Datasets:

hndc

/

AI4Hazard-small

like 0

GLOSSARY 421 identified elsewhere, typically using a scenario-based ha zard evaluation procedure such as a HAZOP Study. Lessons Learned —Applying knowledge gained fr om past incidents in current practices. Likelihood —A measure of the expected pr obability or frequency of occurrence of an event. This may be expressed as an event frequency (e.g., events per year), a probability of occurrence during a time interval (e.g., annual probability) or a conditional probability (e.g., probability of occurrence, given that a pr ecursor event has occurred). Limited impact incidents —Incidents deemed to be controllable with local resources and which have no lasting effects. Lockout/ Tagout — A s a f e w o r k p r a c t i c e i n which energy sources are positively blocked away from a segment of a process with a locking mechanism and visibly ta gged as such to help ensure worker safety during maintenance and some operations tasks. M a n a g e m e n t o f C h a n g e ( M O C ) —A management system to identify, review, and approve all modificati ons to equipment, procedures, raw materials, and processing conditions, other than replacement in kind, prior to implementation to help en sure that changes to processes are properly analyzed (for example, for potential adverse impacts), documented, and communicat ed to employees affected. M anagement System —A formally established set of activities designed to produce specific results in a consis tent manner on a sustainable basis. M edical Treatment —As defined by OSHA, trea tment (other than first aid) administered by a physician or by registered professional personnel under the standing orders of a physician. M ethodology —The use of a combination of two or more incident investigation tools to analyze the evidence and determine the root causes of the incident. M inor incidents —Incidents with minor actual or potential consequences, including minor injuri es and minor damage. M itigation —Lessening the risk of an acciden t event sequence by acting on the source in a preventive way by re ducing the likelihood of occurrence of the event, or in a protective wa y by reducing the magnitude of the event and/or the exposure of local persons or property.

13 Operational competency development 13.1 Learning objectives of this Chapter Competency development is composed of learning opportunities, including on- the-job learning, coaching and training. Th e aim is to increase people’s knowledge and skills levels. Competency development focuses on individual and team (e.g., crew, shift or team) competency. Similar to individual competency, team competency is also based on the level or type of risks associat ed with tasks, and the complexity of the work. Therefore, the process usually invol ves all individuals on the task, including any contractors. Individual gaps in competency are filled through learning. Collective or team competency gaps can be filled through adding team members or third-party service providers that po ssess the missing skills or knowledge. By the end of this chapter, the reader should be able to: • Understand the process of developing and maintaining employee competency. • Identify suitable learning approaches for particular type of human performance and types of competency. • Have greater awareness of effective learning opportunity design. 13.2 Good practice in learning 13.2.1 Facilitate Learning, develop and assess Individuals develop competency over time, through a combination of structured learning opportunities, including on-the-job training, apprenticeship, mentorship, assessment feedback, and formal qua lifications programs. As individual competency develops from basic applicatio n to advanced, or from awareness to mastery, so does their ability to work in dependently. The progress from “knowing how” to “being able to put knowledge in to practice” happens slowly, bit by bit. Learning is a gradual process, and it builds competency over time. Various approaches to support learning help to develop knowledge from the lower competency levels, e.g., from Pr imary Well Control and Secondary Well Control concepts including pressure cont rol and role of equipment to higher competency levels such as functions and maintenance of Blow Out Preventer, as shown in Figure 13-1. Human Factors Handbook For Process Plant Operations: Improving Process Safety and System Performance CCPS. © 2022 CCPS. Published 2022 The American Institute of Chemical Engineers.

136 | REAL Model Scenario: Leaking Hoses and Unexpected Impacts of Change hose hanging about a meter above grade could be subject to impact from foot and vehicular traffic. If the hose had been strung up above a roadway between two posts, however, it could have swayed in the wind, causing it to rub at the points of contact with the posts. Moreover, why was the Kawasaki plant using a hose when a hard-piped connection seemed more appropriate for an application like this? Either way, the Kawasaki plant did not seem to have completed a thorough hazard evaluation, running the hose without considering the possible impact of foot traffic, stresses, or rubbing on sharp corners. Festus and Swansea clearly had two different causes. In Festus, the vendor mislabeled the hose and the plant took the vendor’s word for it. In Swansea, the label was correct, but the driver ignored the label. João was already checking whether the correct hoses were being used, but could hoses be coming in mislabeled? He recalled that they’d changed vendors in the past few years. Could the new vendors’ hoses have different colors than the old ones, creating confusion? In the cases of Belle and Anonymous 1, problems clearly existed in both plants’ asset integrity management systems. Hoses used well beyond their service life represented an asset integrity failure. Juliana was already looking into their own testing and inspection frequencies. But the use of duct tape in Anonymous 1 worried him. Could the failures of some hoses have been delayed by temporary (and forbidden) patching? And only reported to management grouped in with later failures? The operators would never do this, would they? It would be easy enough to check. Just walk into the control room and ask innocently, “Can I borrow your duct tape, friends?” They had no reason to keep duct tape on hand. At first, Atchison and Anonymous 2 didn’t seem to apply to the current problem, although Antônio made a note to do a human factors evaluation of the storage tank unloading system. However, when he thought about it more deeply, connecting a hose to the wrong reactor port could cause unexpected chemicals to get into the hose. He made a note to follow up on that possibility. One of the other incidents in the Belle report, the oleum leak, caught Antônio’s attention. The piping was made of the correct material for transferring oleum and could certainly resist the moisture in ambient air and steam. However, trace oleum vapors had escaped through a flange and mixed with moisture in the air, forming sulfuric acid of less than 98% concentration. That composition was very corrosive to oleum piping! Could the failing hoses

280 | Appendix D High Reliability Organizations HROs also actively seek to learn and im prove. Frequent training is aimed at building deep technical com petence, enabling personnel to better recognize hazards and respond to unexpected problem s. Training also helps build trust and credibility am ong coworkers. Incident and near-m iss investigations are treated as an opportunity to learn, and learnings are openly shared across the organization. Procedures are updated based on learning acquired. HROs recognize that com m unications are vital, and use m ultiple channels to com municate safety critical inform ation to ensure it is delivered and received, especially in emergencies. For exam ple, nuclear powered aircraft carriers have twenty different com munication devices. HROs exhibit m indful leadership including engaging often with front line staff through site visits and active encouragement of bottom-up com m unication of bad news. They proactively conduct m anagem ent system audits, often in response to incidents that occur in other sim ilar industries. They also invest resources in safety m anagement and can balance profits with safety. Another characteristic of HROs is resilience, the ability to recover from errors. Despite their low incident rates, HROs are not error-free. Rather, they remain preoccupied with failure to better anticipate them and recover from errors when they occur. Most of the attributes discussed above should sound fam iliar to readers of this book. The m ain differences arise from the natures of the organizations considered to be HROs com pared to chem icals, oil, and gas. These differences m ay m ake som e aspects of safety culture easier to attain and others m ore difficult. For exam ple, the commercial terms of nuclear power facilities are heavily regulated, with strict controls on costs, rates, and profits. In some ways, this can reduce safety vs. profitability conflicts. However, regulations are subject to politics. When

12. Tools for IS Implementation 12.1 IS REVIEW METHODS – OVERVIEW 12.1.1 Three Approaches Inherent safety reviews of new and existing processes form the foundation of an IS management prog ram. IS reviews for new processes or modifications to existing proces ses present the best opportunity to identify possible design features th at are inherently safer than those proposed. For existing processes, th ese opportunities are more limited, due to the cost and feasibility involv ed in making major modifications. However, whenever a change or even an in-kind replacement is planned for an existing facility, an IS review should be conducted to identify any potential IS alternatives that can be incorporated into the design. Many of the more qualitative haza rd analysis tools can be adapted to incorporate the principles of inherent safety, and three basic approaches to conducting IS reviews have emerged in the chemical process industries. These are similar to PHA approaches that have been in use for many years: HAZOP What-If? Checklist The Checklist approach can be used in combination with both the HAZOP and What-If methodologies. Each of these techniques is described below. Because of the importance of consid ering inherent safety early in the design sequence when changes can most readily be made, inherent safety considerations are particular ly important in conducting hazard reviews (i.e., a preliminary hazard analysis). The above methods can be applied to these reviews. 302 (VJEFMJOFT
GPS
*OIFSFOUMZ
4BGFS
$IFNJDBM
1SPDFTTFT
"
-JGF
$ZDMF
"QQSPBDI #Z
$$14 ¥

5IF
"NFSJDBO
*OTUJUVUF
PG
$IFNJDBM
&OHJOFFST

52 PROCESS SAFETY IN UPSTREAM OIL & GAS annual plan to correct the deficiency. Wher e goals are achieved, then new actions or investments can be agreed upon so that process safety performance is not static but improves over time. This is continual improvement. 3.3 CONCLUSION The RBPS process safety management structur e is an accepted industry practice. It provides a useful basis for upstream process safety or for enhancing upstream safety management systems via the appropriate inclusion of better tools, techniques, and systems from RBPS. It is non-prescriptive and can be incorporat ed into a company- specific management system compatible with regulatory requirements globally. There are many other management systems that have varying numbers of elements. Some combine topics into fewer elements; others expand these out giving more elements. The number and arrangement are not important so long as the topics are covered. RBPS highlights some key topics not included in other systems. As is noted in the pillar Learn from Experience, those interested in process safety should strive to learn from their own experience as well as the experience of others. All of the aspects of RBPS should be considered as an opportunity to learn and to improve process safety performance, whether or not they are regulated at your specific operating location.

48 temperature, will be increased when th e device dimensions are reduced, although viscous losses will increa se. Secondly, the surface area-to- volume ratio of the system increase s as dimensions decrease, thereby increasing the interface area per uni t mass or volume. This increases the mass and heat transfer rates. Both fa ctors together can create extremely efficient mixing devices and heat e xchangers. Since volume is a three- dimensional property, the volume of the micro-device, and hence the amount of material in it, is reduced by the inverse of the third power of its characteristic dimensions. In doing so, material inventory is greatly reduced, as are concerns about la rge temperature and concentration differences because of the greatly reduced response time of the micro-device (Ref 3.20 Stankiewicz). 3.3 CONTINUOUS STIRRED TANK REACTORS A continuous stirred tank reactor (CST R) is usually much smaller than a batch reactor for a spec ific production rate. In addition to reduced inventory, using a CSTR usually resu lts in other benefits that enhance safety, reduce costs, and improve the product quality. For example: Mixing in the smaller reactor is generally better. Improved mixing may improve product uniformity and reduce by-product formation. Controlling temperature is easier and the risk of thermal runaway is reduced. A smaller reactor provides greater heat transfer surface per unit of reactor volume. Containing a runaway reaction is more practical by building a smaller but stronger reactor rated for higher pressure. In considering the relative safety of batch and continuous processing, it is important to fully understand any differences in the chemistry and processing conditions that may outweigh the benefits of a continuous reactor’s reduced size . Englund (Ref 3.7 Englund 1991b) describes continuous latex proces ses that have enough unreacted monomer in the continuous reactor to be less safe than a well-designed batch process.

18. Capturing, challenging and correcting operational error 217 • Failure in task verification – if task verification had been conducted correctly (by a peer or through ot her independent verification) the mistakes in the drawings or installa tion would likely have been detected. • Failure in error challenge skills – in the instance that someone had previously detected the fault in the P&ID and procedures, it could be that they felt unable to challenge or report the error, due to fear of repercussion, or lack of communicati on skills or error challenge skills. A useful source of information on latent error detection is provided by Saward and Stanton’s book - ‘Individual Latent Error Detection (I-LED): Making Systems Safer’ [72]. 18.3 Why do we fail to capture, ch allenge, and correct errors? Human performance is affected by cogn itive ability. Human errors, related to cognitive ability, can be grouped into four categories or stages – sensory, memory, decision, and action [73]. These are shown in Figure 18-2 and explained next. Figure 18-2: Categories of cognitive error 1. People first process information through sensing what is happening around them (use of sight, hearing, smell, taste, touch, and balance). 2. Next, information is retrieved from the memory. 3. Then, a decision is made on how to respond. 4. Finally, actions are initiated based on the decisions made. An error can occur at any of these stag es. Understanding this process is an important step in learning to manage er ror effectively. Reasons for failures in detecting, correcting, and/or challeng ing errors are shown in Figure 18-3.

74 Human Factors Handbook Figure 7-3: Task walk-through process (Compiled by CCPS) 7.5 Validation of job aids Even with the involvement of operat ional and maintenance teams in the development process, it is possible that new or amended job aid or SOPs may not be practical or may need improvement. This is especially true for new or upgraded equipment and processes. It is important to validate job aids on their first use and on an ongoing basis. An operational validation includes a r eview with operational and maintenance The walk-through process

9 • Other Transition Time Considerations 162 (Adapted from [21, p. Figure 5.1] ; and [85, p. Figure 6] ). Figure 9.2 Example project life cy cle stages in context of the equipme nt or process life cycle stages.

14.2 Seek Learnings | 181 Chen and Winston agreed with Wai-Kee and Mei. But Chen wanted more details. He said, “Come up with a plan including a timeline and what resources you think you will need. We’ll put it on the agenda for next Monday’s meeting.” Wai-Kee and Mei left the meeting and decided that they would meet at lunch every day to work on developing the plan. Wai-Kee was well versed in facility siting, having read CCPS’s Guidelines for Siting and Layout of Facilities (CCPS 2018) and taken training courses to keep himself up to date. While he worked on the siting analysis, Mei developed communications plans for the public and for first responders. By the end of the week, they were ready to present their plan to Chen. When Mei and Wai-Kee gave their presentation at the weekly management meeting, Chen and Winston were impressed by their thoroughness. They approved the plan and required that the two give regular updates at future management meetings. 14.2 Seek Learnings With everyone on board, Wai-Kee enlisted the help of one of his new hires, Anna, who had recently graduated at the top of her class. Wai-Kee said, “Anna, I need you to review the public literature and see what we can learn from past incidents.” Anna responded positively and made it a goal to review public databases from around the world for any ammonium nitrate explosions. She found numerous incidents involving ammonium nitrate, many with severe consequences. Anna tabulated all the data she could readily find and summarized the more notable incidents in chronological order (Table 14.1). Table 14.1 Ammonium Nitrate Incidents Year Place Country Site Tons NH4NO3 explodedFatalitiesInjuries 1921 Oppau GermanyPlant 450 561 1,952 1924 Nixon, NJ USA Plant — 18 100 1940 Miramas France Plant 240 1942 Tessenderlo Belgium Plant 15 189 900 1947 Texas City, TX USA Ship 2,300 581 3,500 1947 Brest France Port 3,000 26 5,000 1972 Taroom, QLD Australia Truck 18.5 3 — 1988 Kansas City, MO USA Truck 29 23 —

CONTINUOUS IMPROVEMENT 157 Other examples of metrics (both le ading and lagging) that facilities should consider include: Overdue reviews of operating proced ures, safe work practices, and maintenance practices that could lead to abnormal situations Number of times upper and lower operating limits are exceeded Overdue training, especially for op erators, on troubleshooting and managing abnormal situations Number of abnormal situations that occurred that were not already covered in HIRA studies Duration of inhibited or bypassed safety-critical elements/ equipment, instrumentation, and alarms that are essential for warning and managing abnormal situations Number of outstanding recommendations from management of change, auditing, and incident investigations that are relevant to reducing abnormal situations Incident and near-miss rates, espe cially high potential near-misses, arising from abnormal situations Number of repeat incidents associated with abnormal situations Number of risk evaluations completed and number of risks requiring action identified Number of risks requiring action that have been mitigated versus the total number identified Percent of control panel operators trained on recognizing abnormal process situations Percent of control panel operators trained on the alarm system and individual key alarms Number of alarms suppressed Number of alarms that are in constant alarm (stale alarms) Number of gaps in information communicated via shift handover, from checklist-based audits MOCs: percent of overdue action items MOOCs: percent of overdue actions on organizational changes Number of open action items from Pre-Startup Safety Reviews (PSSRs)

2 • Defining the Transition Times 25 Figure 2.2 Timeline used to determine which tran sient operating mode applies

W ITNESS M ANAGEM ENT 115 Drafting a list of potential witnesses at the start of the investigation is helpful, as the list can be modified as the investigation progresses and more witnesses come to light. Sources of information on possible witnesses include: • List of people associated with the facility • Operator’s and other logs • Permits to Work • Work schedules • Computer access records • Employee and visitor sign-in sheets • Names of personnel on work orders and procedures / risk assessments • Purchasing records • Design and drawin g documentation • Training documentation • Organizational charts • Lockout/tagout records • Audit records • Hospital admission records • Phone logs or records • Referrals made by current witnesses • List of personnel respon ding to the emergency • Contact with people outside of the facility • Responses to public advertisin g for the need for anyone with related information to come forward, possibly including people who have posted on social media 7.3 W ITNESS INTERVIEW S 7.3.1 H uman Factors Related to Interviews Humans are unable to record and pl ayback occurrences in perfect detail. Eyewitness accounts should be cons idered incomplete. Most of us have received little formal training in observation techniques. The common optical illusion amusements in Figure 7.3 remind us that our minds will often complete the expected or anticipated picture or image, even if it is not necessarily present. Consider the text in Figure 7.3. Most people will miss the repeated extra word. Similarly, witnesses may fill in data that are missing from their recollection of an occurrence or overlook data due to oversight or

Overview of the PHA Revalidation Process 13 Evaluate the Cumulative Effect of Changes in the Process, Equipment, or Personnel. Process, equipment, and organizational changes must be managed. Most PSM models, such as the 20-elemen t RBPS model developed by the CCPS, include a management of change (MOC) element to ensure that changes are thoroughly evaluated and properly authorized before implementation. However, MOC requests are often evaluated within a rather narrow context (i.e., "What are the potential consequences of wh at we intend to do right now?"). In cases where processes undergo frequent changes (e.g., a multi-purpose unit), several sequential changes (e.g., a phased equipment installation), or multiple simultaneous changes (e.g., during a unit shutdown/turnaround), the significance of a particular change may not be accurately assessed. In addition, subtle/creeping changes, or the inte raction of various changes, may be overlooked. Finally, not all changes may have been captured and evaluated under the MOC program, or temporary changes may not have been reverted to normal service. Thus, PHA revalidation offe rs an opportunity, on a periodic basis, to perform an integrated evaluation of the cumulative (and potentially interrelated) impact of all of change s, both controlled and uncontrolled. Correct Gaps and Deficien cies in the Prior PHA. Gaps are errors of omission (i.e., failures to address the establishe d requirements applicable to the prior PHA). Some companies have documented requirements for what a PHA must address. In addition, some regulations, as will be discussed in Chapter 2, are very specific in defining issues that must be addressed during PHAs (e.g., the consideration of human factors). Failure to address human factors, or some other requisite company or regulatory co nsideration, would be a gap that must be filled during the PHA revalidation. Deficiencies, on the other hand, are errors in applying the PHA methodology. For example, the prior PHA team may not have consistently (1) traced the effects of loss scenarios to their ultimate consequences, (2) judged the severity or risk of similar scenarios, or (3) documented all the engineering and administrative controls upon which their risk judgments were based. Revalidation teams, often including several or entirely new members, may identify hazards that the prior team did not document. Incorporate New Knowledge and Operating Experience. PHA revalidation teams may have access to information th at the prior PHA team members did not have available to them. Such inform ation might come from new company research, from work done by others and reported in industry literature, from recently issued or revised RAGAGEPs, from internal and external incident investigation learnings, from the work activities discussed in the “Process Knowledge Management” element of the RBPS framework, or through development of process safety information (PSI) as defined in United States

3.2 Characteristics of Leadership and Management in Process Safety Culture |83 3.2 CHARACTERISTICS OF LEADERSHIP AN D M AN AGEMEN T IN PROCESS SAFETY CULTURE Several key characteristics emerge from the Section 3.1 discussion of the basic themes of leadership in general and process safety leadership in particular. Like other core principals of process safety, the principal Provide Strong Leadership overlaps with other principals. Where appropriate in the following discussion, the overlap is noted. Set the Tone First and foremost, strong leaders/managers should set the overall process safety tone for the workplace. When leaders say, “Nothing is m ore important than safety,” they should m ean it. They should say it with a sense of vulnerability, as with the understanding of everything that process safety requires. This will help everyone else believe that the senior m anagem ent believes fully in the importance of process safety. Without this belief, little else will be possible. Only the senior managers can establish this belief and it must be created in both word and deed. It not only starts with management/leadership, but it continues with them as well. A single verbal message without follow-up actions, or no sustained transm ission of m essages will erode this belief. Also, inappropriate workplace behavior such as harassing behavior of any kind, unequal treatment by supervision or management such as favoritism or nepotism, or any other behavior that does not value and respect the people in the organization should not be tolerated in any way. It does not matter whether the behavior is face-to-face or occurs online. The existence of this type of workplace is a key cultural warning sign (Ref 3.16) of potential catastrophic incidents.

208 | Appendix: Index of Publicly Evaluated Incidents Section 3: Selected Causal Factors (Continued) Reactivity Hazards—Primary Findings (Continued) J155, J157, J158, J159, J160, J161, J162, J163, J169, J171, J187, J193, J197, J264, J267 S4, S11, S12, S14 Reactivity Hazards—Secondary Findings C30, C48, C50, C63 J8, J10, J12, J39, J44, J45, J54, J55, J65, J81, J86, J103, J105, J108, J114, J117, J127, J137, J153, J156, J252, J262, J271 S7 Relief System Design—Primary Findings A2, A5, A6, A10 C11, C59, C73 J48, J77, J153, J175, J192, J227, J258 S4, S16, Relief System Design—Secondary Findings C21, C37, C62 J10, J22, J44, J98, J116, J136, J142, J251 Safe Design/Error in Design—Primary Findings A2, A4, A5, A6, A7, A10 C2, C7, C13, C14, C26, C34, C39, C45, C46, C49, C56, C69, C70, C76 D20, D32 HA1, HA2, HA5, HA8, HB4, HB5 J19, J20, J21, J23, J29, J48, J57, J60, J62, J68, J69, J78, J80, J83, J87, J88, J90, J91, J93, J97, J103, J104, J105, J106, J107, J110, J111, J115, J119, J123, J124, J125, J126, J131, J132, J133, J134, J145, J151, J153, J155, J156, J158, J159, J160, J161, J163, J167, J169, J173, J176, J179, J182, J202, J207, J210, J212, J214, J215, J216, J218, J219, J226, J227, J228, J232, J235, J236, J237, J238, J239, J241, J244, J245, J250, J257, J258, J260, J265, J268, J269, J270 S3, S8, S9, S10, S16, Safe Design/Error in Design—Secondary Findings C3, C11, C15, C16, C20, C23, C24, C29, C32, C33, C36, C37, C38, C40, C41, C42, C44, C47, C48, C50, C52, C75 HA10, HB6, HB7, HB9 J9, J15, J37, J41, J42, J43, J44, J46, J49, J58, J61, J63, J66, J75, J77, J81, J84, J85, J86, J89, J99, J100, J101, J112, J116, J117, J118, J122, J139, J144, J149, J157, J164, J183, J206, J252, J261 S11, S13

236 | 6 Where do you Start? Explain the Personal B enefits When implementing any change, nearly everyone in the organization will want to know how the change will impact them . Leaders should explain to personnel their new expectations and should help personnel understand how everyone will benefit. Short-term, measurable goals should be set, and then progress reported so everyone can have a sense of accom plishment (Ref 6.5). 6.4 SUMM ARY Im proving the process safety culture of a facility starts with leaders understanding there is a problem and an improvement opportunity that it is worthy of the organization’s attention. The case should be built on facts as well as on costs and benefits for improving the culture. Once the case has been established, a baseline should be established through a culture assessment. The assessment should be built on interviews and record reviews, followed by focus groups to test improvement ideas. The formal improvement process should start by exam ining the state of the culture com pared to the culture core principles. The core principles should be considered roughly in the order presented, and addressed in small steps rather than trying to fix everything at once. In considering improvements, keep things as simple as possible, and use metrics to help reinforce both the vision and progress towards it. Above all, felt leadership – consistent and involved – needs to be sustained always.

224 | 6 Where do you Start? com munications during workers’ normal activities. Activities to observe include: Safe work practices: Hot work, line-breaking, equipment- opening, confined space entry, etc. Pre-start-up safety review meetings. Shift changes: If facility has separate shift changes for control room operators and field operators, observe both types. MOC review meetings. Contractor safety training (the assessment team itself m ight be subjected to this training to begin the assessment). Daily production meetings (meetings where operations and m aintenance activities are discussed, scheduled, and prioritized). Non-routine operations. Safety meetings or similar events where process safety issues are on the agenda. Once a pattern of behavior has been determ ined, assessors should engage in conversations with those being observed (Ref 6.3). Workers in organizations with Behavior-Based Safety (BB S) program s will be used to this for occupational safety. The goal of these discussions is to validate what was learned from surveys, interviews, and focus groups, and to identify specific opportunities to im prove process safety culture or the PSMS. Evaluate Symptom s and Causal Factors Observations generally start by recognizing sym ptom s of culture gaps. From there, assessors should focus on identifying the causal or contributing factors of the symptoms recognized. Causal factors generally are determ ined by finding the underlying reason for the sym ptom s. As discussed above, avoid confrontational questions that can put the interviewee on the defensive. Once the immediate underlying reasons are known, • • • • • • • •

150 | 11 REAL Model Scenario: Culture Regression own procedures.” Lucas drove the point home. “You don’t want to wait until you get to that point,” he said. “By then, it’s too late. As the senior people on this rig, we have to remember that the decisions we make affect our crew.” Lucas said, “Charlotte has some other findings that I think may interest you.” Charlotte said, “The human fatigue survey that we conducted with your crew provided some interesting results. It’s clear that the twelve-hour shifts are taking a toll on the crew. People are just not getting enough sleep.” Oliver said defensively, “Tell me something I don’t know.” Charlotte said, “On my survey, I asked if people would be willing to work a longer rotation, but with shorter work hours. Say either eight or ten hours instead of twelve. The response was overwhelmingly yes. It’s a win-win. They work fewer hours, but they stay longer, so there’s more continuity in personnel.” “That’s an interesting compromise,” Oliver said. “But I need to get a more detailed picture as to how it would affect my crew and potentially production.” Lucas responded, “We’ll get a couple of scenarios worked out in the next few weeks, and then we can get back together again.” Oliver nodded. “Sounds like a plan.” Lucas said, “One last thing. We really need a united front to bring all of this to Mason. Can we count on you?” Oliver said, “One step at a time. Let’s see what you come up with, and then we’ll talk. You’ve given me a lot to think about.” 11.6 Prepare Over the next few weeks, Charlotte and Lucas worked on a more detailed plan to show Oliver and then Mason. Charlotte worked on developing several scenarios for shorter shift hours but longer rotation. Lucas worked on ways to create a sense of vulnerability on the rig. At the follow-up meeting, Charlotte said, “I looked into various shift scenarios and recommend that we move forward with the ten-hour shift.” She based her recommendation on API Recommended Practice 755 (API 2019): • Work sets shall not exceed 9 consecutive day or night shifts. • There shall be 36 hours off after a work set, or 48 hours after a work set containing 4 or more night shifts. • Shifts are routinely scheduled for 10 hours and holdover periods should not exceed 2 hours and, where possible, occur at the end of the day shift.

106 Human Factors Handbook 9.7.2.3 Support the mind (consider the mental aspects of work) People are “programed” from birth to respond to information in the world in a certain way. Good design should work with rather against these natural tendencies. For example: • Stereotypes and conventions – follow known stereotypes or conventions, such as color-coding, opening and closing of things, turning items off and on. Ensure that this convention is applied consistently across the design of all equipment to avoid co nfusion. Thinking specifically of electrical panels with indicator lighting either green or amber, indicating active / energized or deactivated - this should be consistently applied across panels, buildings, etc. • Affordance – use the form of equipment or an object in the way it should be used. For example, buttons “afford” pushing, handles should be pulled as noted in section 9.4. • Uncluttered information – cluttered information can be off-putting and difficult to read. Ensure appropriate spacing and the use of blank areas, especially for Graphical User Interfac e (GUIs) displays, to help people view relevant information. • Simplicity – only provide information or controls that a person needs to do their job. • Co-location (items are near one another) – ensure related controls and displays are located adjacent to each other. • Consistency – ensure equipment design is consistent. For example, all hard-wired interfaces conform to the same rules e.g., the emergency shut-down button is always top right, and all local control panels have buttons in the same layout. • Feedback – provide timely information on user input so that users can tell when the system has been changed or is doing something. For example, with touchscreens (where th e feel of a physical button is not present), feedback can be provided with the touchscreen image changing color or flashing, or with “haptic” feedback such as vibration. • Natural mapping – set out information mimics (mimics are an exact or approximate graphical representation of a process plant with integrated indicators and instrumentation) and displays, so that they correspond to things in the real world. For exam ple, a touchscreen control panel to move a crane should be orientated to correspond with the crane’s actual movements. More information on na tural mapping is provided in section 9.7.2.4.

CASE STUDIES/LESSONS LEARNED 179 The associated HMIs were also key factors including: The PF and PNF have displays with airspeed indicators from different pitot tubes, with the potential to cause confusion when they provide different readings. It is not possible for the PF or the PNF to observe the position of both sidesticks simultaneous ly. Thus, unless the pilots communicate clearly, it is diffic ult for one pilot to understand the other’s control input. Several alarm messages indicate d on the ECAM (Electronic Centralized Aircraft Monitoring) system, and the report stated that these were read out by the PNF in a “disorganized manner”. Seven lines are available on the ECAM for message display and once those lines are full, a green arrow points downwards to indicate other messages of lower priority that ha ve not been displayed. To view those messages, the pilot must clea r the earlier mess ages, although it was not possible to determine if any of the crew cleared one or more ECAM messages during th e incident. No announcement, however, to this effect was made. The report states: No ECAM message enabled a rapid diagnosis of the situation to be made initiating the appropriate procedure. Information on the angle of attack is not directly available to the pilots. The sidesticks have no artificial feel but they do have a spring centering device when the stick is released. Theoretically, both the PF and the PNF can make simultaneous inputs to the sidestick, in which case, the flight system sums the input from both sidesticks, up to a pre-set limit. Simultaneous inputs from both sidesticks would trigger an audible alarm and a light on the instrument panel, although there was no report of simultaneous inputs being made during this incident. Nevertheless, if the PF and the PNF had different understanding of the situation, it would be difficult for the PNF to know the control input to the sidestick without good verbal communication. The loss of the Normal Flight Control laws meant that the systems that usually prevent the pilots from making control changes outside the operating envelope no longer existed. This is even more critical at high

132 Guidelines for Revalidating a Process Hazard Analysis As discussed in Chapter 1, any PHA can be divided into three parts: 1. Core Methodology/Core Analysis. T h e c o r e o f a P H A i s t h e identification of hazards and safeguards, typically using the HAZOP Study, What-If/Checklist, or FMEA technique. This is the foundational analysis that identifies the hazards of the process, the engineered and administrative risk controls (safeguards), and the worst credible consequences if all the safeguards failed. 2. Complementary Analyses. The PHA often includes additional analyses focused on specific topics that warrant further consideration using techniques other than those for the core analysis. These complementary analyses are often checklists and include studies such as facility siting studies, human factors analyses, damage mechanism reviews, and dust hazard analyses. 3. Supplemental Risk Assessments. Some PHAs include supple- mental risk assessments of selected loss scenarios. Often these are Layers of Protection Analyses (LOPAs), but they may include Bow Tie, Fault Tree, Event Tree, or Human Reliability Analyses. The distinction between these three parts of a PHA is important to the following discussions because the variou s parts may be revalidated differently, depending on the approach chosen. 7.1.1 Revalidation of the Core Analysis Redoing the Co re Analysis. The Redo approach is selected when there are many changes in the process equipment or procedures, the core methodology, existing node definitions or risk tolera nce criteria, or when the organization desires an independent assessment of the hazards. If the Redo approach was selected, the revalidation is conducted in almost exactly the same manner as the initial PHA. The process is divided into nodes (if the nodes used in the prior PHA are not suitable), and worksheets are pr epared. The team identifies potential loss scenarios for each node, and those sc enarios with consequences of interest (as defined in the revalidation scope) are documented. At a minimum, the documentation lists the hazards, the existi ng risk controls (safeguards), and the credible unmitigated consequences if the risk controls were to fail. The documentation usually includes a repr esentative list of specific hazardous events (causes, failure modes) that the te am used as the basis for its frequency judgment. The team then decides whether the risk of the scenario exceeds the organization’s risk tolerance. If so, the team notes the need for further risk reduction and may offer sp ecific suggestions to achieve tolerable risk.

Pipes 75 Because the pipe spec is just an acronym for the pipe feature, one may expect another document that outlines the detail of each pipe class. This is the Piping Material Specification Table. There could be less than 20 to more than few hundreds pipe specs in a Piping Material Specification Table of a project or plant. Figure  6.10 shows one page of a piping specification table, which belongs to imaginary pipe spec of A0. A Piping Material Specification Table is a large docu- ment, and thus it has a table of contents. The content are called a Piping Material Spec Summary, which is shown in Figure 6.11. The duty of the designer is choosing the suitable pipe spec for each pipe in the P&ID. To do that, three pieces of information are needed: the name of flowing fluid, its required design temperature, and its required design pressure (Figure 6.12). The designer may start with checking the piping spec summary to find the available specs for the commodity of interest. He/she may find two to three different suita-ble pipe specs suitable based on the commodity and the temperature range. Then the designer needs to pick one Table 6.4 Commodity acronyms and their meaning. Acronym Fluid name NG Natural gas FW Fire water IA Instrument air UA Utility air PA Plant air D Drain V Vent UW Utility water LPS Low‐pressure steam MPS Medium‐pressure steam HPS High‐pressure steam HGS Heating glycol supply HGR Heating glycol return CWS Cooling water supply CWR Cooling water return Figure 6.10 An e xcerpt of A0 pipe spec table.

Piping and Instrumentation Diagram Development 50 vessels. This frame is not for gases. Figure  5.9 shows a liquid/solid‐level frame with some examples for differ - ent levels. Such a frame is more common for liquids and because of that there can be an extra L in the acronym for different levels that represent liquid. For solids, only a few levels can be used in the frame. It is mainly because of difficulties in measuring the accurate level of solids in silos, where there may only be a high‐high level and low‐low level. 5.3.2.4 Flow Levels The level frame for flow is not as common as tempera-ture and pressure. Flow is a parameter that is mostly defined for pipes. However, in cases where it is defined for equipment, it refers to the flow from the piping to the equipment. Figure  5.10 shows a flow‐level frame with some examples for different flow levels. Among the five‐level frames of normal, high, high‐high, low, and low‐low levels, the last two are the most com-mon. This is because equipment is generally more tolerant of high flow rates but may be more sensitive to low flows. From a purely theoretical point of view, every piece of equipment can handle every flow rate, even huge ones! With higher flow rates, there is only a higher pressure drop. If a piece of equipment cannot handle a specific huge flow rate, it is not because of the flow rate, but because the flow does not have enough pressure to over - come the pressure drop within the equipment. For example, the low‐low set point flow for pipes is the minimum flow that keeps the pipe full of fluid. Partial flow in a pipe not only is uneconomical but also can cause corrosion problems. One exception is gravity flow, in which a partial flow must be maintained. However, “seal flow” is so low that it is rare that anyone specifies low‐low flow for pipes. For some equipment, low or low‐low flow rates are specified by the manufacturer. With a centrifugal pump, for example, a low flow rate is the flow rate generally specified by the manufacturer as the “minimum flow rate, ” which means any flow rate less than that and pump will be instable because of internal flow circula-tion. Low‐low flow rate could be defined as the flow rate that is not even able to fill the pump casing. Generally pump manufacturers do not bother to report this low‐low flow rate because of its rarity. A low flow rate can also be a problem in fired heaters because the tubes can burn out. 5.3.2.5 Analyte Levels A similar frame for analyte or composition can be defined. It is, however, not as common as other process parameters. Figure 5.11 shows such a frame for pH of a water stream. 5.3.3 Par ameter Levels versus Control System The control system will be discussed in more detail in Chapters 13–16. How can a control system work in a process plant? Let’s consider temperature as a process parameter in a warm lime softener as shown in Table 5.1. HSI le velOverflo w wor ks here NPSHA of downstreampump is calculated basedon this Not applicableHHL HL NL LL LLL LSI le vel Figure 5.9 Liquid/solid levels . HSI le vel HHF HF NF LF LLF LSI le velRedirecting flow to outside Rated capacit y of a piece of equipment Minimum flow of centrifugal pumps Seal flow of pipes Not applicable Figure 5.10 Flo w levels. HSI le vel HHpH LpHNpH Normal pHHpH LLpH LSI le vel Figure 5.11 Analyt e levels.

EQUIPMENT FAILURE 187 Figure 11.6. Pump explosion from running isolated (CCPS 2002 b) Design considerations for process safety. Two different types of pumps and compressors include centrifugal and positive displacement. When pumps and other rotating equipment are running, the process fluid can le ak from between the rotating shaft and the body of the pump. Leaks can result in fires or toxi c releases if the fluids are flammable or toxic. Different types of seal configurations are ava ilable to prevent these leaks. The selection of pump and seal type is usually dependent on pr ocess considerations. Every type of pump and seal has process safety considerations. With compressors, liquid entry into the compressor can cause catastrophic failure. Protection should be provided upstream of compressors to remove liquids and associated shutdown systems should also be provided. Centrifugal pumps (Figure 11.7) are susceptible to leaks, deadheading, running isolated, cavitation and reverse flow. Design configurat ions that have two pumps in parallel can be especially vulnerable to these failure modes because the possibility of starting the wrong pump. Centrifugal pumps, as with other rotating equipment, need shaft seals between the process fluid and the external environment. The simplest form of a seal is packing material. This can degrade with time and leak. Mechanical seals are the next type. In a mechanical seal pump, a seal face is kept in contact between the shaft and casing. These seals leak less than packing but do require a lubricating fluid that must be compatible with the process fluid. Mechanical seals can be single or double (Figure 11.8). In a double mechanical seal, two seals sit back to back inside a chamber external to the pump. The seal chamber is flushed with a fluid, and leaks are contained and can be detected in this fluid. Double mechanical seals are better at preventing leaks but require more complex maintenance.

1 • Introduction 4 the case of maintenance, good communication is essential so than everyone knows what others ar e doing at any time and what responsibilities each has. Split or unclear responsibilities are a recipe for disaster. [3, pp. 3-4] “ “To [help] prevent these ty pes of incidents from occurring, facilities should employ ef fective communications, provide workers with appropri ate training, and have in place strong and up-to-date policies and procedures for hazardous operations such as start-ups and shut-downs [5, p. 1] .” When the equipment is being shut down (the transition time), the word shut-down is used in this guideline. This spelling, with the hyphen between the words, distinguishes the activities taken on operating equipment during a transition (when the equipment is being shut-down ) and the activities conducted when the equipment is not operating (during a shutdown – no hyphen). In particular, the hyphen is essential when discussing incidents, as many incident reports designate “shutdown” without distinguishing between the shut-down mode and a project-related shutdown. Thus, the incidents in this guideline will focus on those that occurred when the equipment was being shut- down . For consistency within this gu ideline as well, when the process equipment or process is starting up, start-up with a hyphen will be used (recognizing that another commo n spelling is “startup”). This guideline will discuss two general transient operating modes (the time when a process is in t ransition between its idle and its operating states.) Specifically, the normal start-up and normal shut- down times are defined as: 1. Start-up time—from an idle, safe, and at-rest state to normal operations, and 2. Shut-down time—returning from normal operations to its normal idle, safe, and at-rest state. If there is an emergency shut-down, the controlled end state of the process equipment should be idle, sa fe, and at-rest. If the emergency

394 Figure 15.2 Modified process 15.3 CASE STUDIES FROM CARRITHERS The case studies and examples in the following section are taken from a presentation entitled, “It's Never Too La te for Inherent Safety,” by G.W. Carrithers, A.M. Dowell, and D.C. Hendershot (Ref 15.2 Carrithers).

360 | Appendix F Process Safety Culture Assessment Protocol 138. Is there a system , with effective accountabilities, for ensuring that recommendations from risk assessments are implemented in a tim ely fashion, and that the actions taken achieve the intent of the original recommendation? 139. Are the hazard/risk analysis performed as part of the MOC process adequate? Has this part of the MOC review process become somewhat pro-form a with little effort beyond routing the MOC to someone in the Safety Departm ent for a routine review? 140. Are conflicts of interest allowed in the assignment of HIRA/PHA team leaders? For exam ple, the process/project engineer who is responsible for the unit/system being studied should not lead the PHA on that process but should be a team m em ber. 141. Does a questioning attitude prevail at all levels of the organization regarding the hazards/risks? 142. Are process safety risks and related controls communicated throughout the organization and beyond (contractors, other com panies)? 143. Does m anagem ent “face the facts” when necessary in response to process safety issues? Conversely, are difficult decisions regarding process safety issues routinely deferred hoping that the situation will be resolved in a different way? 144. Has the As Low As Reasonably Practicable (ALARP) principle been applied in m aking decisions about hazard/risk abatem ent? Has the ALARP principle been applied reasonably and consistently? 145. Have form al definitions of tolerable risks that have been agreed-to by the entire organization been adhered to without regard to their ramifications? For example, if a risk based inspection (RB I) program has been implem ented have the ITPM frequencies that allow the process safety risk to remain at a tolerable level been followed even if this requires that equipment be shutdown unexpectedly to perform a needed test or inspection?

22 INVESTIGATING PROCESS SAFETY INCIDENTS Although these are the result of action s or inactions by people, this does not imply that people are to blame. In reality, human factors are a contributing or intermediate causatio n, but it is weaknesses in the management system(s) that have allowed contributions, such as those listed above, to exist. 2.2.4 Multiple Causation Incidents are generally not the result of a single cause or act, unless an individual deliberately decides to work unsafely or damage/sabotage a chemical process. Even in such ex treme deliberate acts, engineering and management controls that might have minimized the probability and/or consequence of the act should be considered as part of security vulnerability assessments. Most incidents have multiple root causes, and certain combinations of those causes can give rise to accidents or near-misses. Some of these causes may have resulted in near-misses or minor incidents on previous occasions, i.e., less severe precursors such as scenarios when a barrier failed but the event did not propagate to adverse co nsequences. A thorou gh investigation of these types of events will not only find the root causes of the subject incident, but will also find other root causes that were near-misses. It is therefore an avoidable mistake to stop an investigation after identifying only one root cause. If the near-misses ar e not investigated, they may cause a future incident even if the root causes of the subject inci dent are corrected. 2.2.5 Events vs Root Causes An event (including a non-event, i.e., an omission) cannot be a root cause because it is either a causal factor or the consequential result or symptom that follows a root cause. For example, the operator opened the drain valve is an event that led to a spillage of hazardous material. In this case, the root cause is related to why the operator opened the dr ain valve – was it due to inadequate training, hu man error, or anothe r cause? Similarly, failure to follow procedure is not a root cause. It is a symptom of an underlying cause. W e are too much accustomed to attribute to a single cause that which is the product of several, and the majority of our controversies come from that. Marcus Aurelius

Appendix A – Concluding Exercises These concluding exercises bring together topi cs from multiple chapters of this book illustrating how the various proc ess safety elements are rele vant to a single facility. Exercise 1: LNG Value Chain The LNG value chain involves offshore production of gas, pipeline transportation of the gas to shore, treatment and liquefaction of the gas, storage and loading of the gas onto ships, shipping the LNG, and finally a receiving terminal where the gas will be used. You are involved in the project to design and construct the LNG receiving terminal. The chemical process is relatively simple. The LNG is offloaded from ships into large LNG storage tanks. The LNG storage tanks consists of a stainle ss steel inner tank that contains the LNG, about 1 meter thickness of insulation, and an outer reinforced concrete tank to provide secondary containment and store the LNG at atmospheric pressure and - 260 F. From the tanks, the LNG is vaporize d, also referred to as regassified. Before the natural gas is delivered by pipelin e to the customer, it is odorized with an unpleasant smelling odor to aid in leak detection. References that may be helpful for this ex ercise include those from The International Group of Liquefied Natural Gas Importers available at these links. https://giignl.org/sites/default/files/PUBLIC _AREA/About_LNG/3_LNG_Safety/giignl2019_infop apers4.pdf https://giignl.org/sites/default/files/PUBLIC_A REA/About_LNG/4_LNG_Basics/giignl2019_infop apers2.pdf 1. Name 2 codes or standards that might apply to this project. One code should be specific to the design of the facility. The other should address management of the risks. 2. Describe the physical properties of LNG. Is it hazardous? Cite your sources. 3. Beyond chemical hazards, what other hazards might warrant consideration? 4. Have there been any LNG accidents in industry that you can learn from? 5. Make a plan for what process safety studies and activities you will do, or have done, at what stage of the project. 6. As this is early in the project and details ar e not available, a Preliminary Hazards Analysis is being conducted. List 10 specific questions that should be considered. For each, identify potential consequences. 7. For the scenarios identified in the Preliminar y Hazards Analysis, list potential methods to prevent or mitigate the consequences. 8. What inherently safer design options might be considered for this project?

3.3 Leadership vs. Management |97 Strong process safety leadership refers to the ability of a person to convince his/her reports and peers of the right process safety thoughts and actions – winning their hearts as well as their m inds. Senior m anagers should be process safety leaders. Additionally, in a strong process safety culture, m id-level m anagers, supervisors, technical specialist, and even front-line employees can and should be leaders also. True leadership is not conveyed by one’s position on an organization chart. Effective leaders inspire their reports and co-workers and earn their respect with direction and advice that is sound and consistent. Leaders accept direct accountability for all things that occur within their sphere of responsibility. They do not attem pt to publicly place blam e on their subordinates when things go wrong. More than anything, subordinates will not want to disappoint som eone who has earned their respect as a leader. Visionary and inspiring managers who are also good leaders are comm itted to doing what is right, and demonstrate their values through their com munications, actions, priorities, and provision of resources. 3.4 CON SISTEN CY OF PROCESS SAFETY MESSAGES In its investigation of process safety culture in BP’s USA refineries, the B aker Panel (Ref 3.23) found that workers had received many m essages from management over the years addressing m any values. These tended to dilute the im portance of any value generally, and certainly of process safety. This happens in many companies. Leadership communicates m ission and vision statements, core values, central tenets, and overarching principles, wishing to better define what their organizations stand for and how they operate. Many times, these

156 Guidelines for Revalidating a Process Hazard Analysis The PHA revalidation report format and content may be prescribed by facility or company requirements. This chapter provides suggested docu- mentation practices for the PHA revalidation and associated records that could be used, if specific local requirements are not provided. 8.1 DOCUMENTATION APPROACHES The spectrum of approaches to reva lidating a PHA naturally results in a spectrum of documentation styles. For convenience and consistency, the terms Redo and Update are used herein to compare and contrast the documentation that typically results from those approaches. But, just like the revalidation activities themselves, the documentation styles are not mutually exclusive, and practical applications usually result in a combination of both being used for different portions of the PHA. It is possible that a team using the Update approach might copy all relevant information from prior PHA(s) into the current report, so the reader would not need to refer back to the older documents. In that case, the Update analysis would be documented in the Redo style. With Redo documentation, a new and detailed PHA report, similar in format and content to the initial PHA report, is prepared. This single document: • Identifies the process unit exam ined and the reason(s) various portions of it were included in the PHA (e.g., Nodes 1-26 are required by national regulations; Nodes 27, 32, and 41-48 are required by local regulations; and all other nodes are required by company policy) • Lists meeting participants and who, specifically, filled roles required by regulation or policy. Some organizations include a log of the activities that occurred on each meeting day Terminology Note In the first edition of this book, the terms “ basic” and “ evergreen” were used to describe revalidation documentation styles. Those terms were eliminated in this edition and replaced with Update and Redo , corresponding to the revalidation approaches. Some organizations may continue to use the older terms because they are embedded in their internal guidance. The term “evergreen” mentioned in Section 1.6 refers to an ongoing revision of the current PHA analysis worksheets, NOT a revalidation.

FIRE AND EXPLOSION HAZARDS 67 If a fuel/air mixture forms a vapor cloud and finds an ignition source, it could create a flash fire back to the release point. Alternatively, where the conditions are suitable, it could create a vapor cloud explosion. An explosion is describe d by a pressure-time curve as shown in Figure 4.12. The damage caused by an explosion is du e to the pressure and the impulse, described by the area under the pressure-time curve. The curve is at a stationary point at some point away from the explosion center. An explosion in which the reaction front is less than the speed of sound is a deflagration. Where it is greater than the speed of sound, it is a detonation. The overpressures, and potential resultant damage, are significantly greater for a detonation than for a deflagration. The speed of the reaction front is influenced by three main factors: fuel reactivity, congestion or obstruction, and confinement. Th e fuel reactivity is related to the laminar burning velocity. Congestion or obstruction desc ribes the size and number of blockages in the path of the reaction front (Figure 4.13). Confin ement describes the limits on how an explosion can propagate. For example, an explosion in an open field can expand in more directions than one confined between two plates or ultimately, on e that is confined in a pipe shape, like the barrel of a gun. Figure 4.12. Explosion pressure-time curve (FEMA) Vapor Cloud Explosion - The explosion resulting from the ignition of a cloud of flammable vapor, gas, or mist in which flame speeds accelerate to sufficiently high velocities to produce significant overpressure. (CCPS Glossary)

152 Guidelines for Revalidating a Process Hazard Analysis • Reviewing specific changes (esp ecially changes made due to recommendation or incident) since the prior PHA when performing a Redo . Important insights may be missed if changes resulting from incidents are not specifically evaluated for their effectiveness in preventing recurrence • Fully describing the scenarios, including the intermediate events that led to the consequences, so that future revalidation teams can understand the current PHA team assessment • Using the experience of the PHA team to revalidate supplemental risk assessments (e.g., LOPA/QRA), as well. The team can provide insight as to the functionality and reliability of critical safeguards used in these assessments and Update them accordingly • Being vigilant for unresolved reco mmendations from the prior PHA • Using checklists for special or uni que requirements that apply to processes to remind PHA revalidation team members of key points to verify Obstacles to Success: • Using an inadequate team and/or facilitator expertise • Failing to obtain stakeholder agreement on the purpose, scope, methodology, and schedule prior to the first team meeting • Repeatedly using the same study leader and/or team to Update the PHA of a particular unit • Relying on team members with t oo much personal investment in the prior PHA (e.g., they do not think anything should be changed or that any recommendations should be made), despite evidence of inaccuracy or inconsistency • Failing to manage conflicting priori ties impacting the team meeting schedule and the participation of required personnel • Using inadequate PSI (e.g., incomplete P&IDs or operating procedures) • Failing to revalidate complementary and supplemental analyses, in addition to the core analysis • Failing to consider past incidents and operating experience, including those incidents and changes that were not controlled • Failing to evaluate how adequately the recommendations of the prior PHA were resolved • Failing to maintain consistency with the prior PHA when Updating causes, consequences, safegu ards, or risk rankings

Evaluating Operating Experience Since the Prior PHA 75 4.2.2 Incident Reports Incident investigation reports are an im portant type of operational experience that will strongly influence the reva lidation approach. The term incident encompasses events with actual process safety losses and/or the reasonable potential for process safety losses (“near misses”). Some companies broaden the definition further to include any demand on, or activation of, an engineered safety feature (e.g., a relief valve lif t or a deluge activation, performance deficiencies in those systems, or exceed ances of safe operating limits). The goal o f p r o c e s s h a z a r d a n a l y s i s i s t o r e d u c e t h e r i s k o f p r o c e s s s a f e t y l o s s e s t o tolerable levels, so major loss events should be relatively rare or nonexistent. To supplement the reported incidents, the revalidation leader should solicit the recollections of team members about abnormal situations where any safeguard failed, but a loss was averted. Evalua tions of emergency response drills and exercises involving on-site and/or community resources may also provide valuable insights. Lessons learned fr om major incidents involving similar chemicals, equipment, or processes in industry are often available from reputable sources, such as those listed in Table 4-2, as well as from books such as Incidents That Define Process Safety [25]. Table 4-2 Example Sources of Incident Information from Industry Regardless of the source, incident reports provide a wealth of information that is useful to the revalidation team. Therefore, when looking at a unit’s incident history and contemplating revalidation, answers to the following questions are crucial: • Did the PHA identify the loss scenario that occurred? If not, was that an isolated oversight or indicative that the PHA team overlooked other scenarios? The PHA team may have been CCPS Process Safety Beacon CCPS Process Safety Incident Database European Process Safety Centre EUROPA – eMARS Dashboard – Major Accident Reporting System of the European Commission United Kingdom Health and Safety Executive United States Chemical Safety and Hazard Investigation Board

214 Security risk can therefore be cons idered a function of consequence (C), threat (T), vulnerability (V) and attractiveness (A) or: Risk (R) Security = [Consequence (C), Threat (T), Vulnerability (V), Attractiveness (A)] With regard to security, it is appropriate to define the terms of “consequence,” “threat,” “vulnerability,” and “attractiveness.” Consequence is the severity of loss or damage that can be expected from a successful attack against an asset. Examples of relevant consequences include: Injuries to the public or to workers Significant environmental damage (such as contamination of drinking water) Direct and indirect significant financial losses to the company Disruption to the national, regional, or local operations and economy Loss of business viability The US Department of Homeland Security (DHS) has identified five potential consequence categories or se curity issues asso ciated with the production, use, storage or distribu tion of chemicals: (Ref 9.4 CCPS) 1.Release . Toxic, flammable, or explosive chemicals or materials that, if released from a facility, could create significant adverse consequences for human life or health. 2.Theft or Diversion . Chemicals or materials that, if stolen or diverted, could be used as weap ons or easily converted into weapons using simple chemistry, equipment or techniques. 3.Sabotage or Contamination . Chemicals or materials that, if mixed with readily available materials, could release poisonous gasses or create other significant ad verse consequences for human life or health. 4.Critical Relationship to Corporate or Government Mission . Chemicals, materials, or facilit ies that, if unavailable, could

3 • Normal Operations 41 Similar to the normal process shut-downs, normal process start- ups will be process-specific, and for that reason, no approach described herein would apply to eve ry process start-up, either. An example approach for a safe start -up for a continuous process may involve the following steps: 1. Pressuring up the equipment (i.e., to prevent flashing of high pressure feed sources), 2. Adding external heat and/or cooling to the equipment, 3. Introducing the feed stream(s), 4. Adding heating and cooling sou rces to the streams, and 5. Bringing the equipment to the normal processing operating conditions. Although each process will have sp ecific procedures, the steps for a normal start-up procedure may includ e the same steps in reverse of the steps used in the process’s normal shut-down procedure. 3.6 Incidents and lessons learned Note: All incidents which occur during normal operations are, by definition, covered in abnormal and emergency operations (Part II of this guideline). Before describing incidents from normal shut-downs and start- ups, it is well known that deviations from the established written start-up procedures, especially when co mbined with bypassing critical safety devices, can result in incide nts. For example, an incident that injured four employees occurre d when an automated process controlling the sequence from star t to finish was interrupted and the supervisor allowed maintenance technicians troubleshooting the process to override the computer safeguards [26]. A frequent cause described in incident case studies for semi-batch processes with a high exotherm reaction was lack of agitat ion or delayed agitation when the

137 It should be noted here that ther e is a tendency fo r some inherent risks to be volume or quantity depend ent. A material that is considered “safe” at lab scale may carry hi gh risk when managed at bulk manufacturing quantities, or even at pilot plant scale. Eventual scale must be considered when evaluati ng risks even at the research and development phase. For example, do ing a process hazards review of a pilot plant run is essential for safe pilot plant operations, but may not reveal inherent hazards of bulk sc ale production. Virtual “scaling-up” must be part of the early life cycle inherent safety review. A similar issue may arise as a result of lab equipment being different from manufacturing equipment. A reaction th at is safe in lab equipment may not be in the plant. These issues need to be taken into consideration in moving from lab scale to pilot plant to full-scale manufacturing. As described in Chapter 2, the life cycle of chemical processes consists of eight stages. Figure 8.1 repeated from Chapter 2 depicts these life cycle stages. 8.2 CONCEPT Processes are conceived in a combination of technical activities “on paper” and in the laboratory. It is du ring the Concept stage that a product is conceived based on a perceived or real industrial or business need. This translates to the need for particular physical and chemical properties to achieve the desired compound. This, in turn, begins to narrow the choice of feed materials and possible intermediates to arrive at a desired final product as the basic chemistry is worked out. Final, or even initial equipment de signs, are not chos en during this stage of the process life cycle, so the IS strategies of Minimization , Simplification , and to a large degree, Moderation may not be applicable during this stage. For example, an objective to “intensify” a process and reduce the scale poses a challenge to the designers to meet that goal and results in reduced quantities of chemicals and potential exposures. However, this is the stage when Substitution can have its largest impact. Selecting an intrinsically safer, le ss hazardous chemical to provide the needed reaction(s), if possible, at th is stage is far easier than during subsequent life cycle stages.

Hall, N. (7 October, 1994). Chem ists clean up synthesis with one-pot reactions. Science, 266, 32-34. Hannon, J. (1992). Clean technology through process intensification. IChemE North Western Branch Symposium Papers, 3 , 7.1- 7.6. Harris, C. (1993). Containment enclosures. In V. M. Fthenakis (Ed.). Prevention and Control of Accidental Releases of Hazardous Gases (pp. 404-410). New York: Van Nostrand Reinhold. Hawksley, J. L., and. Preston, M.L (1996). Inherent SHE—20 Years of Evolution. In Internatio nal Conference and Workshop on Process Safety Management and Inhe rently Safer Processes, October 8- 11, 1996, Orlando, FL (pp.183-196). New York: American Institute of Chemical Engineers. Heil, J.A. (1995). Inherent Safety Characteristics of Innovative Nuclear Reactors . Petten, Netherlands: Netherlands Energy Res. Foundation. Hendershot, D. C. (1987). Safety considerations in the design of batch processing plants. In J.L. Woodward (Ed.). Proceedings of the International Symposium on Preventing Major Chemical Accidents , February 3-5, 1987, Washington, D.C . (pp. 3.2-3.16). New York: American Institute of Chemical Engineers. Hendershot, D.C. (1988). Risk reduction alternatives for hazardous material storage. In Proc. 1988 Hazardous Materials Spills Conference, May 16-19, 1988, Chicago, IL (pp. 611-618). New York: American Institute of Chemical Engineers. Hendershot, D.C. (1988). Alternat ives for reducing the risks of hazardous material storage facilities. Environmental Progress 7 (3), 180- 184. Hendershot, D.C. (1991). “Design of inherently safer chemical processing facilities.” Presented at Texas Chemical Council Safety Seminar, June 11, 1991, Galveston, TX, Session D. Hendershot, D.C. (1993). Inherently safer plants. In Guidelines for Engineering Design for Process Safety (pp. 5-52). New York: American Institute of Chemical Engineers. 479

214 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Rollover, the spontaneous and sudden movement of a large mass of liquid from the bottom to the top of a storage tank due to heat from the fire changing the fluid density gradient Example 1. The Buncefield explosion and fire described at the start of this Chapter is an example of a gasoline storage tank overfill inci dent. The ensuing fire engulfed more than 20 large storage tanks over a large part of the Buncefield depot. Example 2. The Cleveland LNG Tank failure in 1944 was the worst LNG accident in the USA. Wartime shortage of nickel led to use of a lower nickel stainless steel material than the now standard 9% Ni steel. This lower nickel st eel was subject to brittle fracture, and this occurred leading to a catastrophic tank failure and total loss of the LNG tank contents. The LNG spilled into the tank yard and boiled off to me thane. This cold methane, as liquid and cold vapor, was not lighter than the surrounding air. Thus, it flowed into the nearby neighborhood down streets and in sewer lines resulting in130 fatalities. Example 3. In 1919 a 8,700 cubic meters (2.3 milli on gallon) tank of molasses suddenly broke apart, releasing its contents into the City of Boston. A wave of molasses over 5 m (15 ft) high and 50 m (1600 ft) wide surged through the streets at an estimated speed of 60 kph (35 mph) for more than 2 city blocks (Figure 11.28) . The incident led to 21 fatalities and over 150 injuries. The tank was not properly inspected du ring construction and not hydrotested before filling it. Leaks between the welds had been ob served, but no action was taken. (CCPS 2007) Figure 11.28. Molasses tank failure; before and after (CCPS e) Example 4 . A delivery truck arrived at a plant with a solution of nickel nitrate and phosphoric acid named “Chemfos 700” by the supplier. A plant employee directed the truck driver to the unloading location and sent a pipefitter to help unload. The pipefitter opened a panel containing 6 pipe connections, each of which fed to a different storage tank. Each unloading connection was labeled with the plant’ s name for the material stored in the tank. The driver told the pipefitter he was delivering Chemfos 700. Unfortunately, the pipefitter connected the truc k unloading hose to the pipe adjacent to the Chemfos 700 pipe, labeled “Chemfos Liq. Add. ” (Figure 11.29). This is similar to the human factor issue in the Formosa Plastics explosio n (Chapter 16). The “Chemfos Liq. Add.” tank

148 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION continually improve this performa nce. Metrics such as the Recordable Incident Rate, and Days Away from Work Cases have become common safe ty performance metrics. The focus of these metrics, however, is solely on occupational safety. Industry recognized the value of the occupati onal safety performance metrics and saw the need to improve process safety performance by creating process safety-specific metrics. In 2007, CCPS published a document defini ng process safety metrics. API RP 754 Process Safety Performance Indicators for the Refining and Petrochemical Industries was originally published in 2010. (API RP 754) The upstream oil and gas industry aligned with refining and petrochemical in creating a parallel document IOGP Report 456 – Process safety – recommended practice on key performance indicators . (IOGP 456) This allowed integrated oil and gas companies to collect consistent data statistics for upstream and downstream operations. Since the original metrics documents were issu ed, CCPS, API, and IOGP have updated their relevant documents while keeping them aligned. An effort is underway to align the metrics with the UN, Globally Harmonized System of Classifi cation and Labelling of Chemicals (GHS) . (UN) The metrics alignment has enabled companies to: Measure process safety performance, identify underperformance, and drive continuous improvement, and Compare company performance to indu stry performance, past company performance, or intra-company performa nce, and continuously improve process safety performance. The system used in the CCPS, API, and IOGP documents is a four-tier system as shown in Figure 9.3. This approach is similar to Heinrich ’s incident pyramid that depicts a larger, bottom- level of minor incidents (a larger area of the tr iangle representing a larger number of incidents), a mid-level of incidents, and a top, small leve l of more serious accidents (the smaller area representing less incidents). In the process sa fety pyramid, Tier 1 and Tier 2 have been designated Process Safety Events (PSE) that have occurred. Tier 1 PSEs are of greater consequence; Tier 2 PSEs of lesser, but still serious, consequence.

86 4.32 Starks, C.M., Phase transfer catalysis: An overview. In C.M. Starks (Ed.). Phase Transfer Catalysis: New Chemistry, Catalysts and Applications , September 8, 1985 (ACS Symposium Series No. 326). Washington, D.C.: American Chemical Society, 1987. 4.33 Sundell, M.J., and Nasman, J.H., Anchoring catalytic functionality on a polymer. Chemtech, 23 (12), 16-23, 1993. 4.34 Tietze, L.F., Domino reactions in organic synthesis . Chemistry & Industry, 453-457, 19 June 1995. 4.35 U.S. Environmental Protection Agency, Design for the Environment Alternatives Assessments, www.epa.gov/saferchoice/design- environment-alternatives-assessments. 4.36 U.S. Environmental Protec tion Agency and Occupational Safety and Health Administration, Chemical Safety Alert: Safer Technology and Alternatives, June, 2015. 4.37 U.S. Environmental Protection Agency, Office of Pollution Prevention & Toxics, Design for the Environment Program Alternatives: Assessment Criteria for Hazard Evaluation, v2.0 , August, 2011. 4.38 U.S. Environmental Protection Agency, Presidential Green Chemistry Challenge : 1996 Greener R eaction Conditions Award – The Dow Chemical Company, www.epa.gov/greenchemis try/presidential-green- chemistry-challenge-1996-greener -reaction-conditions-award 4.39 U.S. Environmental Protection Agency, Presidential Green Chemistry Challenge: 1996 Greener Synt hetic Pathways Award (Monsanto), www.epa.gov/greenchemistry/presidential-green-chemistry-challenge-1996-greener-synthetic-pathways-award 4.40 U.S. Occupational Safety and Health Administration, Transitioning to Safer Chemicals: A Toolkit for Employers and Workers (www.osha.gov/dsg/safer_c hemicals/index.html). 4.41 Walsh, F., and Mills, G., Electrochemical techniques for a cleaner environment. Chemistry and Industry, 576-579, 2 August 1993.

140 | REAL Model Scenario: Leaking Hoses and Unexpected Impacts of Change After discussing the contents of the modified policy with Antônio, Márcia developed a poster showing a monkey climbing, with a big red X through it, with text explaining that the only appropriate places to stand were the floor, a ladder, a scaffold, or a person-lift. The posters were distributed around the plant. An image of the poster was also incorporated into the plant training video. 10.8 Embed and Refresh About 18 months later, Antônio pulled into a carpark spot next to Adriana. After they’d wished each other “Bom dia,” Adriana said, “I’ve been meaning to ask you, Antônio. You addressed the problem of stepping on hoses by changing the work-at-heights standard. And I guess that was a good thing because it also protects insulation, piping, and cables. But it really doesn’t prohibit us from stepping on a hose that’s on the floor.” They passed through the gatehouse, exchanging hugs with the guard. As they stepped back outside, Antônio smiled at her. Adriana was becoming a seasoned operator. Someday she’d take João’s job for sure. They stopped in front of the monkey poster. He pointed at it and said, “Doesn’t say you can stand on a hose, does it?” “So, walking on it should be OK!” she exclaimed, and slipped through the locker room door. Antônio got a double coffee from the cafeteria and sat alone at the back, thinking about Adriana’s joke. It was a joke, wasn’t it? He thought about the new reactor system that would be added into the high bay. They would be hiring new operators and running more complicated processes. There would be a detailed HAZOP, ongoing construction activities, and new operations. Lots of opportunities for things to fall through the cracks. He reflected on the work they’d done to figure out the reason the hoses were failing. Initially, they’d checked everything that anyone could expect to be a problem and found nothing. Who would have expected that a change in lighting would cause operators to deviate from their normally diligent performance?

41 2.2 American Chemical Society (ACS) Public Policy Statement 2015-2018, Inherently Safer Techno logy for Chemical and Related Industrial Operations, 2015 . 2.3 Amyotte, P, et. al., Reductio n of Dust Explosion Hazard by Fuel Substitution in Po wer Plants, Trans IChemE (81), Institution of Chemical Engineers, 2003. 2.4 Amyotte, P.R. Goraya, A.U, He ndershot, D.C., and Khan, F.I. (2006). Incorporation of inherent sa fety principles in process safety management. In Proceedings of 21s t Annual International Conference– Process Safety Challenges in a Glob al Economy, World Dolphin Hotel, Orlando, Florida, April 23-27, 2006 (pp.175-207) New York: American Institute of Chemical Engineers. 2.5 Belgian Administration Of Labour Safety, Technical Inspectorate, Chemical Risks Directorate, Process Safety Study: Practical guideline for analysing and managing chemical process risks, 2001. 2.6 Berger, S, Hendershot, D., Fa mini, G., Emmett, G., Defining Inherently Safer Technology to Fo cus Process Safety and Security Improvements, Chemical News, August 2010. 2.7 Carrithers, G., Dowell, A., Hendershot, D., It’s Never Too Late for Inherent Safety, International Co nference and Workshop on Process Safety Management and I nherently Safer Processes, American Institute of Chemical Engineers, 1996. 2.8 Center for Chemical Proc ess Safety, Guidelines for Engineering Design for Process Safety , American Institute of Chemical Engineers, 1993. 2.9 Center for Chemical Process Sa fety, Guidelines For Initiating Events And Independent Protection Layers In Layer Of Protection Analysis, American Institute of Chemical Engineers, 2014. 2.10 Center for Chemical Proc ess Safety, Hazard Evaluation Procedures, 3 rd Ed., American Institute of Chemical Engineers, 2008. 2.11 Center for Chemical Process Safety, Layer of Protection Analysis -Simplified Process Risk As sessment, American Institute of Chemical Engineers, 2001.

334 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION a. No known ignition sources are present b. Fired heater is located within the potential area of the flammable cloud Estimate the procedure reliability for a si mple well-documented procedure assuming the operator is well-trained, under low stre ss, and sufficient time to diagnose the problem and execute the appropriate actions. Also assume no feedback such as a buzzer or bell indicates that the procedure has been executed properly. Explain your results. Explain the difference between individual risk and societal risk. References AE Solutions, AEShield, https://www.aeshield.com/online-store. Bloch 2019, “Looking Back at the Phillips 66 Expl osion in Pasadena, Te xas: 30 years later”, Bloch, K. P., Vaughen, B. K., Hydrocarbon Processing, American Institute of Chemical Engineers, New York, N.Y., October. CCPS Glossary, “CCPS Process Safety Glossary ”, Center for Chemical Process Safety, https://www.aiche.org/ccps/resources/glossary . CCPS 1989, Guidelines for Process Equipment Reliability Data , Center for Chemical Process Safety, John Wiley & Sons, Hoboken, N.J. CCPS, 1999, Guidelines for Chemical Processe s Quantitative Risk Assessment , Center for Chemical Process Safety, John Wiley & Sons, Hoboken, N.J. CCPS 2001, Layer of Protection Analysis - Simplified Process Risk Assessment , Center for Chemical Process Safety, John Wiley & Sons, Hoboken, N.J. CCPS 2008, Incidents That Define Process Safety , Center for Chemical Process Safety, John Wiley & Sons, Hoboken, N.J. CCPS 2009, Guidelines for Developing Quantitative Safety Risk Criteria , Center for Chemical Process Safety, John Wiley & Sons, Hoboken, N.J. CCPS 2013, Guidelines for Enabling Conditions and Conditi onal Modifiers in Layers of Protection Analysis , Center for Chemical Process Safety , John Wiley & Sons, Hoboken, N.J. CCPS 2015, Guidelines for Initiating Ev ents and Independent Protection Layers in Layer of Protection Analysis , Center for Chemical Process Safety , John Wiley & Sons, Hoboken, N.J. DNV GL, SAFETI, https://www.dnvgl.com/services/qra-software-safeti-1715 . DOD 2012, “Standard Practice, System Safety, M il-STD-882E”, U.S. Department of Defense. EPD, https://www.epd.gov.hk/eia/register/report/e iareport/eia_1252006/html/eiareport/Part3/Sect ion13/Sec3_13.htm HSE a, “Guidance on ALARP Decision in COMAH”, https://www.hse.gov.uk/foi/internalops/ hid_circs/permissioning/spc_perm_37/. HSE b, “ALARP at a glance, https://www.hse.gov.uk/managi ng/theory/alarpglance.htm

126 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS 5.4.1 Expectations of Policies and Administrative Procedures Clear and accepted policies and admi nistrative procedures are essential to establishing the minimum expectat ions of personnel. For example, are the policies clear to all personne l regarding their authority to make timely decisions? Note this incl udes the “Stop Work Authority”, which allows any personnel to request a ha lt in procedures or operations if there are safety concerns. Are organizational responsibilities documented? Are communication requirements between operating shifts, between operating teams and maintenance, and between operating teams and leadership wri tten down and followed? Formal, written policies and procedures are recommended over depending only on guidelines. Often, guidelines ca n imply that the expectation of conformance to the regulations is op tional or that they are merely suggestions. 5.4.2 The Relationship of Policies to Abnormal Situation Management Policies and administrative procedures evolve over time as a company or plant site matures. Most companies now have established policy manuals that cover safety and envi ronmental procedures, as well as onsite and offsite emergency response procedures. Management guidelines are then written to ma nage changes to these policies and procedures. Policies also establish a working culture, for better or worse. Cultural issues are a recurring theme in some of the example incidents in Chapter 3. A positive working culture indicates: An environment that respects and supports each team member and decisions that are made, without the benefit of hindsight. Admitting, and sharing/learning from, mistakes. Proactive seeking of learnings from others. Open and structured communications between operation teams and others. It is not always easy to establish an effective relationship between teams. Plant operating teams/shifts tend to build autonomy among themselves especially in the presence of a strong or dominant team leader. While this can promote team pride and build intra-team

191 experiences excessive corrosion an d must be replaced, alloy(s) that have been shown to resist or prevent the corrosion mechanism in question should be used. This decision depends on the remaining life of the equipment and the possible severity of the consequences associated with the release of the material. This is an example of making th e process more inherently robust (Ref 8.46 Hurme). These suggestions for improvin g operational and maintenance practices can also be considered as ap plications of the four IS strategies, although they do not directly relate to the chemical hazards or physical design of the process itself. S ee Chapter 14 for a more detailed discussion of the application of IS concepts and strategies to the programmatic aspects of other pr ocess safety program elements. 8.8 CHANGE MANAGEMENT Management of Change (MOC) progra ms are used to manage process changes on process equipment (n ot replacement-in-kind), utility systems, operating and other procedur es, as well as su pporting policies, practices, and procedures. They ar e elements of all process safety programs required by regulation , and are often key elements of voluntary, consensus process safety programs. MOC programs impose a review and approval process that helps ensure that changes are thoroughly vetted before they are physically implemented to manage risk. MOC programs should also preserve and protect against elimination of inherently safer features. For ex ample, debottlenecking projects are intended to increase throughput and process efficiency. However, increasing unit or plant throughput requires that equipment size be increased at various points in the pr ocess(es). Increasing the size of a valve or installing a larger pump could result in high pressure in a vessel, thus increasing the risk of a release. Sanders (Ref 8.73 Sanders) presents a number of examples of changes affecting the safety of a plant. MOC procedures and forms should be modified to include appropriate reviews and verifications that confirm: that inherently safer strategies previously incorporated have not been compromised, and,

5. Human performance and job aids 51 Figure 5-1: Overview of Human Factors aspects of developing a job aid Technically validate & approve Operationally validate Update & maintain job aid Task characterization Select type of job Task analysis & task walkthrough Hazard Identification & Risk Analysis Draft the job aid Engage Users Apply Human Factors guidance

7 Developing content of a job aid 7.1 Learning objectives of this Chapter By the end of this chapter, the reader should be able to: • Understand the use of task analysis and the results of Hazard Identification and Risk Analysis in development of job aids. • Understand the role of worker involvement in the development of job aids. This Chapter builds on the CCPS guide “Guidance for Writing Effective Operating and Maintenance Procedures” [25]. In particular, this Chapter cites the use of task analysis, the output from Hazard Identification and Risk Analysis (HIRA) and task walk-throughs to help produce job aids. The Energy Institute provides a detail guide on how to perform task analysis [29]. 7.2 Outputs from task analysis Task analysis includes identifying the task steps, describing the task actions, and assessing the judgments and decisions n eeded to perform a task. The outputs include: • A detailed step-by-step record of tasks and sub-tasks. • An estimated time to complete the task. Task analysis can also be used to: • Help judge the minimum number of people needed to perform a task. • Identify the competences and skills required to carry out a task. • Develop training and selection requirements, and • Support error analysis. A detailed description of task steps can be used to write the steps in a procedure or in another form of job aid. This may be in the form of a diagram or a table (list). The development of task analysis can be time consuming, but it can be used for other purposes. For example: • As a source for writing a step-by-step SOP. • As a training aid – to visually show the sequence of steps and to help develop learning objectives. • To support a critical review of the ta sk. Is it possible to do the task in fewer steps or with fewer people? Human Factors Handbook For Process Plant Operations: Improving Process Safety and System Performance CCPS. © 2022 CCPS. Published 2022 The American Institute of Chemical Engineers.

312 | Appendix E Process Safety Culture Case Histories fire drifted over the local community, and as a precaution, local authorities ordered comm unity evacuation lasting two days. It was the third incident with a similar cause experienced by the com pany. The incident investigation found that while cleaning an out-of- service reactor, an operator forced open the bottom valve of the wrong reactor, bypassing a critical safety interlock by attaching an air hose adapted to fit an instrument air connection to the “open” port of the valve. A label on the hose described it as an “Em ergency Air” hose. The contents of this reactor, hot reacting vinyl chloride monomer and partially formed PVC, drained onto the floor. Shortly afterwards, the flamm able m ixture ignited. The resulting fire killed the operator, his supervisor, and 3 other operators. The investigation found that the “Em ergency Air” line was provided to allow operators to drain the reactor in a runaway reaction scenario in case the norm al vent and relief system alone were not sufficient to control the pressure during a runaway reaction. It seems clear in hindsight that the drained m ixture would have ignited as occurred in this incident and therefore m ay not have provided much m itigation benefit. Instead, the “Em ergency Air” line had becom e routinely used for what the operator thought he was doing – forcing open the bottom valve of a reactor being cleaned, rather than opening it according to procedure, from the panel board on a higher floor. The incident was investigated by the US Chemical Safety B oard (CSB ). In their report, CSB pointed out that the com pany’s learning culture m ay not have been sufficiently strong. What other culture gaps m ight have contributed to this incident? Did the PHA team fully understand and act on the hazards and risks of the emergency procedure to drain hot, reacting VCM from the bottom valve using an “emergency air line”? Did operators feel any sense of vulnerability that they m ight open the wrong valve

INTRODUCTION 9 timelines, fault trees, logic trees, predef ined trees, checklists , and application of human factors. Examples are included to demonstrate how they apply to the types of incidents readers are likely to encounter. Chapter 11—The Impact of Human Factors This chapter describes human factor co nsiderations in incident investigation. It provides insight and tools to identi fy and address applicable human factor issues throughout an investigation. Practical models are presented along with examples. Chapter 12—Developing Effective Recommendations Once the likely causes of an incident have been identified, investigation teams evaluate what can be done to help prevent recurrence or mitigate consequences. The incident investigat ion recommendations are the product of this evaluation. This chapter addresses types of recommendations, attributes of high quality recommendations, methods to document and present recommendations, and related management responsibilities. Chapter 13—Preparing the Final Report In the case of incident investigation, a major milestone is completed when the final incident investigation report is submitted. The incident report documents the investigation team’s findings, conclusions, and recommendations. This chapter describes practical considerations for writing formal incident reports, and discusses the attributes of quality reports and differences among incident notifications, interim reports, and a final report. Considerations and associated practical techniques are provided for stating report scope, preparing preliminary no tices, documenting the investigation process and results, developing a report format, and performing a quality assurance check that includes management review and approval. Chapter 14—Implementing Recommendations The recommendations generated from an incident investigation when implemented in a timely and effective fashion, decrease the probability of recurrence, and/or reduce the potentia l consequences of an event. This chapter begins with case examples th at underscore key concepts, and then focuses on the critical aspect s of effectively implementing recommendations. It addresses initial resolution of the recommendations, their full implementation , effectiveness of follow-up, and tracking.

15.5 EXAMPLE OF AN INHERENT LY SAFER STUDY OF A STEAM PRODUCTION FACILITY The following comes from a paper written and presented by Karen Study at a 2005 safety conference sponsored by the Mary Kay O’Connor Process Safety Center in College Station, Texas. (Ref 15.13 Study) Choosing an inherently safer altern ative may seem straightforward. However, sometimes what initially seems to be the most obvious IS alternative may not actually provide the best overall risk reduction. In this case study, an “inherently safer” alternative was selected and later discarded due to issues uncovered du ring the detailed design phase. The option ultimately chosen was inherently safer than both the original design and the “inherently safer” alternative. 15.5.1 Facility Description The unit produces a large amount of steam using a multiple burner boiler with natural gas and a low BTU off-gas as its fuel sources. The boiler waste gas (flue gas) is sent to an elevated stack where it is discharged to the environment. This flue gas is mainly nitrogen and water, with oxygen and ca rbon dioxide. As with all boilers, there is also NOx present in the flue gas. A team was formed to assess different NOx reduction options. After evaluating several options to achieve the required NOx emission reduction ta rgets, the design team chose to install a Selective Catalytic Reactor (SCR). 15.5.2 Initial Design Proposal (Liquid Anhydrous Ammonia) To supply ammonia to the SCR, the de sign team chose to tap into an existing liquid anhydrous ammonia piping header that supplied a nearby processing unit. Piping was minimized as much as possible, to ~600 feet of 2 inch pipe. A vaporizer skid, whic h used steam to vaporize the liquid ammonia prior to injecting into the SCR, was to be installed near the boiler. See Figure 15.9 for a high-l evel overview of this option. After the option was selected, the process safety group was consulted to provide input. Due to concerns regarding incrementally increasing risks associated with the current liquid anhydrous ammonia piping system, the safety group recommended using aqueous ammonia available from a nearby processing unit. This seemed to be a 412

xxvi PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION GHS Globally Harmonized System HAZID Hazard Identification Study HAZMAT Hazardous Materials HAZOP Hazard and Operability Study HEART Human Error Assessment and Reduction Technique HIRA Hazard Identification and Risk Analysis HRA Human Reliability Analysis HSE Health and Safety Executive (U.K.) HTHA High Temperature Hydrogen Attack HRO High Reliability Organization I&E Instrument and Electrical IDLH Immediately Dangerous to Life and Health IEC International Electrotechnical Commission IOGP International Association of Oil & Gas Producers IOW Integrity Operating Window IPL Independent Protection Layer ISD Inherently Safer Design ISO International Organization for Standardization Isom Isomerization Unit ITPM Inspection Testing, and Preventive Maintenance JSA Job Safety Analysis KPI Key Performance Indicator LFL Lower Flammable Limit LNG Liquefied Natural Gas LOPA Layer of Protection Analysis LOPC Loss of Primary Containment LOTO Lock Out Tag Out LPG Liquefied Petroleum Gas LSIR Location Specific Individual Risk MAWP Maximum Allowable Working Pressure

5. Human performance and job aids 47 Job aids also minimize the potent ial for error. This is because: • Operators often require instruction on the correct way to operate a system. If the system is changed, they need updated instructions and information to help them to correctly operate the changed system. • Process safety requirements (i.e., operating within specific parameters, such as pressure, temperature, flow rate and material composition) are accurately communicated to operators. • Job aids can help people to remember steps in long or repetitive tasks where it can otherwise be easy to forget or unintentionally skip steps, especially if the task is complex, time pressured, performed less frequently and where a risk of distraction or fatigue is present. • Job aids can specify critical steps which are actions, or inactions, that are irreversible and if performed incorrectly can result in significant harm. Several human performance tools are at the disposal of personnel when executing critical tasks, including the work planning, STAR method, and three-way communication. • Job aids can indicate safety critical task steps that should be double- checked or independently verified, helping to spot and/or recover from errors. • Job aids can include “Hold Points”, where work is paused and checked. This can help snap a person out of “fast brain mode” (where someone is skillfully performing a task with little conscious thought), and allows them and others to double check their work with a pair of “cold eyes” (fresh eyes). • Stressful situations, such as emergency response, can reduce the ability to think clearly and accurately (limiting cognitive capacity). In these conditions, job aids can reduce the demand placed on memory and cognitive capacity (information processing and decision-making) and help to ensure successful task performance. Decision trees, alarm response procedure, manuals, process flow diagrams and other process information can help people to understand how equipment and systems work, what the hazards are, how to operate safely and what can lead to “Developing, documenting, and maintaining process knowledge is one of two elements in the Understanding Hazards and Risk Pillar.” ”Documented, current, and accurate operating procedures help ensure that each shift team operates the process in a consistent, safe manner.” CCPS “Guidelines for Risk Based Process Safety” [5]

xxx Human Factors Handbook Engineers, process safety practitioners and regulators who wish to gain an understanding of Human Factors concepts and methods will find much of immediate practical value. This book has been written by a combined panel of plant operations professionals with in-depth knowledge of a wide range of process plants together with very experienced Human Factors experts. It has then been widely peer- reviewed, resulting in a comprehensive han dbook that is easy to follow. Each of the 26 chapters contains essential knowle dge, presented in a straightforward, accessible manner and supported by nu m e r o u s e x a m p l e s t o s h o w w h y t h e concepts are relevant in processing indust ries. A notable feature is the analysis of major accidents from this sector that reveal where human factors contributed to failure or recovery during the event. Practical tools and techniques are provided for each topic area with guidance for application and more experienced pr actitioners will discover new ideas for their portfolio of Human Factors methods. This valuable handbook is definitely recommended reading for those striving to improve the safety and efficien cy of process plant operations. Rhona Flin Professor of Industrial Psychology Aberdeen Business School Robert Gordon University

94 •Atorvastatin calcium is a drug that lowers cholesterol by blocking its synthesis in the liver. The ke y chiral building block in the synthesis of atorvastatin is ethyl ( R)-4-cyano-3-hydroxybutyrate, known as hydroxynitrile (HN). Traditional commercial processes for HN require a resolution step (where the racemic mixture is separated into its two enantiom ers) with 50% maximum yield or syntheses from chiral pool precursors. They also require hydrogen bromide to generate a bromohydrin for cyanation. All previous commercial HN processes ultimately substitute cyanide for the halide under heated alkaline conditions, forming extensive byproducts. These processes also require a difficult high-vacuum fractional distillation to purify the final product, which decreases the yield even fu rther. Codexis has designed an alternative HN process around the exquisite selectivity of enzymes and their ability to catalyze reactions under mild, neutral conditions. The evolved en zymes are so active and stable that Codexis can recover high-quality product by extracting the reaction mixture. The process involves fewer unit operations than earlier processes, most notably averting the need for fractional distillation of th e product. The formation of byproducts and the generation of waste is reduced, avoiding hydrogen gas, and reducing the need for solvents and purification equipment. This process utilizes fewer hazardous materials and allows for more moderate operating conditions (aqueous, pH ~7, 25–40 °C, atmospheric pressure) (Ref 5.9 EPA). 5.4 SECONDARY CONTAINMEN T - DIKES AND CONTAINMENT BUILDINGS Most secondary containment systems are considered passive protective systems. They do not eliminate or prevent a spill or leak, but they can significantly moderate the impact without the need for any active countermeasures. Containment systems can be defeated by manual or active design features. For example, a dike may have a drain valve to remove accumulated rainwater, and the valve could leak or be left open. Another example is a door in a cont ainment building that could be left open.

APPLICATION OF PROCESS SAFETY TO ONSHORE PRODUCTION 103 Barrier Analysis The barrier approach as employed in bow tie analysis is now being used more frequently onshore. It is essential that personnel know what the important barriers against major incidents are and particularly those where they have a role in operating or maintaining them. Process safety events are rare and unless fully explained, personnel may not understand the role of barriers in preventing or mitigating these rare events. This contrasts with occupatio nal safety events wh ere the barriers are relatively easy to identify. Knowledge of barriers also underp ins effective implementation of MOC procedures. Barriers can be degraded during changes and such changes must be managed to ensure the barriers are returned to full effectiveness or are replaced with new ones. 5.3.3 Learning from Experience As with well construction, onsite pr oduction facilities employ process safety programs to ensure the safety and environmental management system is fully functioning. Key RBPS Elements include Incident Investigation, Measurement and Metrics, Management Review and Continual Improvement , and Auditing . To facilitate measurement and improving pr ocess safety performance, industry has developed standard process safety indi cators as described in API 754 or OGP 456. Incidents should be recorded and categorized using these documents into one of four tiers. Many companies share this data (e.g., IOGP, 2019a), and this allows companies to benchmark their performance. Tiers 1 and 2 represent larger and medium size events. Tier 3 represents demands on safety systems (e.g., relief valve openings) and Tier 4 represents deficien cies in activities but no loss of primary containment (e.g., not meeting requirements of safety management system). Many companies have committed to publicly reporting the number of Tier 1 and 2 events, and some are also considerin g reporting Tier 3 events. These direct measures of process safety indicators help to identify areas for improvement. Auditing is a formal review of all the mana gement system elements to ensure that all specified processes are in place and functioning. Management Review is a periodic ongoing process that assesses safety performance against current targets and de termines whether existing controls are adequate or need improvement. If the aim is to drive performance to improve current targets, then that is termed Continual Improvement . 5.3.4 Emergency Management Emergency Response Plans Regulatory requirements for emergency re sponse plans for larger onshore facilities are provided in OSHA PSM and EPA RMP or in the Seveso Directive. Smaller facilities follow state requirements or industry good practice. The RBPS element Emergency Management provides details. The emergency response plan should

208 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS than had been predicted by the si mulations. However this incident demonstrated that the thermal st ability was not actually greater. The design of the 60 Still Base in cluded a thermocouple that, under normal circumstances would measure the temperature of the liquid in the drum. However, it was position ed above the heater batteries and under the situation where the liquids had been removed from the drum; it was in the gas space and did not measure the sludge temperature. This meant that the sludge could have be en at a much higher temperature than the specified 90 °C, and the surface temperature of the steam batteries could have been higher still. The failure of the steam regulator and operating it on bypass should have been the subject of a management of change (MOC) procedure. The effect of the change in design (removal of the early stage to separate nitrocresols) and change in procedures (operating steam regulator on bypass) should have b een reviewed in the light of the operating experience of the unit. Th e changes should also have been a factor in the procedure that was adop ted for the removal of the residues. Principles of Inherently Safe Desi gn would encourage the removal of potentially unstable materials as earl y as possible in the process, or ideally to prevent their formation in the first place. 7.3.7.3 Management of Organizational Change (MOOC) There had been several recent chan ges to the organization and the Castleford site, and the reporting lin es had changed significantly. The report states that the newly designated team leaders had not received adequate training and that there were a number of errors on the permit and preparation of the safe system of work. The newly appointed area manager did not check the system of work, or the permits, and his attention was distracted due to other problems. MOOC should consider the level of labor and skills available to safely deal with abnormal situations, pa rticularly when switching from a hierarchical to a matrix organization. 7.3.7.4 Human Factors and Culture Employees were pressured to use th e 60 Still Base quickly in order to process the high stocks of whizzer oil. It may be that there was pressure to clear out the drum quickly, so that the impact on production would be limited.

W ITNESS M ANAGEM ENT 135 7.5 CONDUCTING FOLLOW -UP INTERVIEW S Following further evidence collectio n and causal analysis/ hypothesis development, more direct and structur ed questions can be developed for follow-up interviews. Conduct these in the same general manner as other interviews, but use a more direct, straight-to-the-point interview style. Initially, the interviewer may use op en-ended questions, but follow-up, closed-ended questions are us ually asked sooner than they would be asked during the initial interview. Focus on the gaps in information and apparent inconsistencies. However, take care to ensure that witnesses do not believe that the follow-up interview indicates the interviewer doubts their credibility; rather, emphasize that the investigation team is simply trying to gain greater clarity. 7.6 RELIABILITY OF W ITNESS STATEM EN TS Some of the details provided by th e witnesses may be inaccurate or inconsistent for various reasons as discu ssed above. It is possible that there may be more than one interviewer leading the various interviews who may record their observations differently. A key challenge is to compile the information received in a consistent ma nner, combine it with other evidence in a timeline, and determine which witness information is reliable and which is not. These issues need to be consi dered as the evidence is analyzed, which is discussed further in Chapter 9. 7.7 SUM M ARY Witness information is vital data and ca n come from a number of individuals and groups. However, it is quite fragile , so great care should be taken to get the most complete and accurate info rmation possible. Human recollection is imperfect and is easily biased, but by applying the techniques described in this chapter, the interview team can extract the best quality information from the witnesses.

Fundamentals of Instrumentation and Control 263 contact‐type level sensors) process fluid through a capillary. There may be two incoming signals into transmitters and if not a simple single parameter is being transmitted. The signal to the transmitter is not only from the sensor or primary element. For example in the case of flow orifice as the flow sensor, two signals go to the transmitter from pressure points upstream and down-stream of the flow orifice. The transmitter output signal is almost always elec - tronic. This signal mostly goes to a controller but it can go to an indicator or other devices too. 13.11.3 C ontrollers The function of the controller block is to compare the input signal against an SP and then generate an output signal that is proportional to the deviation. Controllers are usually located in the control room, so the tag has a divider; because they are “brains, ” they are represented by a circle within a square (Figure 13.35). The input and output signals from the controller are always electronic. They are delicate systems and should be located indoors, inside field cabinets/housings, or in the main control room. The SP is not usually shown on P&IDs. In addition the values of the SPs are unknown and if someone is interested they need to refer to the SP table to see the value for that particular controller. However, if the SP is a “remote set point” and comes from another control system, then it is shown as a soft - ware signal. 13.11.4 I ndicators Indicators are instruments that show parameters any - where, even in remote area like in control rooms. Indicators can be shown in three scenarios (Figure 13.36): 1) Dire ctly on the process stream in the field (Figure 11.36(a)). These indicators are installed in the field and are used for checking a process parameter by the rounding operator in the field. These indicators take the signal from a transmitter, even if transmitters are not shown. The tag doesn’t have a divider because it’s in the field. These indicators can potentially be replaced with gauges. There are plenty of field transmitters that are provided with the capability of indicating. This means the majority of XTs are XITs. There are cases that prevent us from providing an indicator in the field. One of them could be the harsh-ness of environment. 2) Indep endently, but from the control loop and in control room (Figure 11.36(b)). These are the indi-cators that we use these days mainly for process parameters that need to be visible in the control room. Here the indicator tag has a divider because it is located in the control room. 3) As p art of a block with other main functions in the control room (Figure 11.36(c)). These indicators were very popular in the past. In the early days we tried to say, “control this parameter and also show it to me in the control room. ” However these days, with the implementation of HMIs (human machine interfaces), almost everything is already visible in the control room and through the monitors, even if we don’t ask for it. So we no longer need to use tags like TIC or FIC. 13.11.5 Final C ontrol Elements in a BPCS The action of a BPCS can be either regulatory or discrete via “control loops. ” The final control element could be a variety of items but two of the most common final elements are control valves and VSDs on electric motors. The BPCS may also handle discrete actions. This may be for a batch operation such as filtration, where we have Set PointFlow Controller Usually in control room Loop No.FC 1051 Figure 13.35 Con troller block.OR(a) TI 437 FT 1051FT 1051FIT 1021FC 1021FI 1051 FV 1021 TIT 437FIC 265(c)(b) Figure 13.36 Indica tors.

TOOLS AND METHODS FOR MANAGING ABNORMAL SITUATIONS 129 Example Incident 5.3 – Hydrogen-in-Chlorine Explosion (cont.) The incident investigation found several contributing factors: The new pressure transmitters were wired backwards. No PSSR was conducted to chec k the pressure controls before startup. Leadership was under pressure to start up the plant as the schedule was past due. Communication between plant leadership and operating teams was strained due to several issues encountered during the shutdown. Lessons learned in relation to abnormal situation management: Organizational Chain of Command: During this stressful abnormal situation, the unit leaders overrode the chain of command and empowerment of personnel, who did not consider they had “Stop Work Authority”. Management of Change- Pre-Startup Safety Review: Although the MOC was conducted for the changes, the PSSR was not conducted because startup was overdue. Learning from Incidents: History across the chlor-alkali industry has included many hydrogen-in-ch lorine explosions. The highest risk is during startups. This was not considered during this plant’s startup. 5.4.3 Process Metrics Metrics are an important consideration of many business models. Metrics have been created for business goals, quality, safety, environment, security, training, and mechanical integrity. Metrics are also highly relevant in the manageme nt of abnormal situations. It is worth noting that CCPS considers measurements and metrics of such high importance that they are includ ed as one of the four elements in the RBPS pillar of Learning from Experience . (CCPS/RBPS 2007a). CCPS has

68 INVESTIGATING PROCESS SAFETY INCIDENTS investigative tools. It is a good practice to identify internal and external resources available to assist with these tasks. Suggested topics include: • An overview of the company in cident investigation management system • Incident investigation concepts, including the fact-finding, not fault-finding philosophy • Specific investigation techni ques used by the organization • Interviewing techniques • Gathering evidence • Developing and testing hypotheses • Identifying Causal Factors • Using tools to determine caus al factors and root causes • Writing effective recommendations • Documentation and report requirements • The roles of the team members • Confidentiality of the investigation Team member training may also include “role playing” for activities such as witness interviews, conflict resoluti on, and confidentiality issues. Team members should under stand that they are not ex pected to perform at the level of full-time professional investigators. They should feel free to request help or training as soon as they reco gnize a need. After initial training and accreditation, brief periodic refresher-training sessions or tabletop role- playing drills are a good way to reinfo rce the training objectives. Summary training topics may include: • Site-specific incident investigation plan • General roles and responsibilities • Specific assignments for team members such as interviewing, photography, and other roles • Evidence preservation and handling protocols • Locations for evidence storage • Controlling communications from team members Investigation Leaders Some organizations break this training into two or more levels, with team leaders given more training if they will lead investigations of higher level or complex incidents. Leaders learn how to determine the appropriate investigation methodology, how to gather data, how to analyze data for causal factors, how to determine root causes, and how to develop effective recommendations and reports.

INVESTIGATION M ANAGEM ENT SYSTEM 59 4.2.2 Specifying and Managing Documentation The management system should specify documentation requirements for interim data and work products of th e investigation. The company’s legal staff may have a valuable opinion on this guidance or they may offer case- by- case opinions. For example, the legal department may wish to be involved with witness interviews and physical evidence colle ction and management. Certain documents or evi dence may need special attention due to potential litigation. It is important not only to document investigation activities appropriately, but also to properly manage all documents and evidence developed by the investigation team. The team needs to develop a control system to track all documentation and evidence. A log should be developed, and every piece of evidence or docu mentation should be given a unique identifier number/code and entered into the log. Legal counsel should also be consulted on the scope of distribution lists of documents that are prepared by the team. If the investigation is being conducted under attorney–c lient privilege, counsel will determine the scope of those who need to be on distribution lists. Do not forward any documents, emails, communications or information to any other person unless expressly permitted by legal coun sel. Otherwise, such distribution may waive attorney-client privilege and/or work product. It is important to keep control of preliminary copies and draft reports issued for team review and comment. A good practice is to include a full distribution list on each copy, so that receiv ers of the document know who else has been copied. This is especially important on sensit ive documents relat ed to accidents. In addition to the use of headers and footers noting confidentiality, expert investigators include DO NOT COPY on some documents and always use the pagination style that notes the identification “ this is page x of y" markers on certain documents. A chain of custody should be maintained for all evidence that is moved to a different location or transferred to a different party. It is likely that items could be sent for examination by interested parties for testing by a specialist. It is essential to preserve the condition and quality of the evidence as well as to know pr ecisely where it is at any given time. Incident investigation document retention is another important issue to consider. Lawyers and investigation team members are likely to disagree about which documents to keep and how long to keep them. Retained documents may be useful to maintain corporate memory; however, retained documents may also create increased legal liability. Each organization must

4.6 Summary |153 discussed throughout this book, but it is particularly useful in identifying normalization of deviance. Fatigue, resulting from excessive overtime, can lead to conditions conducive to normalization of deviance. Overtime records can be trended in various ways, including cumulative overtime, num ber of extended shifts, and fraction of workers extending their shifts in each time period. Learn to Assess and Advance the Culture In m any ways, the actions taken resulting from m etrics for the other culture core principles indicate how well the facility and company is learning, assessing, and advancing the culture. Participation in voluntary process safety activities within the company and in trade and professional groups indicates the degree to which learning from outside the com pany is being considered. 4.6 SUMM ARY The process safety culture of the organization depends heavily on hum an behavior. Leadership can influence this behavior positively or negatively, as can many outside influences. Ethics can be a motivating force, especially if ethical behavior is modeled by leaders. Compensation can play a role in driving the desire culture, however, leaders should exercise care to prevent com pensation from unintentionally driving undesired behavior. Assessing the existing culture and then im plementing changes to correct it can be challenging, but ultim ately should be done to focus efforts where they can make the biggest difference. Ultimately, the application of the core principles of process safety culture is a journey. Leaders and employees need to put in the work to build a strong culture. There are no shortcuts. • • •

PROCESS SAFETY REGULATIONS, CODES, AND STANDARDS 41 Incident Investigation Compliance Audits Pre startup safety review Emergency planning & Response Trade secrets Employee Participation CCPS built upon these fourteen elements in cr eating the Risk Based Process Safety system. Also, in the US, the EPA “Risk Management Plan (RMP) Rule” is equally a key regulation. It implements Section 112(r) of the 1990 Clean Air Act amendments requiring facilities that use certain hazardous substances to develop a Risk Management Plan. EPA RMP focuses on people and the environment outside of a facility. (EPA a) In the EU, the SEVESO Directive was introduc ed in 1982 following the Seveso incident in Italy in 1976 (see section 1.5 for details) and the Flixborough explosion in 1974 (see section 18.2 for details). The Seveso Directive has been updated and currently Seveso-III is in effect. The Seveso-III-Directive (2012/18/EU) aims at the prevention of major accidents involving hazardous substances and at limiting thei r consequences both to humans and the environment. (EC a) The “Control of Major Accident Hazards (COMAH) Regulation” is the regulation that enforces Seveso Directive in the U.K. (HSE) Offshore in the U.K., the “Offshore Installation s (Safety Case) Regulations” were created in 1992 following the Piper Alpha incident in 1988. (HSE 2015) In the U.S. offshore, following the Deepwater Horizon incident in 2010, the “S afety and Environmental Management Systems Rule”, 30 CFR Part 250 Subpart S, was finalized. (BSEE) Similar process safety management regulati ons cover operations onshore and offshore in Norway, Australia, Canada, and other countri es. In countries where no such regulations exist, many companies choose to follow the re gulations cited previously in this section. Examples of process safety regulations ar ound the world are listed in Table 3.1.

Appendix 223 A.3 Managing the unexpected during transient operating modes Although there were no references to the “expecting the unexpected” concept in the transient operating modes incidents, the concept is not new [121]. Based on the significant process safety incidents that have occurred during all operating modes, there is much room for applying the concept within each mode. Since there should be an understanding what types of unexpected situations that can occur, this section provides more detaile d guidance on how best to anticipate unplanned but “expected” sh ut-down situations for: Loss of utilities (Section A.3.1) Engineering control issues (Section A.3.2) Administrative control issues (Section A.3.3) This section concludes with a brie f discussion on the dangers of “normalizing the deviance” and maintaining a sense of vulnerability when managing the risks of haz ardous processes (Section A.3.4). A.3.1 Anticipating and addressi ng unplanned loss of utilities As was noted in Chapter 7, a hazards analysis team can use a checklist for evaluating the loss of utilities, helping the team anticipate issues and establishing procedures for a safe shut-down and a safe restart. At some point during normal operations, for example, there could be a sudden, unexpected power outage that could shut down the entire facility. For this reason, facilities have developed safeguards to ensure that there is a reliable backup emergency electrical supply that can be used to safely shutdown the affected equipment. A useful checklist for other potential utility losses that could jeopardize the engineering controls required to manage the proc esses safely is provided in Table A.3-1.

74 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS Instrumentation may require different calibration to account for differences in physical properties such as density (See Example Incident 3.3, Texas City) A thorough risk-assessment process, such as a “What-If” PHA should identify these issues, and adequate procedures/ checklists must be put in place. Example Incident 3.15 and Example Incident 3.16 illustrate what can go wrong when operating procedures and design alarm issues are not aligned. Example Incident 3.15 – Batch Reaction Alarms Ignored A polymer plant comprised 12 batch re actors, each with a cycle time of about 8 hours. It was crucial to keep the agitator running, in order to control the exothermic reaction, so a high (top) priority alarm was in place to warn operators if the agitator stopped. However, at the end of every batch, the agitator was turned off by the process control computer, which activated the same high priority alarm, even though it was not re quired under these circumstances. With this alarm sounding about ev ery 40 minutes, the operators soon began to ignore it. Despite report ing the matter to management, the issue was not resolved. Operators go t so weary of hearing the regular sound of the very loud alarm with no volume control, that they wrapped it in several layers of bubble-wrap to silence it. Therefore, the likelihood of a rapid and appropriate response to a “stirrer trip” alarm during the reaction phase was drastically diminished.

7.2 Sustainability of Process Safety Culture |247 m onitored by their supervisors to ensure they are working in accordance with these expectations. Additionally, leaders and supervisors should be alert to new employees being indoctrinated by co-workers in ways that conflict with core cultural expectations. Awareness of this possibility should be a checkpoint for the cultural snapshots mentioned above. Continue Professional Development Learning to assess and advance the culture applies to individual workers as well as the com pany as a whole. Professional development brings in new skills and stimulates novel ideas for improving process safety. It also helps “recharge the batteries” so workers do not feel they are just going through the notions. Professional development goes beyond receiving training. Reading, attending conferences, and m aking presentations are also useful. Additionally, asking em ployees to deliver training is an excellent way to learn even deeper what they already know. CCPS, and other organizations provide m any opportunities to attend conferences, receive training, publish and read articles, and establish peer networks. Reassess periodically Reassessment differs from the snapshots mentioned above in the depth of the assessment. Chapter 6 addressed the culture assessment process and how to determ ine the frequency of assessments. Additionally, culture status can be assessed during regular process safety audits and by investigating trends of process safety leading indicators. These may not reveal the root causes of any problems, but can trigger deeper investigation. As of this writing, only Contra Costa County, California, USA has a statutory requirement to periodically assess process safety culture. The likelihood of other jurisdictions taking up such a

244 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION mitigation measures in the design to reduce the associated risk. Process hazards come from many sources, including the following: Material and chemistry used (e.g., flammability, toxicity, reactivity) Process variables - the way the chemistry works in the process (e.g., pressure, temperature, concentration) Equipment failures Recall from Chapter 1 that a hazard is a chemical or physical characteristic that has the potential for harming people, property, or the environment. Hazard identification involves analyzing a process and thinking about what scenario could exist where a hazard might result in undesired consequences. Specif ically, hazard analyses are used to identify weaknesses in design and operation of facilities that could le ad to a hazardous material release (loss of containment). This Chapter introduces a variety of hazard an alysis techniques that can be used during various stages of the design and during the op eration of a facility. Many regulations require process hazard analysis be conducted for new fa cilities and revalidated, in the case of OSHA PSM every five years. (OSHA) The revalidation may take the form of redoing the study, confirming the previous study is valid and up to date, or a blend of these approaches. Further guidance on revalidation studies is provided in Guidelines for Revalidating Process hazard analysis, 1st edition. (CCPS 2001) The approach of confir ming the continued validity of the previous study might include the following steps. evaluate that the previous hazard identification analysis used an appropriate analysis method evaluate that the previous hazard identifi cation analysis was accurate and complete determine if the previous hazard identifica tion analysis has been updated to reflect changes that went through the management of change process ensure that process safety information is current verify that action items from the previous hazard identification analysis have been resolved ensure that learnings from incident investigations have been implemented document the revalidation Process hazard analysis (PHA) - A n o r g a n i z e d e f f o r t t o i d e n t i f y a n d evaluate hazards associated with processes and operations to enable their control. This review normally involves the use of qualitative techniques to identify and assess the significance of hazards. Conclusions and appropriate recommendations are developed. Occasionally, quantitative methods are used to help prioritized risk reduction. (CCPS Glossary)

92 PROCESS SAFETY IN UPSTREAM OIL & GAS Figure 5-2. Source term pathways to ultimate consequences (IChemE, 1996) buoyant. A greater hazard is associated with the tr eatment process where the regeneration of rich absorbent liquid produces a nearly pure H 2S stream. This is immediately sent to a sulf ur plant that converts H 2S to pure sulfur or injects the acid gas back into the ground. Some facilities use double pipe for added safety for this connection. The toxic criteria used by the EPA for assessing hazardous facilities is ERPG- 2 (Emergency Response Planning Guideline – Level 2), also known as AEGL (Acute Exposure Guideline Levels). This is the concentration that most people can be exposed to for one hour without developing life threatening symptoms. The ERPG-2 for H 2S is 30 ppm. H 2S leaks initially have an associated rotten egg odor, but after a short time olfactory fatigue causes a loss of the sense of smell and people can be exposed to toxic concentrations with out recognizing it. This is a particular hazard to personnel who are close to th e source and the neighboring community since toxic clouds can trav el significant distances.

1 1 OVERVIEW OF THE PHA REVALIDATION PROCESS A process hazard analysis (PHA) is found ational in helping facility management implement and maintain all four of the accident prevention pillars identified within the Guidelines for Risk Based Process Safety (RBPS) [3, p. 3]. It may also be required to comply with applicable pr ocess safety regulations and internal company requirements. The pillars are listed here and discussed in Section 1.8: • Commit to process safety • Understand hazards and risk • Manage risk • Learn from experience Once management has committed the or ganization to pr ocess safety, the next step is to understand what hazards need to be managed. In 2008, the Center for Chemical Process Safety (CCPS) updated and republished its Guidelines for Hazard Evaluation Procedures book, which includes the following definition of a hazard [2, p. 51]: A hazard is a physical or chemic al characteristic that has the potential for causing harm to people, property, or the environment. Thus, hazard identifi cation involves two key tasks: (1) identification of specific undesirable consequences and (2) identification of material , system, process, and plant characteristics that could produce those consequences. The RBPS term “Hazard Identification and Risk Analysis (HIRA)” encompasses the application of a broad range of analytical tools, including those used in a PHA to identify hazards and evaluate risk. A PHA report documents the results from a particular application of HIRA tools inte nded to meet specific requirements for managing risk in a process. These requir ements can be internal (e.g., company policy) and/or external (e.g., regulato ry). Over the life of a process, these requirements and their interpretation may change. There may also be external changes, such as community development or rainfall patterns that affect risk. The company’s experience with and understanding of the process will increase,

NOTIFICATION , CLASSIFICATION & INVESTIGATION 81 5.2 INCIDENT CLASSIFICATION Classifying incidents can assist decision-making regarding their management and investigation. Classifica tion systems can vary depending on the company and the site organization. There is no perfect one-size-fits-all system of classification. Traditionally, cl assification systems assign a category to an incident based on the type of incident or its actual (or potential) severity. In some cases, it may be usef ul to assign a cat egory based on the nature and complexity of the incident (rather than only its severity) to facilitate the selection of lead investigators and team members with the most appropriate skill sets. In a few cases, the local jurisdiction may mandate a specific approach to incident classifi cation as well as the depth of the investigation. Table 5.1 shows variou s incident classification schemes. The incident classification syst em selected should preferably: Be easily understood, Include clear examples, Detail specific mechanisms to authorize an investigation and who may do so, Help identify the investigat ion approach/methodology, and Help determine the composition of the incident investigation team. In practice, whatever method is used, there may be gray areas in every system. Discovery of new information or changes in perspective during the initial stages of an investigation may lead the team or site management to change the incident classification during the course of the investigation. For example, the team investigating an in cident may determine that an actual (or potential) consequence was more severe than first recognized. The management system sh ould provide guidance on how to make changes in incident classifications when appropriate.

19. Communicating Information and Instructions 243 19.6.2 Potential causes of handover error The main type of error that can occur during shift handover is the omission of information, such as: • An informal handover process relying on improvised notes can create the potential to omit cr itical information. Recording may also be unclear. • Failure to communicate the state of the process. For example, equipment faults, product levels within vessels, or the point being reached within a long start-up process. • Failure to communicate the status of isolations, permitted work, temporary workarounds, or overrides. • Failure to communicate abnormal events in the previous shift that may impact operations in the next shift. • Failure to communicate maintenance or contractor activities in the area. • Unnecessary information obscuring ot her more important information. • Unreliable methods of recording, such as poor handwriting. 19.6.3 Good Human Factors of shift handover It is common practice to us e formalized logs and shif t handover forms, either paper-based or electronic. The specific fields will be process specific. The elements of an effective handover are summarized in Table 19-4. In addition to a formal handover pr ocess, people should be trained in: • The importance of accurate handover. • Two-way communication skills. • An open and engaging culture. The handover process should include information such as the reasons for temporary bypasses, process state, and eq uipment faults. Good handover can also include a checklist, especially those that highlight how the operating state of the plant has changed. Failure to include relevant information in a clear and open way will result in a poor shift handover. Other failings that result in poor shift handovers can include not providing enough time on return to work situations or poorly selected areas away from process (e.g., in the control room creating distractions and providing on verbal cues alone). Good handover process should includ e handover between supervisors and between managers. An on-site formal or informal walk-through, is useful in handovers. In addition to communication of process state, supervisors and Incomplete or inaccurate handover can cause the oncoming shift team to lack awareness of process state and equipment condition, creating conditions for them to make mistakes.

401

154 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS 5.9.3 Pre-Startup Safety Review No matter the type of MOC, a Pre- Startup Safety Review (PSSR) should be performed. The purpose of the P SSR is to ensure that the MOC was fully reviewed, the change was communicated to affected personnel, training has been conducted as n eeded, and documentation and records of the changes have been updated. In summary, a variety of tools and methods should be employed as discussed in this chapter to help evaluate abnormal situations and manage their potential co nsequences. Incorporatin g these tools into the normal management prac tices of a facility is therefore recommended. Chapter 6 will examine and provid e guidance on how to measure and continuously improve the ma nagement system for abnormal situations at a facility.

E.41 Who Me? Yeah You. Coundn’t Be. Then Who? |333 Each driver had a different m ethod for measuring the amount removed during the vacuum operation and or detecting the B S& W to hydrocarbon interface. The official amount transported was determ ined only by the owner by level difference after the hauler departed the well site. The hauling com pany and owner both clearly understood that no hydrocarbon should be removed from the tank during the extraction operation, but no check of the extracted material was made to confirm this before transporting. On the day of the incident, investigators concluded that a significant am ount of hydrocarbon was unintentionally extracted. When the truck was being unloaded at the liquid waste injection site, hydrocarbon vapors from the tank were ignited, m ost likely from the idling truck engine. In the ensuing fire, the truck valve opened, draining additional B S& W and hydrocarbon to the unloading pad. This hydrocarbon form ed a pool fire that took nearly an hour to extinguish. The investigators (ref E.8) noted several m anagement system failures as well as regulatory gaps that contributed to the incident. The investigator further noted that the industry generally recognized B S& W as non-hazardous, and that while some in the industry recognized that hydrocarbon that could be present in extracted B S& W could be flamm able, the m ajority did not. This difference could sim ply one of terminology: “flamm ability” is defined as having a flashpoint below 100 oF while liquids with flashpoints not too far above that temperature m ight can burn and can still ignite readily, especially if warmed. Relying on regulatory definitions when they are not accurate, and denial of hazards are clear signs of a weak sense of vulnerability and a weak imperative for safety . What other culture issues m ight have existed in this situation? The well owner clearly empowered the hauler to verify the absence of hydrocarbon in the extracted B S& W, and this would seem to be a culture positive. Likewise, the waste injector

47 3.2 REACTORS Reactors can represent a large portion of the risk in a chemical process. A complete understanding of reacti on mechanism and kinetics is essential to the optimal design of a reactor system. This includes the chemical reactions and mechanisms, as well as phys ical factors, such as mass transfer, heat transfer, and mixi ng. A reactor may be large because t h e c h e m i c a l r e a c t i o n i s s l o w . H o w e v e r , i n m a n y c a s e s , t h e c h e m i c a l reaction occurs very quickly, but it appears to be slow due to inadequate mixing and contacting of the reactant s. Innovative reactor designs that improve mixing may result in much smaller reactors. Such designs are usually cheaper to build and operat e, as well as being safer due to smaller inventory. In many cases, improved product quality and yield also result from better and more unif orm contacting of reactants. With a thorough understanding of the re action, the designer can identify reactor configurations that maximize yield and minimize size, resulting in a more economical process that generates fewer by-products and waste, and increases inherent safety by reducing the reactor size and inventories of all materials. A relatively new development in reactor design is the spinning disk reactor. In this novel design, the reactions take place in an imposed acceleration field, in this case, centrifugal motion. The fluid acceleration greatly enhances both the mass and heat transfer processes, thereby allowing the same reaction rates to occur in a much smaller volume. The rotating surface of revolution (i.e., the spinning disc) creates an ideal environment for the rapid transmissi on of mass, heat, and momentum because the thin liquid films generate d on the disc are highly sheared. This facilitates rapid physical or chemical processes involving liquids (even viscous liquids), such as poly merization, precipitation, and rapid exothermic organic reactions. Fo r example, the manufacture of a pharmaceutical product may require a 2000-liter conventional stirred batch reactor vessel. Using a 30-cm disc reactor 1000 tons/year of the same pharmaceutical product can be produced at a continuous rate of 30 gram/sec. (Ref 3.20 Stankiewicz). Another recent development in re action technology is the micro reaction. The miniaturization of plant mixing, and heat transfer equipment can generate very high mass and heat transfer rates. Firstly, the gradients driving mass and heat transfer, i.e., concentration and

55 transportation sector. The existing bulk storage tank and its supports were also in need of replacemen t due to corrosion. Therefore, the proposed inherently safer modification was economically viable as well. Similarly, hazardous raw material st orage should also be minimized, with greater attention being given to “just in time” supply. Inventory reduction lowers inventory costs, while increasing inherent safety. However, in determining appropriate raw material inventories, the entire supply chain should be considered, as follows: Will the originating facility for the materials, distribution facilities in the value chain, or both ha ve to increase inventories to provide “just-in-time” service, and will this represent a greater risk than a larger inventory at the end-user facility? How much additional burden will “just-in-time” delivery place on operating staff? Will increased nu mber of “just-in-time” deliveries increase the potential for human errors in loading/offloading operations? Will the additional time working in transient operating modes, due to unplanned shutdowns and th e resulting restarts caused by running out of raw materials, increase the risks? Will transportation and temporary storage of raw material, either in parked railroad cars, tank trucks, barges, or other transportation containers along transportation routes or in transportation facilities, present perhaps an even greater risk than on-site storage in a well-designed end-user facility? Does the increased number of deliveries increase the risk in the mobile portions of the value ch ain? Typically, increased rail, truck, or barge shipments increa ses exposure of populations, property, and the environment al ong the transportation routes to potential loss of containmen t events along these routes. Chapter 8 discusses inherently safer options in hazardous materials transportation in more detail. The reduction in inventory resulting from greater attention to plant operations and design of unit intera ctions can be substantial. Several excellent examples are provided here: An acrylonitrile plant eliminat ed 500,000 pounds of in-process s t o r a g e o f h y d r o g e n c y a n i d e b y s h u t t i n g d o w n a n e n t i r e u n i t

88 If a chemical process requires the concentrated form of the material, it may be feasible to store a more dilute form, and th en concentrate the material by distillation or some ot her technique in the plant prior to introduction to the process. The in ventory of material with greater intrinsic hazard (i.e., undiluted) is reduced to the minimum amount required to operate the process, ho wever the tradeoff of including a distillation step may add a new hazardous process. Materials that boil below normal ambient temperature are often stored in pressurized systems under their vapor pressure at the ambient temperature. The pressure in such a storage system can be lowered by diluting the material with a higher bo iling solvent. This reduces the static pressure imposed on the storage co ntainer, as well as the pressure difference between the storage syst em and the outside environment, thereby reducing the rate of release in case of a leak in the system. If there is a loss of containment inci dent, the atmospheric concentration of the hazardous material at th e spill location and the downwind atmospheric concentration and hazard zone are thereby reduced. For example, the use of modified HF wh ich has a lower vapor pressure, and commensurate reduced airborne exposu re hazard, has also come into more common use in recent years. Chemical reactions are sometimes co nducted in a dilute solution to moderate reaction rates, to provide a heat sink for an exothermic reaction, or to limit the maximum reaction temperature by tempering the reaction. In this exam ple, there are conflicting inherent safety goals - the solvent moderates the chemical reaction, but the dilute system will be dimensionally significantly larg er for a given production volume. Careful evaluation of all process risks is required to select the best overall approach. 5.2 REFRIGERATION Many hazardous materials, such as ammonia and chlorine, can be stored at or below their atmospheric bo iling points with refrigeration. Refrigerated storage reduces the ma gnitude of the consequences of a release from a hazardous material storage facility in three ways: 1.By reducing the storage pressure

286 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Table 13.7. Input and output for dense gas dispersion models Input (varies depending on model) Source Term Cloud mass or volume Temperature Concentration Gas density Cloud dimensions (height, width) Local information Wind speed Atmospheric stability Surface roughness Ground heat capacity, thermal conductivity Physical/Chemical Information Molecular weight Atmospheric boiling temperature Latent heat of vaporization LFL Toxic concentration or toxic dose To estimate cloud size or plume generation rate Hole size Release phase (gas, liquid, two-phase) Flash fraction Aerosol and rainout fractions Release duration Pool boiloff (from rainout fraction) Cloud initial dilution Cloud geometry Output (varies depending on the model): source term summary (if calculated by model): jet discharge or pool boiloff rate, temperature, aerosol fraction, rainout, initial density, initial cloud dimensions, time variance. cloud dispersion information: cloud radius and height (or other dimensions as appropriate), density, temperature, concentration, time history at a particular location, concentrations and width at to specified distances. special information: terrain effects, chemical reaction or deposition, toxic load at particular location, mass of flammable material in cloud. Consequence Effect Modeling Many types of consequences are possible for a release including fires, explosions, and toxic cloud dispersion. The potential types of fire and explosions are described in Chapter 4 and summarized in Table 13.8. Table 13.8. Types of fires and explosions Fire Pool Fire Flash Fire Jet Fire Explosion Physical: vessel ruptur e, BLEVE and fireball, rapid phase transition Chemical: thermal or runaway reaction, propagating reaction (confined and unconfined vapor cloud)

68 Guidelines for Revalidating a Process Hazard Analysis 4.1 OPERATING EXPERIENCE INFLUENCE ON REVALIDATION How could operating experience affect the approach to the revalidation? Consider two different situations involving a simple ammonia storage tank at a shipping terminal. In Case 1, the tank had been routinely storing ammonia for years, as designed. There had been no changes, no incidents, and no out-of- specification results or unexpected corrosion mechanisms detected during any of its scheduled tests or inspections. In Case 2, the same tank had been modified several times in response to multiple leaks and overfill incidents, and most inspections and preventive maintenance had been deferred due to financial pressure. In Case 2, the risk judgments made by the previous PHA team should be re-evaluated in light of (1) the multiple incidents experienced in the process and (2) changes in maintenance practi ces that may have allowed the basic integrity of the tank to degrade or rend ered critical safeguards ineffective. Therefore, in Case 2, consideration of the operating experience favors the Redo approach. Conversely, in Case 1, nothin g in the operational experience warrants Redoing the PHA, so the Update approach should be satisfactory. 4.2 TYPES OF OPERATING EXPERI ENCE THAT SHOULD BE CONSIDERED All operational experience should be considered in preparation for a revalidation, including operational experi ence with other similar units at the same site, at other sites, and elsewher e in industry, if known. Operational experience broadly includes startup an d shutdown activities and any special operating activities (e.g., regeneration of catalyst beds, cleaning between batches). The experiences from unit tu rnarounds and maintenance activities should be considered as well. In a few cases, prior operational experi ence may not be relevant and can be excluded. For example, a ba tch reactor may have once been multi-purpose, but is now in dedicated service because of increased customer demand for one of those products. Operational experience with other product lines may have minimal or no relevance to the current revalidation. The following sections discuss several of the most pertinent sources of operational experience. While specif ics regarding the selection of PHA revalidation approach ( Redo or Update ) are contained in Chapter 5, the sections below are written assuming the default approach to the revalidation is to Update the PHA. Wherever possible, the prior PH A will be used as the baseline for the revalidation, and it will be Updated to incorporate changes and experience

Figure 1-2: Overview of the handbook, by chapter Procedures and job aids 5. Human performance and job aids 6. Selecting a type of job aid 7. Developing content of a job aid 8. Format and design of job aids Operational competence 10. Human performance and operational competency 11. Determining operational competency requirements 12. Identifying learning requirements 13. Operational competency development 14. Operational competency assessment Recognizing and learning from performance 25. Indicators of human performance 26. Learning from error and human performance Task support 15. Fatigue and staffing levels 16. Task planning and error assessment 17. Error management in task planning , preparation, and control 18. Capturing, challenging, and correcting operational error 19. Communicating information and instructions Non -technical skills 20. Situation awareness and agile thinking 21. Fostering situation awareness and agile thinking 22. Human Factors in emergencies 9. Human Factors in Equipment design Concepts, principles, and foundational knowledge 2. Human performance and error 3. Options for support ing human performance 4. Supporting human capabilities

2. The Concept of Inherent Safety 2.1 INHERENT SAFETY AND PROCESS RISK MANAGEMENT The modern design and implementati on of chemical process safety programs incorporates risk-based a pproaches. This includes recognition of the hazards posed by the process, and a continual effort to analyze the risks, and to reduce or control them to the lowest levels practical, while balancing other objectives. A ha zard is defined as “An inherent chemical or physical characteristic that has the potential for causing damage to people, property, or the environment” (Ref 2.12 CCPS Glossary). Risk is defined as “A measure of human health effects, environmental damage, or economic lo ss in terms of both the incident likelihood and the magnitude of the lo ss or injury. A simplified version of this relationship expresses risk as the product of the likelihood and the consequences (i.e., Risk = Consequence x Likelihood) of an incident” (Ref. 2.12 CCPS Glossary). In other words, risk is a function of both consequence (hazard) and likelihood (frequency). Safety can be defined as a condit ion of tolerable risk compared to the benefit of the activity. CCPS fo rmally defines safety as: “The expectation that a system does not, under defined condit ions, lead to a state in which human life, economic s or environment are endangered” (Ref 2.12 CCPS Glossary). This cons iders who receives the benefit when compared to who bears the risk. Inherent safety is one of the tools available to improve safety; it is a preferable method, if feasible, but other approaches are valid and succ essful depending on the specific situation. Inherent safety is used wh ere it meets overall safety objectives, where it is sensible given various alternatives to manage risks, and where there is an application for th e inherent safety principles. In general, the application of inherent safety principles has reached a higher level of maturity in situatio ns where high hazards exist and the risk is also high, i.e., in chemical proc ess safety and in the security of such facilities. 12 (VJEFMJOFT
GPS
*OIFSFOUMZ
4BGFS
$IFNJDBM
1SPDFTTFT
"
-JGF
$ZDMF
"QQSPBDI #Z
$$14 ¥

5IF
"NFSJDBO
*OTUJUVUF
PG
$IFNJDBM
&OHJOFFST

155 8 DOCUMENTING AND FOLLOWING UP ON A PHA REVALIDATION A clear, concise, thorough PHA revalidatio n report is essential to the retention and communication of the PHA results, as well as being needed to demonstrate compliance with internal and external PHA requirements. Because PHA is the PSM element through which an organiza tion identifies process hazards and appropriate risk controls, a revalidated PH A report provides an ongoing basis for many other PSM program elements. For ex ample, it identifies the engineered controls that must be included in the mechanical integrity (MI) program for inspection, test, and preventive maintenanc e (ITPM) to maintain the required probability of failure on demand; it identi fies the loss scenarios that must be addressed in the emergency response plan; and it identifies the consequences of deviations that must be addressed in the standard operating procedures. It is also the basis for future management of change (MOC), incident investigation, and revalidation activities, and it is su bject to periodic audits that verify compliance with regulatory and/or internal company PSM requirements. Experience shows that inadequate documentation of the prior PHA is one of the most frustrating issues for the future revalidation team because (1) it causes unnecessary work that consumes additional time and distracts the team from its goal of evaluating hazards and risks and (2) it may require them to Redo a PHA that could otherwise have been Updated . It is likely that the revalidation team will gain, during the course of the PHA revalidation, a greater appreciation of the importance of complete, accurate documentation and document their own work accordingly. Much of the documentation for a PH A revalidation parallels the typical documentation for an initial PHA (e.g., including a current set of drawings showing the node definitions). However, some approaches to revalidation, as suggested in this book, use a number of screening forms and checklists. Those choosing to use forms and checklists (s uch as those in Appendices A and B) should consider including them in the revalidation documentation so that the basis for key decisions (e.g., the basis for the particular revalidation approach chosen) can be clearly communicated to the next revalidation team.

210 | 6 Where do you Start? While the former is helpful and m ay increase the value of the interview, the latter m ay be a clear indicator of a culture problem . The following basic process should be helpful in establishing a framework for the overall process and increasing the effectiveness of the interviewer’s on-site activities. The emphasis is placed on the interaction that develops between interviewer and interviewee rather than strictly on the mechanics of the interview process. Plan the Interviews. The interviewer should identify the personnel to be interviewed in advance, understand the goals of the interview, determ ine the interview questions, and consider how to m aximize the effectiveness of the interviews. Interviews with a selection personnel that span the spectrum of responsibility will be required during a process safety culture assessment. These include representatives of: Senior management including the senior-most, Middle management, The process safety manager and managers of the PSMS elem ents Front line supervisors; and Hourly personnel including operators, maintenance personnel, and others as appropriate. Front line supervisors and hourly personnel should be selected from each of the facility’s shifts. As m uch as possible, set a specific time and duration for each interview and respect the interviewee’s other com mitments and work schedule. Request that the facility provide coverage for operating staff in safety-critical position, and generally limit interviews with operators to 30–45 m inutes to minimize disrupting operations. Arrange a comfortable setting for interviews. Hourly personnel will generally feel more comfortable in their own working environm ent and m ay feel subtly intim idated in • • • • •

18. Capturing, challenging and correcting operational error 223 Figure 18-5: Cognitive skills requ ired for error self-management Attention and vigilance Information gathering and search Plan formulation Problem diagnosis Systematic decision-making Self-monitoring Systematic scans and checks Divergence detection Information management Planning and mental stimulation Monitoring and evaluation

346 | Appendix F Process Safety Culture Assessment Protocol other training and inform ation that show that process safety is a core value? 16. Are process safety performance goals, objectives, and expectations included in performance contracts, em ployee goals and objectives, and discretionary com pensation arrangements for line m anagers, supervisors, and workers? 17. Are the metrics or other means by which process safety perform ance is m easured defined? 18. Do personnel report a pressure to m aintain performance standards, potentially at the cost of safety? 19. Are there comm itm ents to achieving performance goals that are greater than demonstrated for process safety goals? 20. Do operational pressures lead to cutting corners where process safety is concerned? 21. Is process safety improvement a long-term comm itm ent that is not com prom ised by short-term financial goals? 22. Is there sufficient staff in relevant work groups (e.g., operations, inspection, or maintenance) to perform jobs safely? 23. Is the organization is preoccupied with safety and process safety, such that they can anticipate areas of potential failure and can cope and bounce back from errors when they occur? Do they exhibit a resilient nature? Resilience is defined as the ability of systems to survive and return to norm al operation despite challenges. 24. Is process safety m anagement an independent function in the organization? Does the m ain person responsible for process safety report to those who m ight have a conflict of interest with respect to decisions about the process safety impact on operations? Note: In smaller organizations this independence m ay be more difficult to achieve. 25. Are process safety resources are among the first budget line cuts during times of financial difficulty? 26. Is the process safety staff placed in the untenable position of having to prove that an operation is unsafe? Are those desiring

160 INVESTIGATING PROCESS SAFETY INCIDENTS methodologies outlined below would be necessary for a significant incident, or for an incident where multiple parties are involved and litigation is likely to take place at some stage in the future. Once the site has been inspected and its post-incident condition has been recorded and photographed, the ne xt stage for the investigation team is to conduct a more det ailed examination of the physical evidence. Documenting a list of parts, samples, and other physical data that are collected during the investigation, with each part tagged, numbered and/ or permanently marked (where this does not damage evidence) helps prevent mishandling or disposal of the items. Color-coding via tags or paint can be helpful to those engaged in moving or removing debris. One method is to have the demolition crew move only material that has been clearly marked. The guiding rule is: if it is inside the investigation zone and it is not marked, then it is to be left alone. Long, intermittent runs of piping should be marked at regular intervals, especially wher e the piping passes across the boundary of the investigation zone. Tag attachment should be robust and secure, such as plastic tie-wrap type devices. It is a good practice to photograph the item prior to and after attaching the tag to collected items and to log each of the tags. Some evidence will be highly mobile (e.g., small parts of valves and instruments, personal protective equipment and tools belonging to injured workers). Other items will be perishable (e.g., residual liquid and residue inventories for example) and will require careful handling under the guidance of a written protocol. Electronic data may be difficult to download but is easier to duplicate. A good practice is to bring a large capacity storage device such as a solid-state hard driv e to use as a “master” storage device for use by all team members, and which is backed-up on a daily basis. Access to electronic data should be restricted if there is potential for litigation. It is important to set up a numbering system that can be applied to a variety of types of physical and documentary data, such as that shown in Table 8.5.

416 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Figure 20.4. Overview of damaged WFC (CSB 2013) Figure 20.5. Apartment complex damage (CSB 2013)

Piping and Instrumentation Diagram Development 110 Generally the plug route of multi‐port valves is not shown; however, if the multi‐port valve is used in a criti­ cal application, it should be mentioned as a note beside the valve symbol or in the notes area of the P&ID. An example of using four‐port diverting valve is shown in Figure 7.6. It is shown in the figure that this arrangement reverses the flow of cooling water to the heat exchanger. The operator can reverse the flow every few weeks to reverse the flow of the tube sides to remove fouling from the tube internals. Probably using a multi‐port valve in this application is the best because of the following: ●The service fluid (i.e. cooling water) is a non‐dirty service to clog the valve. ●The reverse flow practice initiated every few weeks is not frequent and does not justify the use of several valves instead of one multi‐port valve. The multi‐port valves are available for throttling pur ­ poses as well. The throttling‐type multi‐port valves are generally globe types. Multi‐port throttling valves can be used for combining (mixing) or diverting purposes (Figure 7.7). One example of using multi‐port throttling valves is shown in Figure 7.8. In this application the valve is used to adjust the flow of two streams that are combin­ing. This valve can be replaced with two conventional valves that work in a parallel control system. The parallel control system will be discussed in Chapter  13. Some designers may decide to use only one conventional control valve on the bypass stream. This is acceptable because between two streams, the bypass stream is the lower resistance route and more flow goes through it. Therefore, a single control valve on that stream satisfies the control goal. 7.4.4 Double‐Sea ted Valves In P&IDs double‐seated valves can be shown as a spe­ cialty item and may use a special symbol. In reality double‐seated valves are distinguished from their respective conventional single‐seated valves because of their bulkier body. In double‐seated valves the stream is split into two streams and then goes through a dedi­cated seat for each stream. The main reason for using double‐seated valves is to reduce the torque required to open or close the valve. A double‐seated valve acting as a control valve needs a smaller pneumatic actuator in comparison with its respective control valve. Double‐seated valves, however, have an inherent prob­ lem, which is their passing‐by. Because of the complexity in manufacturing double‐seated valves, they almost always suffer from internal leaks or passing‐by. Therefore double‐seated valves rarely produce a TSO. Double‐seated valves can mainly be used in control valves on high‐pressure streams when there is not enough room for a large pneumatic diaphragm (e.g. in debottleneck projects) as long as internal leaks are not a problem. 7.5 Valve Operators There are two groups of valve operators: manual and automatic. Manual valve operators are the valves that can be field adjusted by an operator, whereas automatic oper ­ ators that are installed on ROT are the valves that are controlled remotely from the control room. Automatic operators are also known as actuators. The type of valve operator, either manual or automatic, totally changes the way it is handled in a process plant. CWR CWS CWR CWS Figure 7.6 Applica tion of four‐way blocking valve on cooling water heat exchanger. Combining Diverting Figure 7.7 Thr ottling three‐way valve.TCCombining Figure 7.8 Applica tion of three‐way throttling valve for heat exchanger control.

6. Selecting a type of job aid 57 6.2.2.2 Task safety criticality The safety criticality of a task can be a ssessed using knowledge of the task-related hazards. The results of Hazard Identificati on and Risk Analysis (HIRA) can be used to rate task risk. A common HIRA approach is to use a qualitative risk matrix to rate the risk from very low to very high. This ri sk matrix approach can be used to rate the risk of a task. If a HIRA has already been completed for a process, the results can be used directly. These risk ratings may be applied to the Task Criticality in the flow chart previously shown in Figure 6-1. The example matrix in Figure 6-1 uses three risk ratings – high, medium and low. HIRA may use a risk matrix, as in Figure 6-2. Figure 6-2 also gives a potential alignment of HIRA ratings to high, medium and low in Figure 6-1, with red cells being high safety criticality, yellow being medium and green being low. Figure 6-2: Using HIRA risk matrix re sults to assess task safety criticality Likelihood Very high High High Moderate Medium Low Very low Low Very low Low Moderate High Very high Consequence Some organizations perform “Safety Critical Task Analysis”. An example is given in Figure 6-3 (this is a new example). This involves identifying safety critical tasks (i.e., those tasks that, if done unsuccessfu lly, will result in a process safety event), one by one, assessing them and deciding what needs to be done to support successful task performance. The example in Figure 6-3 includes identifying “Failure types” using the “mistakes, s lips and lapses” categories, and then identifying existing and additional “contr ols”. Task Criticalit y may be rated using five factors, with ratings color coded red (high) or green (low) in this example, giving a high score of 10. Guidance on Safety Critical Task Analysis is contained in the Energy Institute guide [29].

GLOSSARY xxxvii Probabilistic Risk Assessment (PRA) A commonly used term in the nuclear industry to describe the quantitative evaluation of risk using probability theory. Probability The expression for the likelihood of occurrence of an event or an event sequence during an interval of time, or the likelihood of success or failure of an event on test or on dema nd. Probability is expressed as a dimensionless number ranging from 0 to 1. Process Flow Diagram (PFD) A diagram that shows the material flow from one piece of equipment to the other in a process. It usually provides information about the pressure, temperature, composition, and flow rate of the various streams, heat duties of exchangers, and other such information pertaining to understanding and conceptualizing the process. Process Hazard Analysis (PHA) An organized effort to identify and evaluate hazards associated with processes and operations to enable th eir control. This review normally involves the use of qualitative tec hniques to identify and assess the significance of hazards. Conclusions and appropriate recommendations are developed. Occasionally, quantitative methods are used to help prioritize risk reduction. Process knowledge management A management system element that in cludes work activities to gather, organize, maintain, and provide in formation to other management system elements. Process safety know ledge primarily consists of written documents such as hazard informatio n, process technology information, and equipment-specific information. Process safety knowledge is the product of this management system. Process safety A disciplined framework for managing the integrity of operating systems and processes handling hazardous subs tances by applying good design principles, engineering, and operating practices. Note: Process safety focuses on efforts to reduce process safety risks associated with processes handling hazardous materials and energies. Process safety efforts help reduce the frequency and consequences of potential incidents. These incidents include toxic or flammable material releases (loss events), resulting in toxi c effects, fires, or explosions. The incident impact includes harm to peop le (injuries, fatalities), harm to the environment, property damage, produc tion losses, and adverse business publicity. Process safety culture The common set of values, behaviors, and norms at all levels in a facility or in the wider organization that affect process safety. Process Safety Incident/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 health and en vironmental consequences.

YYYJJ INVESTIGATING PROCESS SAFETY INCIDENTS PL Protection Layer PLC Programmable Logic Controller PM Preventive Maintenance PPE Personal Protective Equipment PSHH Pressure Sensor High High PSI Process Safety Information PSID Process Safety Incident Database PSM Process Safety Management PSM also Canada’s (non-regulatory) standard, individualized by district PSV Pressure Safety Valve (Relief Valve) R Risk RCA Root Cause Analysis RIDDOR Reporting of Injuries , Diseases and Dangerous Occurrence Regulations RMP Risk Management Program (US) RQ Release Quantity RV Relief Valve SAWS China’s regulatory guideline for incident JOWFTUJHBUJPOT SCAT Systematic Cause Analysis Technique SCE Safety Critical Equipment SDS Safety Data Sheets SEMS Safety and Environmental Management System SHE Safety Health & Environment SIF Safety Instrumented Function SIS Safety Instrumented System SMART Specific, Measureable, Agreed/Attainable, and Realistic/Relevant, with Timescales SOL Safe Operating Limit SOP Standard Operating Procedure SOURCE Seeking Out th e Underlying Root Causes of Events SRK Skills, Rules, Knowledge SSDC System Safety Development Center STEP Sequentially Timed Events Plot T Test Interval for the Component or System (hours or ZFBST T 0 starting time Tn ending time PIF Performance Influencing Factor

INVESTIGATION M ANAGEM ENT SYSTEM 65 • Priority preservation of raw ( uncompressed, unaltered buffer) data on an expedited basis, i.e., be fore memory capacity causes overwriting data or averaging data to a historian archive • Preservation of data related to operator control input and associated control element movements • Preservation of data logs, e. g., alarm, programmable logic controller action, safety instrumented system functioning, set point excursions, etc. Specific electronic evidence identification and preservation suggestions are contained in Chapter 8. 4.2.6 Defining Training Requirements M anagement proves its commitment by action. Management committed to learning from incidents will establish a high-quality incident investigation training program. This helps to ensure that the management system is understood and implemented as des igned. Each job position’s training on the incident investigation system will vary in the level of detail and scope. Persons assigned to lead roles on incident investigation teams should be targeted to receive the most concentrated training. Periodic refresher training is an opportunity for management to reinforce commitment, demonstrate support for the organization’s policy and philosophy on incident reporting and investigation, and discuss modifications and improvements in the investigation process based on lessons learned from performing investigations. Typical training agendas for ma nagement and employees who may report an incident but are not intended to be designated investigative team members, can be brief. Special training may be indicated for those employees and functions that will interface with the incident investigation team during an investigation. These may include, for ex ample, emergency response teams, fire brigade, maintenance, security, site safety, site industrial hygiene, public relations, legal, and environmental. Table 4.1 describes general guidelines for the content of training session s for various functions.

Plant Process Control 323 Temperature Control of Reactors Heat transfer doesn’t necessarily happen only in heat exchangers and then temperature control. In reactors, temperature control may be important if the reaction is  largely exothermic (heat‐emitting) or endothermic (heat‐absorbing). In the below schematic examples of reactor tempera- ture control are shown. In some of the examples below, the temperature of the reactor is controlled using a jacket around the reactor, while in the other examples, heating and/or cooling coil(s) are used inside the reactor. In this example, we have a reactor with an exothermic reaction. The reactor has a jacket for cooling water, Figure 15.59. We want to control the reactor temperature, so we have a temperature sensor on the product line, which is connected to a loop to control a valve on the cooling water inlet stream.However, as we know, the response time for a tempera- ture loop is slow, especially in this case, since it involves the temperature of a bulk fluid. This arrangement could be acceptable if the speed of the loop is not an issue, for exam-ple if the reaction is very slow. In the majority of practical cases, the reaction speed (reaction kinetics) is so high that such a simple arrangement wouldn’t provide good control. We can speed up the response time by using another, faster temperature loop to act as a slave for this loop. The temperature loop of jacket water is faster because the volume of water in the jacket is much lower than the fluid in the reactor (a layer of water versus a bulk fluid). In this improved arrangement, the temperature loop on the product line acts as a master to provide a RSP for the tem-perature controller of the slave loop. In effect, we have a temperature‐to‐temperature cascade control system. Instead of having a jacketed reactor, we can replace this arrangement with the one shown in the bottom right of Figure 15.60. To control the temperature, we divert a TTTTTC TCRSP Jacket, Coil or RecirculationQuickSlow! Figure 15.59 Reac tor control – jacketed reactor. TC TT TEPDC Figure 15.58 Heat e xchanger bypass control with a PDC. TC TCRSP Figure 15.60 Reac tor control – external heat exchanger.

158 | 5 Aligning Culture with PSMS Elements Figure 5.1 Risk B ased Process Safety M anagement System Source: D. Guss, Nexen, Inc. "O gg m 3::: Cl'I 3::: Cl'I _ ~::: ill 0 ~ ,.~i z 0 0 !ll ::0 ::0 Cl'I iii O )i! "Cl'I z ~ ?i C) m ::0 i 0 ~,... ;11i ::0 z iii :B ~o m 31: Process Safety Culture Compliance with Standards Process Safety Competency Worllforce Involvement Stakeholder Outreach Process Knowledge Management Hazard Identification and Risk Analyi;_is Operating Procedures Safe Work Practices Asset Integrity and Reliability Contractor Management Training and Performance Assurance Management of Change Operational Readiness Conduct of Operations Emergency Management Incident Investigation Measurement and Metrics Auditing Management Review and Continuous Improvement

Figure 6-5: Mapping of type of job aid to type of task performance Key: CK = Checklist. GC= Grab card. DFC = Diagnostic flow chart DT = Diagnostic tree Info = Process safety information Log = Log books etc. M = Manual PSB = Process Status Board PTW = Permit to work SH = Shift Handover SOP = Standard Operating Procedure WI = Work Instruction

Principles of P&ID Development 61 make the equipment needier for maintenance. These two components are discussed next. The equipment that are static generally need less maintenance. Among nonstatic equipment (i.e. dynamic equipment), the ones with linear (reciprocating) move-ments may need more maintenance attention than the ones with rotary movements (Figure 5.25). Where there is a rotating shaft in a piece of equipment, the high rotational speed shafts (high revolutions per minute [RPMs]) may need more maintenance attention than low RPM shafts. Pieces of equipment that have tight clearances may need more inspection and maintenance. This is especially true if they are being used in services that are not clean. When it comes to process and process conditions, the equipment that works in very high or very low tempera-tures or pressures may need more maintenance atten-tion. The equipment that processes non‐innocent fluids (i.e. highly acidic, precipitating, scaling, fouling, or any other aggressive fluid) may need more maintenance attention. 5.4.7 Oper ability in Absence of One Item The designer needs to decide the repercussions of equip- ment loss, which means in the absence of a piece of equipment, it needs to be decided what will happen to the rest of unit or plant. The wide range of answers and decisions include: 1) Do nothi ng! In this case, the piece of equipment, unit, or even plant should shut down in the absence of a piece of equipment or instrument. This option should be avoided. Sometimes it is inevitable when a piece of equipment of interest is the main or one of the main pieces of equipment of the plant. 2) Acc umulation of fluid in middle containers. In this solution, placing two containers with enough resi-dence times upstream and downstream of the absent component help to prevent the absence of the compo-nent get “visible” by the rest of plant. In this solution, the upstream container allows the accumulation of fluid, and the downstream container provides flow for the downstream units. 3) Re directing the in‐flow to a “reservoir” for later usage. In this solution, the feed to the equipment can be redirected to a temporary reservoir (like waste tank or pond) to be processed later by returning it back to the system. Usually this is solution is not avail-able for gases or vapors. 4) Re directing the in‐flow to an “ultimate disposal” system. This solution is the same as previous one, but the flow sent to the external reservoir cannot be returned. The feed to the equipment can redirected to a waste‐receiving system, like a flare system. This option can be considered if the preceding option is not doable. The previous option is definitely a better option because valuable materials are not lost. 5) Byp assing the absent item. The feed to the equip- ment can be bypassed temporarily with marginal impact on the operation of the system, like bypass - ing a trim heater if being off‐temperature does not hurt the plant for a short time. There are some cases that is decided to bypass the equipment or unit when it is out of operation. This can be done if the lack of equipment or unit does not affect the process in the short term. 6) The ne arly “similar” item in parallel. A nearly simi- lar system in parallel can take care of the flow that used to go to the absent system but not necessarily with the same quality. One example is having a man-ual throttling valve (e.g. globe valve) in a bypass loop of a control valve. The other example is placing a bypass pipe for a pressure safety valve (PSV) together with a pressure gauge (or pressure gauge point) and a globe valve. In the case of pulling the PSV out of oper - ation, an operator will act as a PSV by monitoring the pressure of the container and being prepared to open the valve if it is needed. 7) The e xact “similar” item in parallel. A parallel, exact replica as spare system can take care of the flow that used to go to the absent system. This is the most expensive option. The examples are all spare pumps or spare heat exchangers (in very fouling services). Spare equipment are very common for fluid‐moving equipment as usually the pumps and compressors cannot be handled otherwise. One important exam-ple is having two fire pumps in parallel with two dif - ferent types of drives (i.e. one electromotor and the other one a diesel drive pump). The spare can be in different forms. In Table  5.10, the schematics of these options in the P&ID are shown. 5.4.8 Pr ovision for the Future The other concept that may affect the development of the P&IDs are provisions for the future. The future arrangement of a plant is not necessarily similar to the current arrangement of plant because the future of a market is not always foreseeable, or if it is foreseeable, it is not economically justifiable to incorporate it into the current plant design. However, to minimize the cost of rearrangement of a plant in the future, some items can be placed in the plant design and the P&ID. Therefore, some “footprints” of future on a P&ID may be seen; how - ever, not all plants consider the future.

16.15 Tyler, B.J., Using the Mond Index to measure inherent hazards. Plant/Operations Progress 4 (3), 172-75, 1985. 16.16 Center for Chemical Process Safety (CCPS 2018). Guidelines for Siting and Layout of Facilities, Second Edition. New York: American Institute of Chemical Engineers, 2018. 441