colleague’s judgment? Establish an Imperative for Safety, Maintain a Sense of Vulnerability, Understand and Act Upon Hazards/Risks. E.39 Playing the Odds A young engineer overseeing his first plant trial batch was discussing the first step of the operating instructions with a 35-year experienced operator. “We can skip the inerting step,” the operator said. “That will save us some time to have coffee and eat those nice donuts you brought for me and m y buddies.”Actual Case History

222 Figure 9.2: Elements in a CCPS Security Vulnerability Assessment (Ref 9.4 CCPS)

APPLICATION OF PROCESS SAF ETY TO ENGINEERING DESIGN, CONSTRUCTION AND INSTALLATION 133 The three FEL stages carry out similar process safety activities, but with greater refinement as the project option is select ed and refined. Some companies develop a process safety in design philosophy to ensure the appropriate process safety activities occur at each stage. 7.2.1 FEL-1 The primary activities in FEL-1 relate to Compliance with Standards (particularly the regulatory aspects that must be complied with) and Hazard Identification and Risk Analysis . At this stage, there are multiple options but with few details yet developed. This may appear as simple block diagrams of major parts of the process. Decisions address location, technology, proc ess and how to achieve inherently safer design. Standards and prior projects help with some design assumptions, and HIRA takes the form of simple checklists of potential hazards (e.g., CCPS, 2019b Table 3.1). The aim at this stage is to rank possible project design options based on likely economic advantages, process safety (including environmental impact), and other factors such as project risk. It is good practice to develop a risk register at this stage. This risk register lists hazards, their potential cons equences, their likelihood, and overall risk. The register is used to support project decisions and is updated as the design progresses. This helps to ensure that no identified risk is neglected when there is a design modification and MOC, or if the project team changes or new personnel join, as happens over the life cycle. The risk regi ster may also identify inherently safer options where some risks may not be present at all, and thus not require mitigation. Inherently safer design (ISD) activities start in FEL-1 when there is the greatest potential to eliminate or minimize risks. Onshore this might be to choose options with immediate export of produced gas and liquids and without any local storage. Offshore this might include some partial subsea processing or a very low staffing option for operating “from the beach”. An exam ple is presented later in this chapter that further illustrates ISD concepts. Some projects carry out an initial Concept Risk Analysis (CRA) based on assumptions for inventories and process conditions. CRA can be a simplified QRA or a simpler assessment limited to potential consequences of worst case or maximum credible events that uses basic information only. It may use prior similar designs to allow sensible assumptions. At the FEL-1 stage, the CRA is more of a screening level risk estimate, and it is useful for comparing different options or identifying major risks that may be difficult to mitigate at later design stages or in operation. 7.2.2 FEL-2 The primary objective of FEL-2 is to select the option to take forward in the project. As before, the primary activities in FEL-2 relate to Compliance with Standards and Hazard Identification and Risk Analysis but now sufficient detail is available such that issues related to Asset Integrity and Reliability can be defined.

83 | 6.6 Prepare Ideally, the team will also include one or more individuals who will ultimately be responsible for implementing any changes that result from the team’s work. The team develops recommendations addressing the corporate learning objectives, just as a team investigating an internal incident or near-miss would. The team may also consider developing additional recommendations that address findings of interest from external incidents. CCPS covers the recommendations process in detail (CCPS 2019b). In summary, the recommendations must: • be SMART, that is, specific, measurable, attainable, realistic, and time bound • address the corporate improvement goals developed in Step 1 • address specific root causes and/or causal factors • result in improvements to one or more PSMS elements, standards, policies, or business practices that apply broadly • map the action to the learning and demonstrate how risk will be reduced. Ways to reduce risk include: o correcting an ineffective barrier o reducing potential consequences o reducing the probability of occurrence o a combination of the above. The team may offer more than one recommendation addressing any given corporate improvement objective. The recommendations may be either alternative, additive, or both. Alternative recommendations offer multiple options for addressing a given improvement objective; after evaluation, the best can be selected. Additive options work together, with each option moving the company closer to meeting an improvement objective. 6.6 Prepare In this step, recommendations become action plans. These may involve one or more of the following: • changes to a policy • changes to a standard • capital expenditures • new risk-reduction measures

235 stages), and benefits of ISD—such as reduced inventories, smaller equipment, and fewer add-on safeguards to be maintained, etc.—as well as its limitations. The evaluation of both initial and on-going costs associated with inherently safer de signs can often help enhance this understanding. Many IS improvements can be implemented cost- effectively, particularly incremental improvement in existing plants. Kletz (Ref 10.18 Kletz,) discusses cultural and management barriers to the implementation of inherently safe r designs, as well as the actions needed to overcome these barriers. Khan and Amyotte (Ref 10.17 Khan) also provide guidance for making th e use of inherently safer design principles more routine. Turney (Ref 10.22 Turney) lays ou t five steps, taken from the European Process Safety Centre’s St atement of Good Practice, which are necessary for effective adoption of inherent safety within an organization: Support by a champion Suitable training Application from the earliest stage of a project Reviews throughout project development Recognition and reward of th ose involved in the project By incorporating inherently safer design into the organizational culture, inherent safety becomes an on-going way of examining and addressing processes and their ha zards, and this philosophy then permeates all aspects of the process safety management systems. For example, once its application is fully understood, the concept of simplification with respect to human factors can become an inherent aspect of the writing of standard operating procedures. 10.4.1 Multiple Demands of IS in the PSM program Dowell (Ref 10.11 Dowell) points out that excellence in operations, which must include personnel and process safety, requires a comprehensive management system approach. The challenge is to integrate the many ESH compliance standards (PSM, RM P, Responsible Care®, RP750, company requirements, etc.) and spec ial program activities (ISO 9000, Sustainability, etc.) into the fabric of the system of making chemicals, which, in turn, is grounded in company culture. These activities must go

108 | 4 Applying the Core Pr inciples of Process Safety Culture Human behavior usually contributes to process safety incidents through human errors. This includes acts of omission – som ething a person fails to do – and acts of commission – som ething a person does that they should not. Experience and research has shown that some hum an error is inevitable. B ut hum an error can be affected by so-called perform ance-shaping factors, stresses and influences that increase or decrease human error. Many of the performance-shaping factors can be m anaged through the practice of human factors design, discussed in depth by CCPS (4.1). Culture also plays a significant role. A person could rightfully ask, “Does behavior create the culture, or does the culture create the behavior?” Arguments can be m ade either way. A key prem ise of this book has been that the behavior of people engaged in any set of tasks will be affected by the culture surrounding those tasks. However, the culture itself can be strengthened or weakened by behaviors. Indeed, behaviors aligned to the culture core principles discussed in this book should strengthen culture. Meanwhile, behaviors counter to the culture core principles, such as breaking trust or norm alization of deviance, weaken culture. The research of Daniellou (Ref 4.2) supports this notion of hum an characteristics that can be influenced by culture to drive behavior. These characteristics can be driven by positive culture to create good safety behavior, and can be driven by negative culture to threaten safety. Daniellou noted that hum an errors, though generally unintentional, are made as the result of conscious acts performed without malice. That is, people choose to perform incorrectly, and do so with good or neutral intentions. In this context, associating error with words such as “fault” or “liable” is doubly counter- productive. Not only does this prevent the organization from identifying the real reason, it also prevents open and frank

21. Fostering situation awareness and agile thinking 269 Table 21-3: Clues for recognizing impaired Situation Awareness Impaired situation awareness Communication tips to resolve the loss of situation awareness Ambiguity – information from two or more sources does not agree. Ask probing questions, such as: • Why do you think these two pieces of information are different? • How can we reconcile the differences? • What can cause the different readings? Fixation (tunnel vision) – focusing on one thing and excluding everything else. Use gentle “nudges” or questions, such as: • Have you thought about the bigger picture? • What may be the impact of this failure on the plant? • Which other factors may have caused this? Confusion – uncertainty or bafflement about a situation. Use gentle “nudges” or questions, such as: • Which aspect are you not sure about? • Can you explain to me what is happening? • Why do you think this is happening? Lack of required information. Use of leading questions, such as: • Have you considered checking pressure levels/valve B? • Who is the right person to ask about the situation? • Who would you normally ask for assistance? • Would anyone else have access to this information? Failure to: • Maintain the task. • Meet expected targets or check points. • Resolve discrepancies. Ask a series of questions to identify shared situation awareness, such as: • How is the task progressing? • Why is it not progressing the way it should? • Is there anything I could help you with? • Who else is working with you on this task? • What are your colleagues or teammates views on these discrepancies?

Manual Valves and Automatic Valves 125 It is very confusing that in tagging regulators CV repre­ sents control valve, but it is not the tag for control valve, but for regulator. “Process parameter” here can be any process parameter including “P” for pressure, “F” for flow, “L” for level, or “T” for temperature. The third footprint of regulator on P&ID is their set point. It is again important to recognize the difference between control loop and regulator here. On P&IDs, when there is a control loop, the set point of control loop is not shown on P&ID. The set point of control loops and P&IDs can be found in “Control and alarm set points table. ” The set points of regulators are generally noted on P&IDs. Therefore, if there is a pressure regulator on a P&ID, the set point of the regulator should be men­tioned beside the regulator on P&ID, for example, “set point; 20 kPag. ” For a flow regulator the set point is mentioned beside the flow regulator symbol like this: “set point; 30 m 3 h−1.” Regulators similar to control valves have failure posi­ tion. They could be “FC” or “FO” or “FL. ” However, the difference between regulator failure position and control valve failure position should be recognized. In control valves the designer has opportunity and the freedom to choose his/her favorite failure position. But this is not the case for regulators. Regulators have a natural built‐in failure position that cannot be changed by the process engineer. For example, the failure positon of all pressure regulators are FC. The failure position of all backpressure regulators is FO and the failure position of flow regulators is always FL. Figure 7.25 shows the set point and failure position of regulators on P&IDs. The last thing about regulators is that they do not generally need bypass like what we see in control loops. As it was discussed we generally have a specific arrange­ment that we may need control loop station but there is no such thing for regulators. As a rule of thumb, whenever we need to automatically control a process parameter, we need to put a control loop including sensor, controller, and control valve. However, in some cases, this system can be replaced by a single regulator. Those are the cases in which we are seeking simplicity in the system and the service fluid is not dirty and the pipe size is not very big. Regulators are plugged easily if they are used in dirty services. Regulators also are not available in big sizes, say, larger than 6 in. 7.15.3 Saf ety‐Related Valves The safety‐related valves are the valves that act when a process parameter wildly violates from its normal level. This violation could be in the lower or higher side of the parameter. Therefore unsafe condition happens when a process parameter goes to the higher side of the normal level or the lower side of that. Therefore it can be said that safety‐related valves act when a process parameter goes much higher or much lower than its normal level. As there are five different process parameters, namely, flow, pressure, temperature, level, and composition (Table 7.20), it can be assumed that there are at least five different safety‐related groups of valves, which are roughly correct: Pressure relief valves are the valves that open when pressure increases and reaches a preset value. Vacuum relief valves are the valves that open when pressure decreases and reaches a preset value. Temperature relief valves are the valves that open when temperature increases and reaches a preset value. Table 7.20 Differ ent groups of safety‐related valves. Process parameter Corresponding safety valve Action Flow Excess flow valve Closes when flow goes beyond a preset value Pressure Internal Pressure relief valve Opens when pressure goes beyond a preset value External Vacuum relief valve Opens when vacuum goes beyond a preset value Temperature Temperature relief valve Opens when temperature goes beyond a preset value Level No valve, only overflow nozzle “Opens” when level goes high Composition Not available Not availableSet point: 20 kPag FC Set point: 50 kPag FO FL Set point: 30 m3h–1 Figure 7.25 Regula tor set point and failure position.

22 Guidelines for Revalidating a Process Hazard Analysis HIRA is a term that encompasses all activities involved in identifying haz- ards and evaluating risk at facilities, throughout their life cycle, to make certain that risks to employees, the public, or the environment are con- sistently controlled within the organization’s risk tolerance. A PHA is one form of HIRA and is sometimes required to meet specific regulatory re- quirements. The concepts of revalidation described in this book can be applied to a PHA or HIRA to address the key principles of RBPS: • Maintain a depend- able practice • Identify hazards and evaluate risks • Assess risks and make risk-based decisions • Follow through on assessment results Manage Risk and Learn from Experience. From the time an initial PHA is completed on a process until the PHA is revalidated, risk is being managed throughout the revalidation cycle. Changes are implemented, incidents are investigated, work is performed, operat ing experience is gained, and external conditions may change. The PHA revalidatio n is an opportunity for the facility to look back at this operating experience an d learn from it to reduce risk in the process. Whether the PHA revalidation is a focused Update , a complete Redo , or a combination of these approaches larg ely depends on what has happened in the process since the last PHA. The PHA can be used to: • Identify which safeguards should be classified as critical • Help develop preventive maintenance schedules and priorities for safety- critical instrumentation • Identify scenarios for emergency response training drills • Identify which written operating procedures should be classified as critical • Help develop a training program for new operators • Establish the baseline for MOC reviews

186 Adequate IS related documentation is the first layer of defense against losing the technical rationale for de cision-making that resulted in IS measures. The Management of Change (MOC) program is the next layer of defense. The importance of MO C is addressed in Section 8.7. Using inherently safer principles , there are some Simplification improvements that are possible duri ng construction and commissioning or operations stages of life that have minimal impact of the design and functionality of the process and ca n be accomplished at modest cost. Most of these improvements make the process more tolerant of human error, which is the basic de finition of Simplification: Altering valve design if possible so that the valve position is obvious at a glance. See Chapte r 11 for the example of a ball valve handle that was installed us ing the incorrect convention of the valve handle being out-of- line with the piping and flow direction when the valve is open. Use of different flange sizes or characteristics so that valves cannot be installed backwards, particularly check valves. Replacing slip blinds with spectacl e flanges so that the status of the blind is always obvious at a glance. Replacing FRP (fiber reinforced plastic) tanks and piping with translucent materials so that fl ow and level are visible without the aid of flow or level instrumentation. If non-translucent FRP materials were initially used, this will likely have to wait until the components require replacement due to normal wear, or at the first available planned outage. Replacement of flexible hoses wi th fixed piping where possible. Piping with flexible joints that can accommodate equipment vibration and the use of articulated arms for loading and unloading are examples. Using quick-disconnect couplin gs only where absolutely necessary and replace unnecessary couplings with bolted joints that require a controlled work environment and the issue of a safe work permit to disassemble. These and other examples of Simplification are presented by Kletz and Amyotte (Ref 8.53 Kletz 2010).

5.1 Senior Leader Element Grouping |161 understand these warning signs and opportunities. Establishing a strong process safety culture will help this happen by fostering mutual trust and by ensuring open and frank communication . When senior leaders visit the workplace, they should engage with workers about process safety, put them at ease, and encourage them to speak freely. Then, following the culture principle Understand and Act on Hazards/Risks , the input should be acted upon. Just as im portantly, senior leaders should insist that the other leaders in their organization do the same. This could take the form of ad hoc discussions on the plant floor, or form al workforce involvement m eetings. Such interactions need not focus exclusively on process safety. The principle of workforce involvem ent can also help identify warning signs and improvement opportunities related to quality, productivity, occupational safety, etc. Workforce involvement also has the potential to improve labor relations, easing future negotiations. Workforce Involvement should ensure that the em ployees closest to the process hazards realize protecting their safety and welfare is the primary goal of the PSMS and their input is not only desired, but is im perative to the PSM S being effective. Workforce Involvement and PS Culture work hand-in-glove to build cooperation and trust am ong all workgroups. Organizations that fail to achieve workforce involvem ent stand to lose m ore than first-hand knowledge of warning signs and improvements. Without workforce involvem ent, prime opport- unities to build trust and open com munication channels are lost. This can lead workers to believe that process safety is som eone else’s job, underm ining the imperative for process safety . And as seen in the Columbia case history (Section 2.4), workers may decide not to report an actual serious situation because they believe their report will be ignored.

Piping and Instrumentation Diagram Development 102 6.14.2 Specialty I tems A plant is a combination of three hardware: equipment, instruments, and pipes. During the design of plant, the designers have an opportunity to design the equipment based on their process requirements. The instruments are selected based on the control requirement of the plant by instrument engineers. However, for pipe and pipe appurtenances, there is no such freedom. The pro-cess engineer must use whatever size of pipe or other piping items desired for the plant and put it on the P&ID. The Piping group in each engineering company is responsible for purchasing piping and piping items, and they generally provide a list of all the acceptable standard types and sizes of piping and piping items that can be used in each project. This list is part of a document called a Piping Material Specification. Basically the Piping group tells process engineers that only items on the Piping Material Specification list must be used on a cer - tain project. However, there are some cases that a pro-cess engineer or instrument engineer needs a piece of piping item, and it is not in piping spec. In such cases, the first option is to try to replace the desired item with another item in the piping spec. If that does not work, the Piping group may agree to this as an exception. These exceptions are considered SP items and definitely cannot be a long list. Therefore, it could be every non‐equip-ment item that has not been included in the piping spec. The Piping group does not care for SP items because they already know the complete specification of each item in the piping spec, and it is easy for them to buy the items from the market. However, for each SP item, the person who asked for a specific SP item (more than likely the process engineer) needs to prepare a data sheet for the item and submit it to the Piping group. On the P&ID, a SP item is shown as a little box beside the item, and the acronym SP with a number, which is the a tag number (Figure 6.86). To decide if an element is an SP items, the piping spec of the project should be consulted. That also means that there is no universal rule that a non‐slam check valve is a SP item. However, an experienced engineer may know that, generally speaking, where there is a check valve in the project’s piping specs, it is a conventional check valve, and if non‐slam check valves are required, they are most likely SP items. When a process engineer includes an element as an SP item, he/she needs to be aware that he/she does not have flexibility to choose the sizes needed as SP items are generally off‐the‐shelf items. Figure 6.87 shows few examples. In this figure, a non‐slam check valve, an injection quill for a chemical injection system, and a flexible joint are all SP items. In a P&ID, the designer may decide to put definition of the SP item beside the symbol. This is acceptable if the definition is one to three words in length. Otherwise it is better to not put the SP item definition on a P&ID and keep it only on the SP item data sheet. Sometimes, there is a dispute between the Mechanical group and the Piping group if one specific element is SP item or actually equipment. This is a valid dispute because if the element is a piece of equipment, it should be tagged and bought by the Mechanical group, and if it is an SP item, it should be tagged and bought by the Piping group. As a rule of thumb, SP items should be small items on the pipe, meaning elements with less than 0.5 m3 in volume without any utility connection and without any complexity in its structure. If an element is small but it has a complicated structure, it is better to classify it as equipment rather than an SP item. 6.14.2.1 Flange‐Insulating Gasket Here, insulating refers to electrical insulation and not heat insulation. Galvanic corrosion is arguably the most common type of corrosion in process plants. Galvanic or electrochemical corrosion happens when two dis - similar metals are put in contact with each other. However, there are some cases that mating flanges are not similar from material point of view. These situations happen when there is a spec break in the pipe. The spec break is almost always on a flange. The flange could be for pipe, valve, or any other pipe appurtenance. One side of the flange is from one metallic material, and the other side is from another metallic material. In such cases, an electrically nonconductive gasket should be placed between two mating faces of the flange to prevent elec - trochemical corrosion. Flange‐insulating gaskets can be tagged as SP items. SPSP Figure 6.86 Specialty it em tags. SP SP SP Non-slam Injection quill Flex. joint Figure 6.87 Specialty it em examples.

130 PROCESS SAFETY IN UPSTREAM OIL & GAS ●Compliance with Standards – These include industry standards, regulations and RAGAGEP. There are many design standards based on years of learnings that are available to guide a project design. ●Hazard Identification and Risk Analysis – The project should identify hazards and employ inheren tly safer design approaches as the first step in managing risk across the asset life cycle. Risk analysis aids in decisions on risk reducing measures. ●Asset Integrity and Reliability – The design greatly affects this element. Selection of materials appropriate for the design conditions and designing equipment to facilitate maintenance both serve to maintain integrity. ●Workforce Involvement – Operations personnel should be included in the design team. This helps to ensure that operational experience is included in design decisions, not ju st for hazard identification, but to aid operability and maintainability issues, and to contribute to inherently safer design. ●Management of Change – Changes to design should be reviewed, especially after hazard identification, in case these impact safety or environmental barriers assumed present in the hazard identification study. During both greenfield and brownfield construction, and in initial start-up, additional RBPS elements become important including the following. ●Contractor Management – This addresses construction safety and SIMOPS control. Many people may work on di fferent activities in a small space, and managing this work is key to safety including verifying competencies, required certifications and that personnel understand the interface requirements. ●Operational Readiness – This activity verifies that the facility is complete and ready for operation. Are all the safe ty features agreed in design now implemented in the constructed facility and is it safe to start-up? ●Safe Work Practices – This addresses safety during construction and ongoing operations. Safe Work Practices are essential to support the safe working on the project during construction, commissioning, startup, and into production. ●Operating Procedures – Procedures should be developed for startup and ongoing operations. Before startup, th ese procedures need to be written, available and understood. ●Training and Performance Assurance – This element aims to ensure that personnel are well trained for startup and for operations. This training should ensure understanding of Safe Work Practices and Operating Procedures . 7.1.2 Project Life Cycle Terminology There are multiple terminologies used by different companies for life cycle stages; this chapter uses the terminology from the CCPS Guidelines for Integrating Process Safety into Engineering Projects . The main stages are listed in Table 7-1. The life

EVIDEN CE IDEN TIFICATION , COLLECTION & M ANAGEM ENT 141 For these reasons, the investigation environment can be challenging, and therefore a systemat ic approach is necessary for the successful investigation of major process incidents. 8.1.3 Priorities for Managing an Incident Investigation Team The incident investigation team has the responsibility for determining the root causes of the occurrence and therefore needs access to the incident scene and other sources of information as quickly as possible. The plant and site management have the primary responsibility for preserving data and site evidence and for preventing the destruc tion of any eviden ce. Nevertheless, the investigation team should provide information to management on the evidence to preserve, method of pr eservation, resources needed to collect and test evidence, and other ev idence related activities. One noteworthy example is the preser vation of time-sensitive data from DCS and PLC systems where uncompressed data may be held in a circular buffer that is being continuously ov erwritten and batter y backups have a limited life. However, there are other priorities, espe cially in the early stages of the investigation. (Ferry, 1988). It is extremely important to note that the investigation team’s responsibilities are significantly different from those of an emergency response team or search and rescue team. Some key activities at the incident scene and the responsible parties are listed in Table 8.1. The investigation team may not be on site until several of the issues listed below are resolved.

DETERM INING ROOT CAUSES 247 Unlike the procedure followed in developing logic trees, the investigation team does not construct th e tree. Rather they apply each causal factor to each branch of the predefined tree in turn, and those bran ches that are not relevant to the incident are eliminated. This prescriptive approach offers consistency and repeatability by presenting different investigators with the same standard set of possible root causes for each incident. The consistency offered by predefin ed trees with stan dard categories and subcategories of root causes also facilitates statistical trend analysis. This allows an organization to more eas ily collect and analyze data from the investigation of incidents and near misse s over a period of time to determine any trends not apparent from single incidents. Some organizations deliberately structure th e root cause categori es and subcategories along the lines of their management system i n o r d e r t o f o c u s o n c o m m o n system issues. While the use of predefined trees do es not directly challenge the investigation team to think laterally of other possible causes, many predefined trees present a wide range of causes, some of which the team may not have otherwise considered. It is therefore possible that the incident could involve a novel root cause that was not previously experienced by those who developed the predef ined tree. The addition of a final test based on another tool, such as brainstorming, can overcome this potential weakness 10.8.1 Predefined Tree Methodology Although there are differences between various predefined trees, the basic method to perform a root cause analysis using the trees is similar, whichever tree is used. The following basic steps apply: 1. First, it is necessary to identify the multiple causal factors of the incident. The procedures in Chapte r 8 (Section 8.4-Timelines and Sequence Diagrams) may be used to identify the causal factors from a timeline or sequence diagram. 2. The first causal factor is then analyzed, starting at the top of the pre- defined tree and working down the branches as far as the facts permit. If the category of a particular branch appears to be an appropriate cause of the inciden t, the branch is followed to successively lower levels until a subcategory is identified as an

108 PROCESS SAFETY IN UPSTREAM OIL & GAS An example of a relatively simpler and lower congestion offshore platform is given in Figure 6-1. This platform in 237 ft (72 m) of water has two main decks with a helideck on the left and a flare tower to the right. The upper deck houses living quarters, offices, a galley and various storage and production equipment; and the lower deck houses various other rooms and equipment including the Motor Control Center. Most equipment is outside, not in enclosed modules. An example of a more complex and congested offshore production facility was Piper Alpha, which was located in the North Sea in 474 ft (144 m) of water. The module layout is shown in Figure 6-2. This shows most equipment to be in modules with tight spacing in both the horizontal and vertical directions. In Piper Alpha’s original design, the accommodation was sited as far as possible from major hazard areas, such as the wellhead and separation modules. However, a regulatory change was implemented by the Government to require gas to be recovered rather than flared. A gas compression area was added as a modification and this was close to the accommodation. This module was a major hazard due to large amounts of high- pressure gas being processed. Offshore production usually starts with flow from a reservoir as a mixture of oil, gas, and water. The production facility may be directly over a single wellhead or may be centralized to process the feed from multiple wells and may use a system of gathering pipelines. Subsea systems are growing in complexity from simple mixing of streams to also include some processing. Topside separation facilities are similar to those shown in Figure 5-2 for onshore as the raw feed is similar. Typically, 1st and 2nd stage separators are used at high and low pressure to degas the oil and separate the oil and water. Gas is treated as necessary (e.g., removal of H 2S, dehydration) and compressed for export, used for power generation on board, or Figure 6-1. Example of shallow water facility Gulf of Mexico BSEE Panel Investigation into West Delta Block incident in 2014

83 alternatives have been identified . REACH is discussed further under Regulatory Initiatives in Chapter 12. 4.8 REFERENCES 4.1 Canadian Center for Occupational Health and Safety (CCOHS), Substitution of Chemicals – Considerations for Selection, www.ccohs.ca/oshanswers/chem icals/substitution.html. 4.2 Catanach, J.S., and Hampton, S.W., Solvent and surfactant influence on flash points of pesticide formulations. ASTM Spec. Tech. Publ. 11, 149-57, 1992. 4.3 Dale, S.E., Cost-effective design considerations for safer chemical plants. In J.L. Woodward (ed.). Pr oceedings of the International Symposium on Preventing Major Chem ical Accidents, February 3-5, 1987, Washington, D. C. (pp. 3.79-3.99). New York: American Institute of Chemical Engineers, 1987. 4.4 Davis, G.A., Kincaid, L., Menk e, D. , Griffith, B., Jones, S., Brown, K., and Goergen, M., The Product Side of Pollution Prevention: Evaluating the Potential for Safe Substitutes . Cincinnati, Ohio: Risk Reduction Engineering Laborato ry, Office of Research and Development, U. S. Environmental Protection Agency. 1994. 4.5 DeSimone, J.M., Maury, E.E., Guan, Z., Combes, J.R., Menceloglu, Y.Z., Clark, M.R., et al., Homogeneous and heterogeneous polymerizations in environmentally-responsible carbon dioxide. Preprints of Papers Presented at the 208th AC S National Meeting, August 21-25, 1994, Washington, DC (pp. 212-214). Center for Great Lakes Studies, University of Wisconsin-Milwauke e, Milwaukee, WI: Division of Environmental Chemistry, Amer ican Chemical Society, 1994. 4.6 Edwards, V. and Chosnek, J., Making Your Existing Plant Inherently Safer, Chemical Engineering Progress, January 2012. 4.7 Flam, F., Laser chemistry: The light choice . Science 266, 215- 217, 14 October 1994. 4.8 Govardhan, C.P., and Margolin, A.L. Extremozymes for industry: From nature and by design. Chemistry & Industry, 689-93, 14 October 1994.

Overview of the PHA Revalidation Process 11 Thus, like LOPA, it can be used to es timate the frequency of specific loss scenarios. Additional guidance on the app lication of bow tie barrier analysis can be found in the CCPS concept book Bow Ties in Risk Management: A Concept Book for Process Safety [19]. Quantitative Risk Analyses (QRA). QRA is the collective term for a variety of detailed quantitative analysis tools used in risk calculations. In a QRA, the calculated consequences and the calculat ed frequency of each scenario may be used (separately or combined) to determin e individual scenario risks. Individual scenario risks can be further combined to provide an overall picture of cumulative site risks. For example, QRA consequence analysis tools can account for particular release characteristics, such as the density and state (solid, liquid, gas) of the released material, its direct ion and momentum, its chemical reactions and interaction with humidity in the atmosphere, its interaction with surrounding structures or terrain, the bi ologic effects on potentially exposed individuals, and so forth. Other consequence analysis tools can calculate detailed fire or explosion effects. Simila rly, QRA frequency analysis tools, such as Fault Tree Analysis (FTA), can predic t the frequency of system failures from data on the failure rates of specific components. Event Tree Analysis (ETA) and Human Reliability Analysis (HRA) can calculate the likelihood (probability or frequency) of particular lo ss scenarios. All of these methods require construction of a valid logic model and its reduction to sets of basic causes for which data is available. Failure data is available from sources such as the CCPS book Guidelines for Process Equipment Reliability Data, with Data Tables [20]. These QRA frequency analysis techniques are not encumbered by the conservative assumptions embedded in a simplified QRA technique such as LOPA, so they typically produce better estimates of event frequencies. However, they require significantly more analytical and computational effort than the simplified techniques. QRA tools are often used in facility siting studies, in risk analyses of occupied structures, or in th e design of Safety Instrumented Systems (SISs). Additional guidance on the application of QRA can be found in the CCPS book Guidelines for Chemical Proce ss Quantitative Risk Analysis [21]. 1.4 PHA REVALIDATION OBJECTIVES The primary objective of a PHA revalidat ion is to produce a revised PHA that adequately identifies and evaluates risk controls for the hazards of the process, as they are currently understood .

107 6 Application of Process Safety to Offshore Production 6.1 BACKGROUND Offshore oil production, particularly in deepwater, has become a major part of upstream operations. Offshore is often divided into shallow water and deepwater production. Shallow water is usually cons idered to be anything under 1,000 ft (305 m). Deepwater is anything greater than this and the term “ultradeep water” for fields in greater than 5,000 ft (1,524 m) water depth is also used. In the Gulf of Mexico, many shallow water fields are sm all and are declining in production. Shallow water structures in the Gulf ar e simple steel jackets. When similar designs were first employed in harsher environments of wind and wave, such as the North Sea, some initial failures (e.g., Se a Gem 1965) showed that greater strength was required. This either strengthened the steel structure or introduced novel designs such as gravity based concrete structures. Some designs improved process safety by separating accommodation from processing on bridge linked platforms. As well construction technology advanced and large fields were discovered in deepwater; floating producti on facilities became necessary as fixed structures are impractical. Deepwater floating facilities have several possible designs, each with advantages and disadvantages (see Chapter 2 for examples). All the designs enable oil and gas production and have separation facilities. Export by pipeline is typical while FPSOs allow for storing oil in the hull for subsequent transfer to a shuttle tanker. A more recent development for gas fields is FLNG – Floating Liquefied Natural Gas facility, which includes a liquefaction plant as well as the usual treatment facilities. It also stores its LNG product onboard and transfers this periodically to LNG carriers for export. FLNG facilities add extra complexity to offshore facilities including cryogenic pr ocessing and greater inventories of hydrocarbons. In principle, the process safety risks in deepwater facilities are often higher than for shallow water designs as the economics of deepwater mean the facilities are bigger and more costly to develop and the number of crew on board is greater. However, all facilities are subject to loss of containment incidents that can give rise to serious consequences. A number of historical incidents are discussed later in this section. One feature of offshore facilities important for process safety is that the workforce usually lives on board and thus off-shift personnel are potentially close to hazards. Onshore, off-shift personnel are at home or in accommodation modules located more remotely. On-shift personne l may be close to hazards onshore or offshore, but offshore they may be sited above or adjacent to process hazards, whereas this is less common onshore. Process Safety in Upstream Oil and Gas © 2021 the American Institute of Chemical Engineers

INVESTIGATION M ETHODOLOGIES 29 3.1 HISTORY OF INVESTIGATION M ETHODOLOGIES AND TOOLS Investigation methodologies for process safety incidents have evolved over time, becoming more systematic, objective and scientific. It is relevant to review the history of investigatio n methodologies to learn from the weaknesses of historical methods and appreciate the approaches in modern methods. 3.1.1 One-on-One Interview The historical approach to investigating incidents was an informal, one-on- one interview, typically between the pe rson involved in th e incident and his or her immediate supervisor. This approa ch has generally been less effective than structured investigation methodol ogies for process safety incidents, especially complex incidents resulting in, or having the potential to result in, serious or catastrophic consequences. Informal one-on-one interviews are still often used as an approach for investigating low severity incidents, including minor occupational injuries. The focus of informal, one-on-one investigations has often been limited to determining the immediate remedies that would preven t an exact repeat of the incident circumstances. For example, a common finding may have been that an operator failed to follow an established procedure . Based on that finding, the investigator might have proceeded to evaluate how best to motivate this specific operator to follow the proc edure as a recommendation to prevent recurrence. This informal type of investigation required little time or training, but the weakness of this approach for significant process safety incidents is that it do es not determine the fund amental reason for the occurrence of the incident in the first pl ace. If the fundamental reason (root cause) is not identified, then measur es cannot be taken to address this fundamental reason, and the incident, or a very similar one, may recur. 3.1.2 Brainstorming Brainstorming is essentially an unstructured tool, but it can provide more perspective and experience than one-on-one investigations. Brainstorming brings together a group of people fr om diverse backgrounds to discuss the incident and intuitiv ely determine the causes of the incident. The group will typically understand the sequ ence of occurrences that led up to the incident through a timeline or sequence diagram. The group may also have identified causal factors, and typically focuses on establishing barriers to reduce the risk (probability or consequences) of recurrence.

196 INVESTIGATING PROCESS SAFETY INCIDENTS or unsafe conditions of the inciden t, as an intermediate step before proceeding to determining the root causes. 9.6.1.3 Identifying Causal Factors The simplest technique for identifyin g causal factors involves reviewing each event or condition on the timeline. The investigator repeatedly asks the following question: W ould the result have been significantly different if the event or condition had not existed at the time of the incident? If the answer is YES, that is, the incident would have been prevented or mitigated by the eliminatio n of a negative event or undesirable condition, then the fact is a causal factor. Generally, process safety incidents involve multiple causal factors. This te chnique is equivalent to step #15 in Figure 9.3. Once identified, the causal factors be come the candidates to undergo root cause analysis. The investigator may streamline this technique by focusing upon each unplanned, unintended, and/or adverse fact (negative event or undesirable condition) on the timeline. It is also important to recognize those items that are still speculative and based on an assumption, as these should be tested later to verify if they are accurate facts. It is critically important that the wo rding or the phrasing of each causal factor accurately and clearly describes the factor. Teams will struggle with cause analysis if the causal factor is not crystal clear to all. In the case of an incident arising from work on a pump th at has not been adequately isolated from energy sources, an investigation team may say one causal factor is “no lockout/tagout (LO/TO)”. However, this short statement can be interpreted in a number of ways, depending upon individual team members’ views of the evidence and personal biases. For example, “no lockout/tagout” can mean: • No procedures for LO/TO exist • Procedures exist but the employees involved had no knowledge of them • An attempt was made to perform LO/TO, but it was performed incorrectly • LO/TO was performed on the wron g equipment or missed on one item • No effort was made to perform LO/TO.

12 Identifying learning requirements 12.1 Learning objectives of this Chapter Phase 3 of the Competency Management focuses on assessment of gaps in competency, and limitations in current training. By the end of this Chapter, the reader should be able to: • Understand a Competency Gap Analysis, and a Training Needs Analysis. • Identify and describe learning objectives. Competency Gap Analysis and Training Needs Analysis are important stages in identifying learning objectives, and training needs. Normally, they would be carried out together, as information from the Comp etency Gap Analysis feeds directly into the Training Needs Analysis. 12.2 Competency gap analysis Competency Gap Analysis is discussed in Chapter 11 and makes use of defined competency standards matrices. Competen cy Gap Analysis assesses individual competency against those outlined in the standards. It answers the following questions: • Which competencies do people currently possess? • What is the level of their competency (e.g., awareness, basic application, skilled application, mastery)? • How can the competency gap analysis bridge the gap between the competency people possess now, and the competency needed? Competency Gap Analysis needs to recognize any prior experience and qualifications and consider how they may satisfy parts of competency requirements. Competency Gap Analysis should consider whether any types of experience or qualifications are needed in satisfying a performance standard. Competency Gap Analysis identifies what gaps exist between employees’ current competency and the competen cy required to fulfil performance standards. Human Factors Handbook For Process Plant Operations: Improving Process Safety and System Performance CCPS. © 2022 CCPS. Published 2022 The American Institute of Chemical Engineers.

5 Facility Shutdowns 5.1 Introduction This chapter discusses the transient operating modes associated with a facility shutdown: the shut-down fo r a facility shutdown (mode Type 5, Table 1.1) and the start-up af terward (Type 6, Table 1.1). The considerations and types of larger projects requiri ng a process unit or facility project-related shutdown are described in Section 5.3. This chapter then provides a discussion on preparing for a facility project- related shutdown (Section 5.4), starting up after a shutdown (Section 5.5), and provides an incident that occurred during these transient operating modes with lessons learned . This chapter concludes with a discussion on the applicable RBPS elements for the facility shut-downs and start-ups afterwards (Section 5.7). 5.2 The facility shutdown Larger capital projects that invol ve shutting down the equipment associated with an entire process uni t, an entire facility, or even multiple, interconnected production facilities are designated as a facility shutdown in this guideline. These projects are complex, involve many different groups, and—distingui shing them from the smaller, planned projects discus sed in Chapter 4—have an extended timeline for the work, covering weeks or even months of both the preparation time beforehand and the execution ti me during the project. Often the longer, larger projects that stop op erations for the long period are called a turnaround or an outage, as well. The facility project timeline was shown in Figure 4.3. These proj ects are intense and often stressful times due to their scope (or scopes), and the complex demands can Guidelines for Process Safety During the Transient Operating Mode: Managing Risks during Process Start-ups and Shut-downs . By CCPS. © 2021 the American Institute of Chemical Engineers

EQUIPMENT FAILURE 235 CSB 2015, “Final Investigation Report Chevron Richmond Refinery Pipe Rupture and Fire”, Report No. 2012-03-I-CA, U.S. Chemical Safety and Hazard Investigation Board, Washington, D.C. CSB 2019, “Fire and Explosion at Philadelphia En ergy Solutions Refinery Hydrofluoric Acid Alkylation Unit, Factual Update”, No. 2019- 06-I-PA. U.S. Chemical Safety and Hazard Investigation Board, Washington, D.C. Chu, “Improved Heat Transfer Predictions for Air-Cooled Heat Exchangers ”, Chemical Engineering Progress , November, 2005. Donaldson-Torit, www.donaldson.com/et-us Drogaris, G. Major Accident Reporting System: Lessons Learned from Accidents Notified, Elsevier, Amsterdam, 1993. EI, “Guidelines for the management of flex ible hose assemblies”, Energy Institute, https://publishing.energyinst.org/topics/asset -integrity/guidelines-for-the-management-of- flexible-hose-assemblies. Ender, Christophe and Laird, Dana, “Minimize the Risk of Fire During Column Maintenance”, Chemical Engineering Progress , p. 54-56, September 2003. EPA 2016, “Operating and Maintaining Underg round Storage Tank Systems”, EPA 510-K-16- 001, February, Environmental Protection Agen cy, Washington, D.C. Fike, www.fike.com Garland, R. Wayne, “Root Cause Analysis of Dust Collector Deflagration Incident”, Process Safety Progress , Vol. 29, No. 4, December 2010. Haslego, Haslego and Polley, “Designing Plate-and Frame Heat Exchangers”, Chemical Engineering Progress , September 2002. HEI, Heat Exchange Institute, www.heatexchange.org HSE 1994 The explosion and fires at the Texaco Refinery, Milford Haven. 24th July, Health and Safety Executive, https://www.hse.gov .uk/comah/sragtech/casetexaco94.htm. HSE 2009a, “Buncefield Explosion Mechanism Phase 1, Vols. 1 and 2” , U.K. Health and Safety Executive. HSE 2011, “Buncefield: Why did it happen?”, U.K. Health and Safety Executive. HSE 2017, “Review of vapour cloud explosion inci dents”, U.K. Health and Safety Executive. IEC 61511, “Standard for Safety Instrumented Systems”, International Electrotechnical Commission, Geneva, Switzerland, https://www.iec.ch/. IOGP 2019, “Risk Assessment Data Directory - Report 434-01 Process release frequencies”, International Association of Oil & Gas Producers, https://www.iogp.org/bookstore/product/434-00- risk-assessment-data-directory-overview/. Kelley, J. Howard, “Understand the F undamentals of Centrifugal Pumps”, Chemical Engineering Progress , October 2010.

330 | Appendix E Process Safety Culture Case Histories The engineer shook his head and explained patiently that it was necessary to inert the reactor, because otherwise the flam mable atmosphere could ignite, especially because the solvent was not being fed through a dip-pipe. “Yeah, I’ve heard of that,” the operator said, “but take it from me, it is a waste of time to inert the reactor because 9 times out of 10 it does not explode.” “Uh, let’s have that coffee and talk about it,” the engineer said. They went into the breakroom , took their coffee, and sat across the table from each other with the box of donuts between them. The engineer reached for a donut. “The thing is,” he said, “if it doesn’t explode 9 tim es out of 10, then it does explode that other one time. I don’t know about you, but my goal is this.” He held the donut up in front of him , showing the operator the big sweet 0. The operator grabbed the donut and stuffed it in his m outh. After washing down that donut with a gulp of coffee, he put 2 m ore donuts in his pocket, left the breakroom and started inerting the reactor. The operator appeared to understand the hazard and possibly even the risk. If so, did he need to have it explained to him again? Or did he need som ething else? Did the operator frequently skip other safety steps in procedures? Was this normal behavior within the plant? Should the engineer have questioned the Plant’s imperative for safety ? How did the engineer convince the operator? Was it through a logical argument? Establishing mutual trust ? Or was the operator testing the engineer’s leadership ? Establish an Imperative for Safety, Provide Strong Leadership, Maintain a Sense of Vulnerability, Understand and Act Upon Hazards/Risks, Defer to Expertise, Combat the Normalization of Deviance.

2.2 Resources for Learning | 19 • European Commission Major Accident Reporting System (European Commission 2020) • Infosis ZEMA (ZEMA 2020) • Lessons Learned Database (IChemE 2020) • Marsh 100 Largest Losses in the Hydrocarbon Industry (Marsh 2018). 2.2.3 Publications Whether your bookshelf is on the wall, on a hard drive, in the cloud, or in a virtual or physical corporate library, you probably have a few books describing how to investigate incidents and extract lessons you can learn from them. Table 2.4 describes the most popular of these books. Additionally, most books about process safety use case histories featuring actual incidents to highlight key concepts. Table 2.4 Books About Incidents and Incident Investigation Author and/or Publisher Name of book or series CCPS/AIChE and Wiley Guidelines for Investigating Process Safety Incidents, 3rd ed. Incidents That Define Process Safety More Incidents That Define Process Safety Earl Boebert and James Blossom, Harvard University Press Deepwater Horizon: A Systems Analysis of the Macondo Disaster Andrew Hopkins, CCH Australia Failure to Learn: The BP Texas City Refinery Disaster Lessons from Longford: The ESSO Gas Plant Explosion Trevor Kletz, Butterworth Heinemann/IChemE What Went Wrong? series Frank Lees, Elsevier Loss Prevention in the Process Industries, eth ed. Roy Sanders, Butterworth Heinemann Chemical Process Safety Learning from Case Histories, 3rd ed.

EQUIPMENT FAILURE 237 Patnaik, T., “Solid-Liquid Separation: A Guide to Centrifuge Collection”, Chemical Engineering Progress , July 2012. PSLP a, Jarvis, H.C. “Butadiene Explosion at Texas City-2”, Plant Safety & Loss Prevention , Vol. 5, 1971. PSLP b, “Butadiene Explosion at Texas City-1”, Plant Safety & Loss Prevention , Vol. 5, 1971. PSLP c, Keister, R.G., et al. “But adiene Explosion at Texas City-3”, Plant Safety & Loss Prevention , Vol. 5, 1971. Ramzan, Naveeed, et al, “Root Cause Analysis of Primary Reformer Catastrophic Failure: A Case Study”, Process Safety Progress , Vol. 30, No. 1, March 2011. Sherman, R.E., “Carbon-Initiated Effl uent Tank Overpressure Incident”, Process Safety Progress , Vol. 15, No. 3, Fall 1996. Shutterstock, Royalty-fr ee stock photo ID: 1340068283 Sulzer Chemtech Ltd., www.sulzer.com/en/ TEMA, Tubular Exchanger Manufacturers Association, http://kbcdco.tema.org/ Urban, P.G., Bretherick's Handbook of Reactive Chemical Hazards 7th Edition , Academic Press, New York, NYISBN:978-0-12-372563-9, 2006.

118 INVESTIGATING PROCESS SAFETY INCIDENTS way. Perhaps the supervisors and ma nagement were aware of these types of issues prior to the ev ent and could have done so mething to correct them. If the investigation reveals that staff routinely fails to follow procedures, this may be indicative of more fundamen tal cultural issues that require addressing by management. For example, an operator may skip a pre-operational check of a system because he believes the check will not discover any problems and takes valuable time that could be used to produce product. In other words, the operator believes the check is a waste of time. Perhaps the operator had skipped the preoperational check many times, and it had never caused any problems. His supervisor may have known he normally did not perform the preoperational checks but had said no thing because it resulted in increased production. Skipping the checks was not a malicious act or act of sabotage or even an act of negligence; it was an ‘accepted’ practice. However, this time when the operator skipped the ch eck, the system failed and a release of process chemicals occurred. Will the operator tell the incident investigation team that he skipp ed the preoperation al check? What motivation would there be? What potential punishme nts are there? Unless the operator believes that it is in his own best interest to divulge the information, he probably will not. Unless boisterous play, negligence, or sabotage are clearly involved, individu als should not be punished for the information revealed during incident investigation interviews. If witnesses are aware of this, they are more likely to openly share the information they have. The investigation team’s responsibility is to gather facts and draw conclusions. Punishment is not part of the investigation process. This philosophy should be emph asized as part of the tr aining requirements, as outlined in Chapter 4. Any disciplinary action arising from an investigation is part of a separate process involving Human Resources personnel/policies. Cultural issues, including company cu lture and country/regional culture, should also be considered. This may include factors such as a tendency to agree with whatever is said by someon e perceived to be in authority or a more senior position, and an unwillingness to divulge information that could reflect poorly on a co-worker. 7.3.2 Collecting Information from W itnesses The accuracy and extent of witne ss information is highly dependent on the performance of the interviewer. The interviewer’s ability to establish rapport and create an atmosphere of trust affects the quality and quantity of information disclosed.

Ancillary Systems and Additional Considerations 389 with the pipe it is installed in. If the pipe needs winteri- zation provisions, the inline instrument also needs it. For non‐inline (offline) the requirement depends on whether it is fluid‐in type or not. The fluid‐in instruments like Bourdon tube in pressure gauges may need winterization arrangement. But non‐fluid‐in types definitely don’t need any winterization arrangement. Figure 18.8 shows some examples of winterization of instruments.18.3.3 Deciding on the Ext ent of Insulation From a purely theoretical viewpoint, when it is decided that an “item” is to be insulated, the “whole” item should be insulated. However in reality we don’t always insulate whole the item, unless it is fully justifiable from an economical viewpoint. The full insulation may also make the normal operation of the system of interest problematic. There are cases where the designer needs to decide on the extent of insulation. This means he needs to decide if all the pipes/equipment of interest should be insulated or only a portion of them. The extent of insulation is generally mentioned on P&IDs; on the main body of the P&ID or in the note area. For example, one general question is whether a full tank, including body and roof, should be heat insulated if the tank needs to be insulated or not. All the tank could be insulated if it is decided so, but the cost of insulation could be saved by insulating the body of the tank. As the main reason for this insulation is to keep the liquid content “warm, ” possibly the roof of the tank doesn’t need to be insulated. Even on the body of the tank, only the portion that is in contact with liquid could be decided to be insulated. The additional reason for not insulating the roof is that the heat transfer coefficient of gas/vapors is much less than for liquids, and the heat transfer from the roof is already low. A  company may decide that: “if a tank needs to be heat insulated, only the body of the tank up to the high liquid level should be insulated. ” For pipes almost always all the body of the pipe could be insulated. The only exception could be large bore TW 00081TT 0008088TI 00080 PP *PT 00080FT 00080 I/S O/S ET N Note 3PI 00080FQI 00080 FI 00080 FX 00080 Figure 18.8 Win terization of some instruments.Table 18.8 P&ID presen tation of winterization for different items. Pipe Equipment Instrument W And/or in pipe tag 6/uni2032-HLS-AS-1003-64H-GT P–07111–2–B39W–50H–ET 222–163L–2/uni2033–IhST(w)(1–1/2/uni2033)Not common unless for small equipment E G ST ETS

EVIDEN CE IDEN TIFICATION , COLLECTION & M ANAGEM ENT 171 timing marks. On the other side of the line, imprecise or approximate data is listed within the period in which it occurred. This is usually displayed as an event occurring someti me between two timing marks. Figure 8.7 is an example of a timelin e that uses a mixture of precise and imprecise data from the incident example discussed in detail in Appendix D. One additional benefit of this technique is that the imprecise approximate times can often be narr owed when compared to the precise data. For instance, the operator may realize that when he manually closed valve A, valve B had already been automatica lly closed. Therefore, the period within which he closed valve A is narrowed. Timelines do not have to end at the time of the occurrence or incident. Sometimes post occurrence data can be valuable. Often, it is important to understand how the emergen cy response actions affected the ultimate outcome of the occurrence. This type of data can be used to improve emergency response actions in the future. Also , changes made during emergency response to positions (valves, switches, debris positions, etc.) can be important to interpretation of the data. When timelines are combin ed with simulations, they become powerful tools, both in understanding the sequence of the events leading up to the incident and in the development of accurate recreations. This allows for a more thorough and comprehensive analysis.

70 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS Example Incident 3.13 – Styrene Runaway Reaction and Release, 2020 (cont.) The only temperature indication in the storage tank concerned was at the base, and the pressure safety valve from the tank emitted directly to the atmosphere and not via a flare system or other emission control system. Lessons learned in relation to abnormal situation management: Abnormal Situation Recognition: The temporary closure of a facility requires special consideration, particularly if the duration of the closure is unknown at the start. Organizational Roles: Responsibi lities and Work Processes: With a reduced workforce, revised procedures and responsibilities should have included extra checks on high-risk areas. These considerations should fall under a Management of Operational Change (MOOC) procedure. Procedures: Procedures should have been revised to address high-risk scenarios due to wo rk force and organizational changes. Process Monitoring and Control: Provided the risk had been identified, additional automatic control measures, including better temperature measurement could have been specified. 3.4.2.4 Turnaround/Shutdown/Decommissioning Turnaround, shutdown, and/or de commissioning of process plant equipment is a normal activity during the life cycle of a chemical process. Because of this both scheduled, preventive, and unplanned emergency maintenance of mechanical equipment has become a critical function of highly performing organizations. Re pair costs, production downtime, leak containment integrity, and regu latory inspections are all factors involved in developing and execut ing maintenance shutdown strategy. Advanced mechanical integrity pla nning methods include sophisticated tools such as Risk Based Inspec tions (RBI), Reliability Centered Maintenance (RCM), thickness testing, and monitoring. Asset Integrity engineers and maintenance speciali sts who have received specific

78 Human Factors Handbook Table 8-2: Checklist for layout of job aids 1. Make headings large to help people identify information. 2. Make headings stand out from the surrounding text to help people identify information. 3. Use spacing, images, and/or blank or white areas to reduce clutter, and to make it easier to identify and recognize information. 4. Use bullet points to make reading easier. 5. Number sections and task steps. 6. Communicate one action per task step. 7. Present different types of information in different formats and fonts (with supporting icons) so they are distinct and different (e.g., is it a warning or instruction, or is it background information?). 8. Clearly indicate who should perform th e task step if the task involves more than one person. 9. Present supporting information separa tely from the task instruction. 8.2.2 Examples of job aids Examples are shown in: • Figure 8-1: SOP. Good features include: o One instruction per numbered line; o Use of color and icons (see 8.6) to indicate the status of each point of information; o Use of space to reduce clutter; o Use of bullets to make requirements easier to read. • Figure 8-2: Grab Card, with some key features highlighted. • Figure 8-3: Decision Flow Chart. Good features include: o Short sentences; o Binary decision points; o Unambiguous decision criteria; o Color coded and large font for key text; o Fits on one page.

F.2 Culture Assessment Protocol |359 other measurem ents of hazard/risk in a way that avoids the need for recomm endations? This m ay be a cultural issue with a given HIRA/PHA team, but it m ay also represent a system ic problem in the organization at large. 133. Are HIRAs/PHAs performed “by the num bers” with little free thinking about what can go wrong? In recent years Layer of Protection Analysis (LOPA) has become a prevalent analytical m ethod for determ ining “how safe is safe enough,” particularly for high risk hazard scenarios identified during HIRAs/PHAs. LOPA has provided a consistent and repeatable m ethod for determ ining how m any independent protection layers are necessary to reduce the risk to a tolerable level. While LOPA is very useful analytical technique to analyze a hazard/risk, but it is not a very good technique for identifying a hazard/risk. Therefore, if the HIRA/PHA process at a facility relies solely on m anipulating num erical credits to reach an acceptable cell on a risk m atrix, without the cause and effect analysis and open discussion that occurs during HAZOP or What-If studies, the HIRA/PHA process may have lost an opportunity to identify additional important risks. 134. Are the recommendations emerging from the hazard/risk assessments meaningful? Do they address and reduce the risks identified? 135. Do risk reduction measures in HIRAs/PHAs over rely on hum an based safeguards such as operator training, the experience of personnel, or the existence of written operating procedures? 136. What are the bases for rejecting risk assessm ent recomm endations? Are the reasons for rejection predom inantly driven by cost considerations? 137. Are the risk assessment tools appropriate for the risks being assessed? Are the right tools to assess risks associated with low frequency – high consequence events? Are the tools deemed appropriate by recognized risk assessment professionals?

324 Inherent safety is assessed relative to a particular hazard, or perhaps a group of hazards, but essentially never relative to all hazards. A chemical process is a complex, in terconnected organism in which a change in one area of the system can impact the rest of the system, with effects cascading throughout the pr ocess. These interactions must be understood and evaluated. Similarly, the chemical industry can be viewed as an ecosystem with comp lex interactions, interconnections, and dependencies. Understanding these relationships is necessary in order to reach a well-balanced reso lution when technological options conflict. 13.2 EXAMPLES OF INHERE NT SAFETY CONFLICTS 13.2.1 Continuous vs. batch reactor The use of continuous, rather than ba tch reactors is a strategy that is often proposed for improving the inherent safety of a chemical process (Ref 13.4 CCPS 1993). This modification generally succeeds because a continuous reactor is usually much smaller, reducing the material and energy inventory of the process, in creasing heat transfer per unit of reaction mass, and improving mixing. However, batch reactors may also have safety advantages, and, under the right circumstances, may be judged to be inherently safer. Consider a simple reaction: The reaction is exothermic and pr oceeds virtually instantaneously to complete conversion in the presence of Catalyst C. The process hazard of concern is that the reaction mass becomes extremely unstable if Reactant B is overcharged, or Catalyst C is left out. The resulting buildup of unreacted Reactant B may result in a potentially explosive reaction if Reactant B exceeds a known critical concentration. Two processes are proposed for this reaction, a batch process (Figure 13.1) and a continuous process (Figure 13.2).

412 INVESTIGATING PROCESS SAFETY INCIDENTS Table G.2 Process Safety Inci dents & Severity Categories Severity Level (Note 4) Safety/ Human Health (Note 5) Fire or Explosion (including overpressure) Potential Chemical Impact (Note 3) Community/ Environment Impact (Note 5) NA Does not meet or exceed Level 4 threshold Does not meet or exceed Level 4 threshold Does not meet or exceed Level 4 threshold Does not meet or exceed Level 4 threshold 4 (1 point used in severity rate calculations for each of the attributes which apply to the incident) Injury requiring treatment beyond first aid to employee or contractors (or equivalent, Note 1) associated with a process safety incident (In USA, incidents meeting the definitions of an OSHA recordable injury) Resulting in $25,000 to $100,000 of direct cost Chemical released within secondary containment or contained within the unit - see Note 2A Short-term remediation to address acute environmental impact. No long term cost or company oversight Examples would include spill cleanup, soil and vegetation removal 3 (3 points used in severity rate calculations for each of the attributes which apply to the incident) Resulting in $100,000 to 1MM of direct cost Chemical release outside of containment but retained on company property OR Flammable release without potential for vapor cloud explosives -see Note 2B Minor off-site impact with precautionary shelter-in-place OR Environmental remediation required with cost less than $1MM. No other regulatory oversight required. OR Local media coverage

| 1 IN TRODUCTION 1.1 IM PORTAN CE OF PROCESS SAFETY CULTURE The 2014 FIFA World Cup sem ifinal between Germ any and Brazil featured two of the most technically proficient team s to contest a m atch. Within a half-hour, however, the difference between the two emerged, as Germany scored five goals on a shell-shocked B razil on the way to a 7-1 rout. The difference? Neymar da Silva Santos, the captain, leader, and culture-setter of the B razilian side, had suffered a fractured vertebra in the previous match, and could not even cheer his teammates on from the sidelines. With their culture-leader absent, B razil failed to execute their usually form idable gam e plan and suffered a catastrophic loss. Sim ilarly, process safety cannot succeed without culture leadership. Investigation of numerous incidents in major hazard operations has clearly revealed culture deficiencies. The data show that without a healthy process safety culture, even the most well-intentioned, well-designed process safety management system (PSMS) will be ineffective. For example, Union Carbide was known as a process safety technology leader in the early 1980s. However, weak culture at its Bhopal facility allowed many “Normalization of Deviance” failures leading to the December 3, 1984 tragedy. Simply stated, a strong, positive process safety culture enables the Essential Practices for Creating, Strengthening , and Sustaining Process Safety Culture , First Edition. CCPS . © 2018 AIChE . Published 2018 by John Wiley & Sons, Inc.PSMS = Process Safety M anagement System 1

Figure 15.6: Aqueous Ammonia: Limita tions of Magnitude of Deviations Table 15.3 gives the total vapor pressure of aqueous ammonia for conditions near the design point. The tank design pressure will be 405