Datasets:

hndc

/

AI4Hazard-small

like 0

14 Human Factors Handbook • Fatigue – “…the CSB concludes that fatigue of the operations personnel contributed to overfilling the tower” ( [14] p.289). Several key operational staff had worked between 29 and 37 12-hour shifts in a row. • Not enough staff – “…operator staffing levels below the numbers required for ‘safe staffing’. This involves the day-to-day operation of units with less than the minimum numbers of operators required…” ( [14], p.285). • Supervision – while the start-up shift started with two supervisors, the experienced supervisor left due to a family matter, leaving an inexperienced supervisor alone. The remaining supervisor was busy with several tasks. • Alarm flood – ISOM operators faced hundreds of alarms going off in a short time frame. They were not able to assess the situation or warn others. There were many factors influencing th e operational decisions and actions. Deficiencies in each of these factor s combined to exacerbate operational problems. 2.3.2 Contributing Human Factors “You cannot change the human condition, but you can change the conditions in which people work.” Professor James Reason (Chapter 7, page 96) [17]. A Human Factors principle is that errors and mistakes happen because of a combination of problems in the working environment and due to the support, or lack of it, offered by the organization. It is important to provide a working envi ronment (or set of conditions) that set people up to succeed throughout the lifespa n of an operating facility. Establishing this working environment / set of conditions should begin in the design of each new facility. Design should include proper Human Factors design in the layout of systems and equipment, and process hazard analyses and risk assessments must include consideration for Human Factors. Human error and mistakes are not the root cause of incidents. “…numerous latent conditions and safety system deficiencies at the refinery influenced their (operator) actions and contributed to the accident…” (CSB, 2007, [14] p.69)

138 | REAL Model Scenario: Leaking Hoses and Unexpected Impacts of Change Table 10.1 Results of Study Based on External Incidents (Continued) Study plan action Feijoada plant finding Determine potential for multiple materials to mix inside hoses. If found, test resistance to mixtures. After careful review of procedures for flushing hoses and compliance with the procedures, opportunities for mixing in hoses were not found. Determine if leaks or drips might be impacting hoses. None found. Evaluate where flexible hoses could be replaced with hard-pipe connections. Several opportunities were found, but none of the hoses in these services were leakers. “Maybe we just haven’t been asking the right questions,” Antônio said. “Let’s take one more walk up to the production floor.” As the four colleagues walked into the building, Juliana turned to João. “The lighting in here is so much brighter now since you installed the LED bulbs,” she said. “My mechanics really appreciate how much better they can see what they’re doing.” Just then, Antônio noticed Adriana, a new operator, standing on water feed piping to read a gauge. He opened his notebook to the page where he’d noted the issue about pipe hangers and showed it to his three colleagues. They walked over to Adriana. “Are you having trouble reading that gauge?” Antônio asked her. “Yes,” she replied. “Ever since the new lights were installed, we get a lot of glare off the glass. We have to get closer to read a lot of gauges.” “Does that include the gauges over the raw material manifolds?” She nodded. “Are you going to tell me we shouldn’t be standing on the hoses to read those gauges?” “Or the pipes,” João agreed. 10.6 Prepare On their way back in the conference room, João said, “Why don’t we have the operators figure out the best way to read the gauges without standing on the pipes or hoses?” “That makes sense,” said Juliana. “I’ll ask my lead mechanic Paulo to coordinate to make sure maintenance is not affected.

EVIDEN CE IDEN TIFICATION , COLLECTION & M ANAGEM ENT 149 Furthermore, consideration should be given to the preservation of fragile physical evidence such as cracks, deposits, chemicals and residues. 8.2.3 Paper Evidence and Data Although paper data is not always fragile, investigators should place a high priority on identifying, collecting, and preserving it. Often, the most difficult issues with paper data are locating the required document s and finding the relevant information within them. Analyzing paper data can be a very time- consuming process. Paper data in the form of operator logs, batch sheets and additions sheets or logs may be particularly important if reactive chemistry is suspected. These may hi ghlight the accidental mixing of incompatible materials, improper sequencing of addi tions, or improper addition rates or volumes. The size and scope of the investigation or other factors could mandate a special document control procedure, wherein each document is given a unique identification number. In this wa y, there is a documented chain of custody (e.g., what documents have been collected, the source of the documents, who has possessed the do cuments at any given time, etc.). Maintaining a complete, retained document set can help minimize confusion and a special log can be useful in ma intaining some degree of control over the flow of paper documents and in fi nding the answers to questions in the documents when they arise. This is especially important when legal issues or regulatory agencies are involved. Paper data from older instrumentation systems such as strip or circular chart recorders should be controlled immediatel y after the occurrence. Strip charts and disk recorders will not a ll turn at exactly the same rate, so checking the turn rate can be critical in comp aring the charts. The measurement range and units for each pen must also be asce rtained. For crucial charts, it may be necessary to perform a check of the calibration. If chart recorders are still operating, before removing the charts, mark and document each one with a time, then wait 30 minutes or an hour and mark again. Mark each item with the instrument number or name, the date, time of removal, and the last position of data recording. Make sure that replacem ent charts are re- installed after collecting the original on es; key data pertaining to subsequent occurrences related to the initial event can be lost if the charts are removed too early or not replenished after removal.

Piping and Instrumentation Diagram Development 348 The type of gas detection system depends on the types of gas to detect. In an FGS any aggressive gas can be decided to be measured. The gas of interest could be a flammable gas, toxic gas, or even inert but suffocating gases like nitrogen gas. The anatomy of the FGS is very similar to that of a SIS, except that the final element is not something that interferes with the process (like a switching valve); it is only an alarm. At the P&ID level, there are two aspects of FGS: its sig- nal handling system (control system) and the location of sensors and alarms. The signal handling system can be shown in auxiliary P&IDs. Figure 16.28 shows a typical FGS control system that can appear on an auxiliary P&ID. Figure 16.28 shows a flammable gas detection system that: ●Has a sensor (“ AE”), transmitter/indicator in field (“ AIT”), indicator in control room through PLC system (“ AI” in diamond). ●Alarms when the flammable gas concentration is 20%  of the LEL low explosion level); alarm‐analyte is high level. Alarm(a) (b) (c) (e)(d)TXH 237TAH 237TT 237TAH 237TT 237 TE 237 TAH 237TT 237 TE 237 XL 237TXH 237TAH 237TT 237 TE 237XH 237 TXH 237Horn Lamp RTAH 237TT 237 TE 237TE TE 237Logic Sensor Figure 16.25 Differ ent ways of showing one alarm system on a P&ID.

Evaluating Operating Experience Since the Prior PHA 71 during a defined outage, such as a turnaround, may be analyzed under one “umbrella” MOC, and the potential for complex interactions may similarly require significant analysis by the revalidation team. However, even simple changes can have far-reaching effects, for example: • The decision to introduce a recycle stream could result in unexpected corrosion or fouling in the equipment, or it could result in unexpected flow paths during an upset or shutdown. • A larger pump installed in one uni t might be capable of producing shutoff head pressures that exceed the design pressure ratings of equipment in downstream units. • An interconnecting pipeline might require protection against overpressure scenarios that could have their initiating event at either end of the reversible pipeline. • A liquid nitrogen carryover from the supplier’s facility unit might cause brittle fracture scenarios during purge activities. Batch formulation changes, utility system changes or process changes requiring several new pieces of equipment or extensive piping modifications might also introduce hazards well be yond the physical boundaries of the changes. Even simple changes may create a complex situation. When several small changes are considered collectively, there may be interactions that have a more significant impact on process risk s. Thus, the more complex the process changes, the more likely that the Redo approach is warranted. The degree to which the hazards associated with changes that have been made were analyzed and how they were analyzed. Continuing the previous example, assume the company did add a solvent recovery process and performed the hazard analysis for the MOC in accordance with their core PHA methodology. So rather than simply document the hazards of the change, they documented a comprehensive analysis of the hazards, risk controls, and consequences of their failure. In that ca se, the revalidation team can directly add the nodes in the project hazard review “mini-PHA” into the existing unit PHA as they closely review and Update the prior hazard analysis worksheets. That allows them to identify any oversights while not having to Redo that portion of the PHA. When MOC teams use the same core anal ysis technique as the PHA, meet the same analysis requirements (See Chapter 2.), and document them in the same format as a PHA, the Update approach can be used to efficiently accomplish the revalidation goals. Conversely, if the so lvent recovery MOC only considered the “hazards of the change,” the PHA revalidation team may need to Redo the PHA, at least for the new parts of the process.

REACTIVE CHEMICAL HAZARDS 87 Incident Investigation – If an operation yields an unexpected result, ask why did it do that? Through understanding why, you may recognize that you were heading down the path to an incident. More importantly, now you know how to avoid it. Detailed Description T2 Laboratories Inc. started in 1996 as a solv ent blending business founded by a chemical engineer and a chemist. One of their products was a blend of MCMT, a gasoline additive. In 2004 T2 began producing MCMT, which be came their primary product by 2007. The runaway reaction occurred during the fi rst step of the MCMT process. This was a reaction between methylcyclopentadiene (MCPD) dimer and sodium in diethylene glycol dimethyl ether (diglyme). MCPD and diglyme were charged to a 9.3 m3 (2,450 gal) reactor and sodium metal was added manually through a valve at the top of the reactor, (see Figure 5.3). Heat was applied to the reactor using hot oil set at 182 °C (360 °F) to melt the sodium and start the reaction to make methylcyclopentene. Hydrogen was a bypr oduct, vented through a pressure control valve. At 99 °C (210 °F) the agitator was started (by this time the sodium should have melted). At 149 °C (300 °F) the heat was turned off. The reaction was known to be exothermic and at 182 °C (360 °F) cooling was applied. After eliminating other possible causes, the CS B concluded that loss of cooling was the immediate cause of the runaway reaction. The reac tor was cooled by adding water to the jacket and allowing it to boil off (Figure 5.3). The cooling system, necessary to control th e exothermic reaction, could be totally incapacitated or severely impaired by several single failures: loss of cooling water from supply, a drain valve left open or partially open, failure of the valve actuators, blockage in the supply line, temperature sensor failure, or mine ral build up in the jacket. (CSB 2009) Without cooling, the temperature could continue to rise. Subsequent testing showed that a second exothermic reaction occurred at 199 °C (390 °F). This reaction was more energetic than the first, desired reaction. The owner/oper ators of T2 Laboratories did not know about this second reaction. This reaction generated enough pressure, very rapidly, to burst the reactor, rated for 41.4 bar (600 psig). Lessons Process Safety Culture. In hindsight, it seems the owners of T2 did not have the necessary process safety competency or know how to build a strong process safety culture. Compliance with Standards . T2 was not in compliance with the OSHA Hazard Communication Standard. No written evidence wa s found that T2 had a confined space entry, lock-out/tag-out, personal protective equipmen t program, or employee training program.

1.6 Corporate Climate and Chemistry |17 1.6 CORPORATE CLIMATE AN D CHEMISTRY If process safety culture underpins everything in a PSM S, then what underpins the culture? What conditions either support or inhibit the developm ent, m aintenance, and sustainability of the process safety culture? Mathis and Galloway (Ref 1.5). identify seven milestones on the safety culture im provement journey. Two of those m ilestones are climate and chem istry. Corporate Climate refers to the conditions within an organization as viewed by its em ployees. In the case of process safety, m anagement creates an organization’s climate through four components: Commitment, Caring, Cooperation, and Coaching. Two organizations may have a common set of activities, from which an external viewer might infer the same culture. However, the cultures m ay be very different. For example, soldiers in com bat and participants in a survival reality TV show may share som e common tasks, i.e., surviving in harsh outdoor conditions, but the climate or environment for these two situations are totally different and therefore the cultures will be very different. Corporate Chemistry refers to the structure of the culture. Like the elem ents that make up a molecule or the elements in the soil that nurtures the growth of plants, safety culture is built around the elem ents of Passion, Focus, Expectations, Proactive accountability, Reinforcement, Vulnerability, Communication, Measurement, and Trust. (Ref 1.5). In developing the culture principles presented in this book, CCPS considered both climate and chemistry. Som e culture principles addressed both climate and chem istry, as shown in figure 1.2, facing. 1.7 SUMM ARY Process safety culture has been recognized as a contributing factor in m any significant incidents that have occurred in the processing industries in recent years. Process safety culture in any

Preparing for PHA Revalidation Meetings 109 needed, unless the team that has been studying the other nodes have that knowledge. Another consideration is that in addition to the required skills, there may be other required qualifications. For example, engineering expertise may have to be provided by degreed engineers in particular disciplines or with particular licenses. Operations expertis e may have to be provided by hourly employees or field supervisors. Two options for selection of PHA revalidation team members include (1) for efficiency, keep the same team as before, or (2) for objectivity, select a new team. Either approach can be defended on its merit. If the Update approach is selected, the same team may be preferred, but organizational issues (e.g., loss or reassignment of past team members) often prevail, and the team ends up as a mix of new personnel and personnel who were involved in the prior PHA. If the Redo approach is selected, a substantially or completely different team is more commonly used and sometimes required. In addition, review of the prior PHA may indicate the need to supplement the revalidation team with members having specific areas of expertise not present on the prior team (e.g., metallurgist, control engineer, chemist, loader/unloader). Some of th ese team members may be part-time or on-call members. As is the case with any initial PHA, th e study leader should make every effort to ensure that the proposed revalidation team has a proper mix of knowledge, experience, and training. For the revalidation to be effective, participants should have specific knowledge of the process be ing evaluated in addition to general subject matter knowledge. Thus, most orga nizations have individual or collective experience requirements. Fo r example, at least one of the individuals providing operations expertise might be required to have been a qualified front-line operator for at least five years and to have worked in the subject unit for at least two years. Alternatively, collectively am ong the revalidation team participants, the organization might require at least three years of operating experience on the subject unit. Some organizations also place maximum requirements on individual job roles or experience. The concern is that an individual with decades of experience or a senior management position mi ght dominate the team discussions, particularly if others have minimal or no experience with the subject process. Thus, they write their team composition rules to favor teams with a range of experience, believing that inherently le ads to more diverse opinions and better Team Members for Redo If the PHA is being Redone because of quality or complete- ness issues, it is usually better to select new team members with a different study leader.

EQUIPMENT FAILURE 201 generated more heat and provided fuel for the flames to propagate back to the tank. (Sherman) Design considerations for process safety. Distillation is temperature, pressure, and composition dependent; special care must be taken to fully understand any potential thermal decomposition hazards of the chemicals involved. Columns need adequate instrumentation for monitoring and controlling pressure, temperature, level, and composition. The location of sensing elements in relation to column internals must be considered so that they provide accurate and timely information and are in direct contact with the process streams. A design feature of some columns is to prov ide a tall base (e.g. 3 m (10 ft)) to provide adequate Net Positive Suction Head (NPSH) to ensure that the bottoms pumps do not cavitate and fail. This also reduces the wetted area exposed to a ground-level fire. Leaks from where piping or instrumentation is connected to these vessels is a common failure. Where the material is flammable, a fire can occur that can impact surrounding equipment. Column support structures and skirts should be fireproofed, as they are not cooled by internal fluid flow and a ground fi re can lead to the column collapsing. Overpressurization can result from freezing , plugging, or flooding of condensers, or blocked vapor outlets, if the heat input to the system is not stopped. API RP 520 Sizing, Selection, and Installation of Pressure-Relieving Devices and API RP 521 Pressure-Relieving and Depressurizing Systems provide extensive guidance on the placement and sizing of pressure relief valves and other overpressure protection sy stems. (API RP 520, API RP 521) Emphasis should be placed upon the use of inherently safer design alternatives using concepts such as the following. Limiting the maximum heating medium temperature to safe levels Selecting solvents which do not require removal prior to the next process step Using a heat transfer medium that prevents freezing in the condenser Locating the vessel temperature probe on the bottom head to ensure accurate measurement of temperatures, even at a low liquid level Minimizing column internal inventory Avoiding dead legs that can corrode, plug or freeze To prevent packing fires: Cool columns to ambient temperature before opening Wash the column thoroughly to remove residues and deposits Use chemical neutralization to remove pyrophoric material Purge columns with nitrogen Monitor temperatures of the packing and column as it is opened Minimize the number of open manways to reduce air circulation

176 INVESTIGATING PROCESS SAFETY INCIDENTS Figure 8.9 Sequence Diagra m for Tank Overflow Example 8.5 SUM M ARY Careful, complete and effective eviden ce gathering is ke y to a successful investigation. Evidence can be ph ysical (damaged equipment, parts, materials, residues etc.), paper records, electronic data or position data. Consider the fragility of the evidence when determining priorities for the investigation team. Preservation of fragile evidence, su ch as electronic process data, is a key factor. The te am members may have to work several paths simultaneously and the need for additional skill sets should be identified quickly. Agreement between interested parties about how the evidence is handled can be supported us ing protocols. It is important to establish a system for do cumenting and securing evidence and a chain of custody is required for items that are moved betw een locations or different parties. Photography is used extensivel y to record evidence and can also be an inherent part of the chain of cust ody process. A set of tools and other equipment should available for meas uring, inspecting, recording and preserving evidence. Timelines and se quence diagram are effective tools to document events and cond itions and identify gaps that require further evidence gathering.

306 Human Factors Handbook The last step before signing a Permit to Work should be a field visit of Operations personnel together with the pe rson responsible of the contractor crew to check the real conditions before any activity starts. This provides a last opportunity to identify any unusual situation or hazard. Coordination This can include: • Communicating and coordinating – holding communication and coordination sessions daily, where a ll tasks over the next 48 hours are reviewed for potential conflict or o verlap. Drawings and plans can be shared and discussed. • Zonal control – this is where multiple contractors may be operating in the same area. Identified areas are defined for “zonal control”, with each zone supervised by an identified co ntractor or by client personnel. • Notice boards or asset maps – thes e should show all daily and planned activities, each contractor color-coded, so it is clear which teams are working in which parts of the plant. Double-checking task completion Some examples include: • The client supervisor performing a “w alk-down” before the start of work and at the end of each shift. This is to check the site is safe and that site housekeeping is in order. • Additional checks on the safe completion of tasks such as isolation and whether a system is safe to restore to service. • Higher levels of checks on task completion during peak periods and critical points e.g., restored to service. • Using independent specialists to check work of other specialist contractors. Double-checking tasks is important a) wh ere contractors are scheduled to work on systems that normally contain hazardou s substances or energy sources, and b) where a system is to be started or restor ed to service after a contractor completes a task. It is important that double checking is undertaken by an independent person not part of the contracting team. Ch ecks completed by colleagues may be unreliable because they assume that the other’s work is trustworthy. Demobilization An explicit demobilization activity can help to ensure that site restoration and reinstatement is performed without omission or miscommunication.

463

17. Error management in task pla nning, preparation and control 209 Table 17-5 continued Human Factors good practice for isolation 9. P&ID, schematics, or their equivale nt, should clearly indicate all connections and routes by which produc t may, for example, enter a vessel, to allow all required isolation points to be identified. 10. P&ID, schematics, power supply diag rams, or their equivalent, should clearly indicate all associated equipment that needs to be deactivated to prevent accidentally restoring supply to isolated systems. 11. P&ID should be marked up to show the connected lines/systems, the isolation valves and spading (blinding) points, test points etc. to help visualize the extent of the isolation envelope. 12. The line should be walked as part of the development of the isolation plan, to confirm the plan with the workers. 13. The open/closed position of isolation valves should be visible and unambiguously indicated. It should be possible to verify the isolating valve is 100% closed and not allowing gas or liquid to pass through. 14. It should be possible to tag and lock off manually operated isolation valves, blinds, power supplies, and other isolation equipment. 15. Include a test or “try step” to verify isolation before commencing works. This should be a positive method of verifying isolation or disabling of energy sources. 16. There should be a test/sampling point for each and every section of isolatable pipework, vessels, and other equipment. 17. Testing points, such as sample outl ets and pressure valves, should be accessible. 18. Isolation and bleed points should be close to the point of maintenance (to aid task coordination). 19. Ensure physical access and lighting for all isolation, blinding, and bleed points. 20. The “line” should be walked, by someone who did not perform the isolation. to verify it is in a safe st ate prior to hand back to operations. 21. All Permit to Work (PTWs) and other job aids should include unambiguous Stop or Hold Points for isolation to be verified and recorded before work begins or continues. 22. Documentation and status boards should unambiguously and simply represent the system state. 23. Shift handover systems should clearly communicate the system status. 24. Compliance with isolation requirements should be treated as mandatory. 25. A contemporaneous record of isolations should be documented to reduce the risk of missing an isolation (for example via an isolation certificate/ checklist). 26. Any variation to the isolation procedure must be authorized.

92 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Table 5.1. Chemical Reacti vity types and examples (Crowl 2019) Reactive Type Example Pyrophoric and spontaneously combustible Readily reacts with oxygen, igniting and burning Aluminum alkyls, Raney nickel catalyst Peroxide-forming Reacts with oxygen to form unstable peroxides 1,3, butadiene Water reactive chemical Sodium, titani um tetrachloride, boron trifluoride Oxidizer Readily yields oxygen or other oxidizing gas to promote or initiate combustion Chlorine, hydrogen peroxide, nitric acid, HF Self-reactive Butadiene polymeriza tion, acetylene decomposition, ethylene and propylene oxide, styrene, vinyl acetate Chemical incompatibles Caustic + muriatic acid Impact sensitive or thermally sensitive Trinitrotoluene (TNT) Runaway reactions Ethylene Table 5.2. Some Reactive Functional Groups (Crowl 2019) Some Reactive Functional Groups Azide N3 Diazo -N=N- Nitro -NO2 Nitroso -NO Nitrite -ONO Nitrate -ONO2 Fulminate -ONC Peroxide -O-O- Peracid -CO3H Hydroperoxide -O-O-H Ozonide O3 Amine oxide ≡NO Chlorates ClO3 The key point is to either prevent the chemical re action, or, if it is desired, to ensure that it can be safely contained in the equipment. The fi rst step in managing chemical reactivity is to

92 Guidelines for Revalidating a Process Hazard Analysis 5.1.2 Redo The Redo approach is usually selected when an Update of the prior PHA is impractical or too complex. This is base d on the evaluation of the prior PHA with respect to the criteria as described in Chapter 3 and the evaluation of recent operating experience as described in Chapter 4. In this approach, the revalidation is essentially a new PHA. Th e team starts from the beginning and performs the entire PHA in detail as if it were an initial PHA. Figuratively, the team starts with blank analysis worksheets, but in practice, most use some information from the previous PHA as a reference or guide. A Redo may be the most appropriate choice in situations where significant: • Changes have occurred to PHA requirements or analysis methods • Gaps or deficiencies have been discovered in the process safety information used as the basis for the prior PHA • Gaps or deficiencies exist in the prior PHA documentation, analysis method, or the manner in which it was conducted • Changes, particularly un- controlled changes, have occurred in the process or equipment since the prior PHA was conducted • Incidents show gaps or errors in the prior PHA The Redo approach involves conducting a PHA of the process using one or more methodologies that are appropriate for the hazards and complexity of the process. The requirements and activities for a Redo generally parallel those of an initial PHA. Significant instructional information is available elsewhere (e.g., in the CCPS book Guidelines for Hazard Evaluation Procedures [2]) on the conduct of initial PHAs. Changing the Core PHA Methodology The decision to Redo a PHA offers the opportunity to reconsider the choice of hazard evaluation core methodology to be used (e.g., HAZOP Study, rather than What- If/Checklist). If a methodology substitution is made, it should be with the intent of enhancing the quality of the PHA, and not at the expense of accuracy or thoroughness. Section 3.1.1 discusses technique selection in light of the nature of the process and its hazards.

24. Human Factors of operational level change 311 24.4 Recognizing operational level changes that impact human performance 24.4.1 The need for early recognition In any organization, changes should be recognized and managed. Some impacts are less obvious (latent as in Figure 24-1) or may be delayed. If “hard evidence” of immediate adverse impacts on human performance is demanded before taking action, this may cause failure to manage changes effectively. 24.4.2 Immediate impacts Some changes and their potential impacts are more obvious and immediate. These include the following examples: • Retraining of people on a new control system, a major change to a process, or new operations. • A major change to staffing levels, such as moving from 10 to seven people per shift team. • A major change to supervision, such as removing team leaders. • The change of the person staffing a key role such as Emergency Response Commander. • A change in piping and instrumentation, requiring a revision of Piping and Instrumentation Diagrams. 24.4.3 Latent impacts Some impacts may not be immediate nor obvious. The impact is hidden until circumstances occur to reveal it, such as a process upset. For example: • The installation of new, high reliability, or automated process equipment ma y reduce the frequency of maintenance and emergency response . This could lead to a lower frequency of performing tasks, whic h can cause skill fade. This increases the likelihood of mistakes when the task is performed. • The centralization of control rooms ma y reduce the frequency of control room operators working outside, and lead to a gradual reduction in their level of local plant knowledge. This would only be revealed when a process upset occurred. Latent impacts occur sometime after the change has occurred. Immediate Latent Figure 24-1: Types of change and impact

34 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS History indicates that abnormal si tuations that are not immediately recognized or effectively addressed ca n escalate at a facility, resulting in one or more of the following: Chemical release or fire leading to injuries or fatalities Equipment and property damage Environmental impact or non-compliance Business interruption Loss of community confidence/ reputational damage Two often-referenced abnormal situation incidents are Union Carbide’s chemical release in Bhopal, India in 1984 and the BP Texas City refinery isomerization unit explos ion in 2005 (CCPS 2008a). As summarized in Example Incident 3.2 and Example Incident 3.3, both events included significant precurso rs that were allowed to continue, resulting in abnormal situations that led to major consequences. Example Incident 3.2 – Union Carbide, Bhopal, India 1984 On December 3, 1984, the Union Carbide plant located about 3 to 4 miles (5 to 7 km) outside the cent er of the city of Bhopal, India accidentally released into the at mosphere approximately 40 metric t o n s o f m e t h y l i s o c y a n a t e ( M I C ) , an intermediate chemical used in production of carbaryl (a pesticid e). The incident resulted in the fatalities of approximately 4,000 people living near the plant and the subsequent fatality of over 16,000, as well as injuries in 200,000 and genetic mutations to affect severa l generations of offspring (Gupta 2004). The incident originated from one of the three MIC storage tanks, which were refrigerated, partially buried, and equipped with relief valves that discharged to a flar e tower through a caustic soda scrubber. At the time of the acci dent, the MIC plant had been shut down for over a month, but carbaryl production was allowed to continue, using the inventory left in the storage tanks.

11.2 Seek Learnings | 145 Lucas next set up a meeting with Mason to discuss the survey. He was able to win Mason over with his focus on the business case for process safety. Like Oliver, Mason told Lucas that the survey should not impact operational efficiency. Lucas told Charlotte of the compromise, and she quickly went to work setting up a survey. She thought to herself, “While I’m at it, I’ll throw in a couple of questions to see what people think of the process safety culture on this rig.” A few weeks later, the results were in. Charlotte analyzed the results and immediately took them to Lucas. She said, “Looks like we have a problem here. Fatigue is definitely an issue and perhaps even more problematic is the process safety culture. Seems like folks are willing to cut corners to meet production numbers.” Lucas was distressed by the results. He thought of his chats with Oliver and Mason, when he had reminded them that fatigue was often a cause of major accidents. Poor process safety culture falls into that category too. He told Charlotte, “We have to make a case for improving our process safety culture before we meet with Mason. Let’s start by looking for external incidents and see what we can learn.” Lucas continued, “Charlotte, you’re probably too young to remember the Space Shuttle disasters. But let’s put it this way, there was so much pressure to get the shuttle off the ground that the management ignored the engineers’ warnings. We can’t let that happen to us on this rig.” Charlotte nodded in agreement. She may not have been born when the Challenger exploded, but she remembered her parents talking about the incident when she was growing up. 11.2 Seek Learnings Charlotte quickly found several relevant major offshore incidents, including some that were in their region: Piper Alpha, North Sea, UK, 1988 Communications broke down from one shift to the next aboard the Piper Alpha oil and gas rig. One of the pumps was shut down for maintenance and had its pressure relief valve removed. A work permit for this pump was neither clearly displayed, nor communicated during a shift change. When a blockage occurred in the other pump, the pump that was undergoing maintenance was put See Appendix index entry S1

51 Table 3.1: Effect of Reactor Design on Size and Productivity for a Gas- Liquid Reaction (Ref 3.2 CCPS) Reactor Type Batch Stirred Tank Reactor Loop Reactor Reactor Size (l) 8000 2500 Chlorination Time (hr) 16 4 Productivity (kg/hr) 370 530 Chlorine Usage (kg per 100 kg product) 33 22 Caustic Usage in Vent Scrubber (kg per 100 kg product) 31 5 3.6 REACTIVE DISTILLATION The combination of several unit oper ations into a single piece of equipment can eliminate equipment, minimize the inventory of reactants, and simplify a process. However, there may be inherent safety conflicts resulting from this strategy (see Chapter 12). Combining several operations into a single device incr eases the complexity of that device, but it also reduces the number of ve ssels or other pieces of equipment required for the process. Careful eval uation of the options with respect to all hazards is necessary to select the inherently safer overall option. Reactive distillation is a techni que for combining several process operations in a single device. One company has developed a reactive distillation process for the manufacture of methyl acetate that reduces the number of distillation columns fr om eight to three, while also eliminating an extraction column and a separate reactor (Ref 3.1 Agreda; Ref 3.4 Doherty; Ref 3.19 Siirola). Inventory is reduced and auxiliary equipment, such as reboilers, cond ensers, pumps, and heat exchangers

90 Guidelines for Revalidating a Process Hazard Analysis 5.1 REVALIDATION APPROACHES The two revalidation approaches (introdu ced in Section 1.5) are described in detail in the following sections. 5.1.1 Update The Update approach is an incremental revision to an existing analysis to reflect the operational experience since the prio r PHA was conducted, as discussed in Chapter 4. An Update is generally a less laborious option than a Redo and is a viable option for high-quality PHAs (as di scussed in Chapter 3) of processes at facilities where changes have been effect ively managed and few or no significant incidents have occurred. In these si tuations, the revalidation team can systematically reaffirm the validity of risk judgments documented in the prior PHA and Update i t w i t h t h e r e s u l t s f r o m r e l e v a n t m a n a g e m e n t o f c h a n g e (MOC)/pre-startup safety review (PSSR) reviews and incident investigations. Two common ways of Updating a PHA include: Detailed Review. The detailed review is similar to Redoing the PHA, but should entail less time and effort, because the content of the PHA report is only amended on an “as-needed” basis and not always developed from first principles. Each section or node of the core hazard analysis, as well as any complementary analysis (such as a huma n factors checklist), in the PHA is reviewed in detail. Changes are discussed to determine their potential impact on previously documented scenarios, or to determine if new scenarios need to be added. (See Section 4.2.1.) Changes implemented in response to prior PHA recommendations or incident investigation s are included in this review. (See Section 3.3.1.) They typically result in revisions to the list of safeguards, but they may also affect other portions of the PHA worksheets. Those scenarios affected by the changes, new learnings, or incidents need to be thoroughly addressed, and necessary changes in the PHA are recorded in the PHA documentation (e.g., HAZOP worksheets, Checklist responses.) Change and Incident Review. Alternatively, the revalidation team may choose to review only the changes and incidents that have occurred since the prior PHA. The revalidation team, with appropriate reference to the prior PHA, discusses the significance of the changes and incidents, and then documents their deliberations and decisions. The revalidation team may use a separate worksheet, and the information recorded us ually includes a description of each change (or incident), its process safety significance, and any risk management decisions or recommendations made. (See Se ctions 3.3.1, 4.2.1, and 4.2.2.) Either

Table 14-1: Suitability of and differences between competency assessments Assessment Method Type of human performance Advantages Disadvantages Issues to consider Verbal Questioning “What if” scenarios Suitable for knowledge- based competency • Useful for investigating knowledge • Can be standardized • Valuable tool for collecting evidence across activities • Not sufficient enough in itself to demonstrate competency • Least likely to be representative of real work conditions • Assessors may answer their own questions • Assessors should be trained and experienced in the use of questioning techniques Written exam Suitable for knowledge- based competency • Valuable for knowledge- based activities • Can be well structured and standardized • Requires assessment time for those being assessed and for the assessors, and time for the scorers • Requires time away from job • Requires skilled assessors to assess the outcome • Danger that knowing is confused with being able to do • Provides supplementary evidence of actual performance

94 | 7 Keeping Learning Fresh Another visual-spatial technique is a drawn diagram. T.J. Larkin, a well- known safety communicator, advocates hand-drawn diagrams over photographs for highlighting the issue for the same reason that animation is preferred over actual video of an incident–their simplicity makes it easier to focus on the important details (Larkin 2012). Repsol provided an example of such a diagram addressing a utility service contamination event, an excerpt of which is reproduced as Figure 7.1. Figure 7.1 Simple Drawn Diagram Example (Source: Repsol, reproduced with permission) Finally, it can be remarkably effective to leave evidence of past incidents in place to serve as a daily reminder of why we must do what we do. One company left shrapnel from a significant explosion lodged in a wall, with a sign commemorating the event. Another company has a framed photograph showing the plant on fire hanging by the employee entrance, with the caption, “Never again. We all know what we must do.” The learning model scenarios in Chapters 9, 10, and 11 use posters with graphics and/or videos to stress the importance of proper safety practices. To communicate visually about responding to person-down incidents you could create a video, as mentioned in Section 7.1. Alternatively, you could draw a simple poster, such as shown in Figure 7.2 (facing page).

Pumps and Compressors 179 to the control valve to open and open further, and this is why it doesn’t keep up with the low flow. The other option is a loop similar to the flow control loop but a pressure control loop. In this design, instead of a flow sensor we install a pressure sensor and the loop is a pressure loop. This control system doesn’t work for all conditions as it only can work when the pump care is adequately steep. It is generally justifiable to use a control system on the minimum flow protection pipe for larger pumps, say larger than 35 hp. For smaller pumps a similar system could be used, but instead of a control system we can have an on/off system. This means the flow sensor is still used but instead of a control valve a searching valve can be used. Whenever the pump flow goes below the minimum flow this valve will be opened to recirculate some flow to increase the pump flow rate to a value higher than the minimum flow. As you can see this is not a very accurate way to adjust the flow to protect the pump against the minimum flow conditions. However, for smaller pumps, say less than 35 down to about 20 hp, it is not justifiable to use a more expensive control loop and instead this switching loop is used (Figure 10.9). For smaller pumps less than 20 hp a continuous mini- mum flow is implemented for the pump. In such cases just a restrictive orifice (RO) could be used on the minimum flow protection pipe (Figure 10.10). For very small pumps, say less than 5 hp, as was men- tioned there is no need to consider implementing a minimum flow pipe at all. It is important to realize that a minimum flow protec - tion pipe is not working during the majority of pump operation. It is because in the majority of times the con-trol valve on the minimum flow protection pipe is closed. This valve will be open only when the flow to the pump goes below the minimum flow value (Figure 10.11). This is the reason that some companies violate their pipe sizing criteria and use higher velocity design bases for a pipe to come up with a narrower and cheaper pipe for this purpose. However, the control valve on minimum flow protec - tion pipe should be “failed‐open” (FO). This valve should be FO because it’s function is important to keep the pump operational. When losing the instrument air if this valve is “failed‐closed” (FC) the pump will start to vibrate if the flow goes below the minimum flow value. So to keep the pump operational this valve should be FO.FC Figure 10.9 Minimum flo w protection pipe with switching loop. RO Figure 10.10 Minimum flo w protection pipe with restrictive orifice. Senses more than 10 0 m3/hr so closes the va lve Senses less than 10 0 m3/hr so opens the va lve165 m3/hr0 m3/hr FV FVFC FCFX FXFT FTFIC FIC 165 m3/hrMinimum flow: 100 m3/hr 20 m3/hr 80 m3/hr 80 m3/hrFigure 10.11 Func tioning of a minimum flow protection pipe.

20. Situation awareness and agile thinking 253 • An overreliance on general expe ctations about how the system functions, in the absence of real-time data. • Several pieces of information may not have been registered by the operator due to cognitive capacity limitations (e.g., limits to working memory). 20.4 Causes of poor situation awareness and rigid thinking 20.4.1 Factors influencing situation awareness Operators’ situation awareness is influenc ed by several factors. These factors may reduce operators’ cognitive capacity and impair their performance. They include: • Experience and training. • Time pressure and workload. • Motivation, stress levels, and work fatigue. • Coordination with team members, and dependence on others. • Weather conditions (visibility), eq uipment, and process noise. • Complexity of the process. • Location or site of the plant. • Abnormal situations. Mica Endsley and Debra Jones in their book “SA Demons: The Enemies of Situation Awareness” [84] identified ei ght causes responsible for failures in situation awareness. Those were termed as demons of situation awareness and include: 1. Attention tunneling - focus on certain type of information and excluding the rest. 2. Requisite Memory Trap – reliance on memory information, despite human memory limitations. 3. Stress, anxiety, fatigue and other stressors – stress and fatigue impair working memory. 4. Data overload- more data is ava ilable than the human cognition can process. 5. Misplaced salience –the way information is presented (e.g., bright colors and flashing lights) overwhelm and misdirect operators' attention. 6. Complexity creep - the more complex the system the more difficult it is for operators to develop accurate comprehension of the situation. 7. Errant mental models – incorrect mental model may lead to inaccurate interpretation of data. 8. Out-of-the -loop syndrome – highly automated systems may result in operator shaving low awareness of the systems.

1.7 Summary |19 facility form s the foundation of the PSM S, regardless of what is written. The quality of demonstrated leadership directly affects the strength and quality of the process safety culture, and the quality and health of the process safety program itself. While m anagers of PSMS’s clearly serve as process safety leaders, all m anagers and executives in enterprises that manage m ajor process hazards can and indeed should demonstrate process safety leadership and help set a strong, positive culture. CCPS defines process safety culture as: “The pattern of shared written and unwritten attitudes and behavioral norms that positively influence how a facility or company collectively supports the development of and successful execution of the management systems that comprise its process safety management system, resulting in the prevention of process safety incidents.” Other definitions related to safety culture and particularly process safety culture can be found in the literature. There is no single definition of safety or process safety culture. Numerous definitions have been presented in the literature in recent year. The CCPS definition embodies all the lessons learned in the literature to produce a definition serving the m ajor hazard industries ranging from upstream oil and gas, through refining, chem icals and pharm aceuticals to manufacturers who handle chem icals and practice chem istry in other industries. Process safety culture and the organizational culture that it fits into are strongly linked. The process safety/safety culture of an organization cannot exist in a vacuum. Any problem s or issues in the organizational culture will also show up in the process safety culture, and elsewhere in the organization as well. Likewise, efforts to improve process safety culture can spill over to positively impact the overall culture of the organization. Organizations that have a strong overall culture and strong process

6. Simplify – An Inherently Safer Strategy In the context of inherent safety, simplify means designing and/or operating the process to reduce or eliminate unnecessary complexity in order to reduce or eliminate the ch emical hazard. Reducing complexity helps to accomplish a number of goals, for example: 1) minimization or elimination of extra equipment that can fail and contribute to a process safety incident, and 2) reducing or eliminating extra processing steps with hazardous conditions or whic h may result in a release of the hazardous chemicals or energy, or (3 ) reducing the use or the number of chemicals necessary in a process. A simpler process is generally safer and more cost-effective than a compl ex one. Kletz (Ref 6.9 Kletz 1998) offered several reasons why proce ss designs are unnecessarily complex: The need to control hazards . Instead of avoiding hazards using inherently safer design principles, most designers choose to control them actively using controls, alarms, and safety instrumented systems. The desire for technical elegance . To some designers, simple equates to crude or primitive, whereas, if carefully designed, a simple process can achieve what it needs to do without excess equipment. A simple process design including only the essential elements to safely carry out its intended task(s) is more elegant than a complicated process th at does the same thing. The failure to carry out hazard an alyses until late in the design . PHAs and similar studies performed late in the design usually result in more active controls and equipment rather than more inherently safer solutions. If a preliminary PHA or similar hazard analysis technique is performed at the conc eptual stage of the project, it may be easier to incorporate IS techniques, especially substitution or elimination. Following standards and specifications that are no longer appropriate or not completely applicable . Active solutions to potential hazards that are sometimes contained in design / engineering standards and specifications can accumulate in a 103 (VJEFMJOFT
GPS
*OIFSFOUMZ
4BGFS
$IFNJDBM
1SPDFTTFT
"
-JGF
$ZDMF
"QQSPBDI #Z
$$14 ¥

5IF
"NFSJDBO
*OTUJUVUF
PG
$IFNJDBM
&OHJOFFST

Appendices 191 Q T R • Contamination of utilities (e.g., water)? • Contamination from spills or runoff? • Noise? • Transport of hazardous materials from other sites? • Flooding (e.g., ruptured storage tank)? VII. Unit Layout Are large inventories or release po ints for HHCs located away from vehicular traffic within the plant? Could specific siting hazards be posed to the site from credible external forces such as high winds, earthquakes and other earth movement, utility failure from outside sources, flooding, natural fires, ice accumulation, and fog? Is there adequate access for emerge ncy vehicles (e.g., fire trucks)? Are access roads free of the possibility of being blocked by trains, highway congestion, spotting of rail cars, etc.? Are access roads well engineered to avoid sharp curves? Are traffic signs provided? Is vehicular traffic appropriately restricted from areas where pedestrians could be injured or equipment damaged? Are cooling towers located such that fog that is generated by them will not be a hazard? Are the ends of horizontal vessels fa cing away from personnel areas? Is hydrocarbon-handling equi pment located outdoors? Are pipe bridges located such that they are not over equipment, including occupied structures and administration buildings? Is piping design adequate to withstand potential liquid load? VIII. Location of the Unit Relative to On-site and Off-site Surroundings Is a system in place to notify neighboring units, facilities, and residents if a release occurs? Are the detection systems and/or alarms in place to assist in warning neighboring units, facilities, and residents if a release occurs? Do neighbors (including units, facilit ies, and residents) know how to respond when notified of a release? Do they know how to shelter-in- place and when to evacuate? Are large inventories or release po ints for HHCs located away from publicly accessible roads? Is the unit, or can the unit be, located to minimize the need for off- site or intra-site transporta tion of hazardous materials? Are workers in this unit protected from the effects of adjacent units or facilities (and vice versa), and are environmental receptors and the public also protected from the following:

GLOSSARY xxv Integrity Operating Window An Integrity Operating Window (IOW) is a set of limits used to determine th e different variables that could affect the integrity and reliability of a process unit. An IOW is the set of limits under which a process, piece of equipment, or unit operation can operate safely. Working outside of IOWs may cause otherwise preventable damage or failure. Lagging Metric A retrospective set of me trics based on incidents that meet an established threshold of severity. Layer of Protection Analysis (LOPA) An approach that analyzes one incident scenario (cause-consequence pa ir) at a time, using predefined values for the in itiating event frequency, independent protection layer failure probabilities, and consequence severity, in order to compare a scenario risk estimate to ri sk criteria for determining where additional risk reduction or more detailed analysis is needed. Sc enarios are identified elsewhere, typically using a scenario-based hazard evaluation procedure s uch as a HAZOP Study. Leading Metric A forward-looking set of metrics that indicate the performance of the key wo rk processes, operating discipline, or layers of protection that prevent incidents. Loss of Primary Containment An unplanned or uncontrolled release of material from primary containment, including non-toxic and non-flammable materials (e.g., steam, hot condensate, nitrogen, compressed CO 2 or compressed air). Management of Change A management system to identify, review, and approve all modifications to equipment, procedures, raw materials, and processing conditions, other than replacement in kind, prior to implementation to help ensure that changes to processes are properly analyzed (for example, for potential adverse impacts), documented, and communicated to employees affected.

3.7 References |105 3.11 Smith, J., & Foti, R., A pattern approach to the study of leader emergence , The Leadership Quarterly, Vol. 9, 1998. 3.12 Foti, R., & Hauenstein, N., Pattern and variable approaches in leadership emergence and effectiveness , Journal of Applied Psychology, Vol. 92, 2007. 3.13 Scouller, J. (2011). The Three Levels of Leadership: How to Develop Your Leadership Presence, Knowhow and Skill . Cirencester: M anagement B ooks 2000. 3.14 Stricoff, S., What Process Safety Needs in a Leader , Safety + Health, 2013. 3.15 United Kingdom Health and Safety Executive (HSE), Safety and environmental standards for fuel storage sites , Process Safety Leadership Group , Final report , 2009. 3.16 CCPS, Recognizing Catastrophic Incident Warning Signs in the Process Industries, American Institute of Chemical Engineers, New York, 2012. 3.17 Mathis, T., Galloway, S., STEPS to Safety Culture ExcellenceSM, Wiley, 2013. 3.18 Paradies, M., Has Process Safety Management Missed the Boat? AIChE, Process Safety Progress, Vol. 30, No. 4, 2011. 3.19 Elliot, M ., et. al., Linking OII and RMP data: does everyday safety prevent catastrophic loss? International. Journal of Risk Assessment and M anagement, Vol. 10, Nos. 1/2, 2008. 3.20 Organization for Economic Cooperation and Development (OECD), Corporate Governance for Process Safety, OECD Guidance for Senior Leaders in High Hazard Industries, J une 2012. 3.21 United Kingdom Health and Safety Executive (HSE) Health & Safety Laboratory, Safety Culture: A review of the literature , HSL/2002/25, 2002. 3.22 Oracle, An Oracle White Paper, Seven Steps for Effective Leadership Development , J une 2012. 3.23 Baker, J .A. et al., The Report of BP U.S. Refiner ies Independent Safety Review Panel , J anuary 2007 (B aker Panel Report). 3.24 HR Council, HR Planning, Succession Planning (http://hrcouncil.ca/hr-toolkit/planning-succession.cfm)

68 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Figure 4.13. Degrees of congestion from low to high (left to right) (BakerRisk 2001) Understanding, modeling, preventing, and mitiga ting vapor cloud explosions is an area of significant focus in industrial facilities handlin g flammable materials due to the harm they can potentially cause. Many signific ant process safety . events have highlighted the importance of this focus and the importance of locating and de signing buildings so as to protect occupants. API RP 752, 753 and 756 provide guidance on blast-resistant design of permanent structures, portable structures, and tents, respectively. Th e CCPS “Guidelines for Evaluating Process Plants Buildings for External Explosions, Fires, and Toxic Releases” provides detailed guidance on multiple explosion modeling methods. Methods us ed in modeling explosions are discussed in Chapter 13. Physical Explosions Overpressure – Any pressure above atmospheric caused by a blast. (CCPS Glossary) Impulse - The area under the overpressure-time curve for explosions. The area can be calculated for the positive phase or negative phase of the blast. (CCPS Glossary) Deflagration - A combustion that propagates by heat and mass transfer through the un-reacted medium at a velocity less than the speed of sound. (CCPS Glossary) Detonation - A release of energy caused by the propagation of a chemical reaction in which the reaction front advances into the unreacted substance at greater than sonic velocity in the unreacted material. (CCPS Glossary) Physical explosions are caused by the rapid release of mechanical energy, and include vessel ruptures, BLEVEs and rapid phase transition. A vessel rupture occurs when the internal pressure exceeds the mechanical strength of the vessel. The vessel strength is sometimes weakened due to a material defect or corrosio n. A rapid phase transition can occur when a material is exposed to a heat source. This in creases the material’s volume, increasing the pressure in the container. Figure 4.14 summarizes the various explosion types and terminology. It is possible for several to occur with any incident.

Table A.3 IST Checklist Moderate Questions 3 MODERATE Questions: 3.1 Is it possible to limit the supply pressure of (hazardous) raw materials to less than the maximu m allowable working pressure of the vessels to which they are delivered? 3.2 Is it possible to make reaction conditions (for hazardous reactants or products) (temperature, pressure) less severe by using a catalyst, or a better catalyst (e.g., structured or monolithic vs. packed-bed)? 3.3 Can the process be operated at less severe conditions (for hazardous reactants or products) by considering: • Improved thermodynamics or kinetics to reduce operating temperatures or pressures • Changes in reaction phase (e.g., liquid/liquid, gas/liquid, or gas/gas) • Changes in the order in which raw materials are added • Raw material recycle to compensate for reduced yield or conversion • Operating at lower pressure to limit potential release rate • Operating at lower temperature to prevent runaway reactions or material failure 3.4 Is it possible to use less concentrated hazardous raw materials to reduce the hazard potential? • Aqueous ammonia and/or HCl instead of anhydrous • Sulfuric acid instead of oleum • Dilute nitric acid instead of concentrated fuming nitric acid • Wet benzoyl peroxide instead of dry 3.5 Is it possible to use larger partic le size/reduced dust forming solids to minimize potential for dust explosions? 3.6 Are all process materials (e.g., he ating/cooling media) compatible with process materials in event of inadvertent contamination (e.g., due to a tank coil or heat exchanger tube failure)? 3.7 Is it possible to add an ingredie nt to volatile hazardous materials that will reduce its vapor pressure? 447

16. Task planning and error assessment 181 A resilient and methodical approach to task planning The production of work instructions can be a high frequency and knowledge- based activity during which mistakes may occur. The identification of task-specific risks and safety requirements draws on knowledge of the process, process hazards, and safety procedures. Task planning needs to be methodical and account for all relevant hazards. If task planning is not done well, hazards may be ignored or overlooked, safety procedures may be incomplete, and instructions can be unclear. The people that are planning the tasks must be competent, and the system of task planning should be resilient. Optimism bias in task planning Tasks often have completion deadlines. Ad ditional pressures include restoring production quickly. This can contribute to “optimism bias” within task planning, for example, being overly optimistic about how long a task will take. It can also involve downplaying the challenges and risks in performing a task. This can contribute to insufficient time being scheduled for a task and insufficient preparation of the team. Including error assessme nt in task planning Many tasks are complex, with many task steps. These tasks can occur over many hours, they may involve many people, and they may involve work in physically separate locations. There may be unanticipated events (such as equipment failure) that may increase task complexity and require a change of plans. These conditions create increased variability and the need for operators to adapt, creating the potential for errors and mistakes. In addition, motivated staff can be very “task focused” and intent on completing the task and solving the problems. This can create a risk of losing awareness of the situation, improvising unsafe ways of completing a task, and overlooking unexpected events or conditions that re quire a change in their actions. When people are task focused, they can miss “weak signals” around them that the situation is unsafe or is changing. Task planning should: •Identify the potential for error •Identify the means to support a successful task performance

Chapter No.: 1 Title Name: Toghraei c15.indd Comp. by: ISAKIAMMAL Date: 25 Feb 2019 Time: 12:30:36 PM Stage: Proof WorkFlow: CSW Page Number: 293 293 Piping and Instrumentation Diagram Development, First Edition. Moe Toghraei. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.Companion website: www.wiley.com/go/Toghraei_PID 15.1 Introduction What is the “plant control system, ” and how we can implement it? In more technical terms, how can we implement a BPCS (basic process control system) in a plant? A BPCS, or regulatory control system, can be divided into two main levels in plants: 1) Pl ant‐wide control. This is used to provide overall control for the entire plant. Some people refer to this as “heat and material balance control, ” since its primary role is to ensure that the plant produces the product in the predicted quantity and with the predicted quality. This control basically creates a link between a plant and its H&MB (heat and mass balance) table. Not all processes and unit operations are individually and directly connected to the plant‐wide control system. 2) Equipment c ontrol (unit operation control). Each “unit” within the process may need to be controlled via its own BPCS with corresponding control loop(s). Most pieces of equipment don’t have an operating “point”; rather, they have an operating “window” . This is not necessarily because of an inherent weak - ness of the equipment; this is something we like and gives the equipment the capability to “fluctuate” under different process conditions. The main duty of the unit control is to bring the unit to its optimum point within its operating window in each different set of process conditions. It is important to mention that the classification of the BPCS control into two levels is only based on their concepts; generally there is no difference in control hardware in a plant, and these two groups cannot be recognized or differentiated in P&IDs. Each of the above concepts carries one aspect of plant control. Plant‐wide control assures the “attachment” of the plant to its capacity and the quality of product(s), whereas equipment‐level control tries to bring a piece of equipment to its best operating point within its operating window and also protect the equipment at its weak points. While in a P&ID all control systems can be traced, a PFD generally shows only plant‐wide control. However, some P&IDs show some major elements of equipment‐level control too. Plant‐wide control is discussed in Sections 15.2–15.4 of this chapter while equipment‐wise control will be discussed in Section 15.5 and after. 15.2 Plant‐Wide Control There are two purposes to installing plant‐wide control: 1) The main pur pose is to link the plant to its heat and material balance table. 2) The s econd purpose is surge or disturbance management. 15.3 Heat and Mass Balance Control The first purpose of plant‐wide control is heat and mate- rial control. Some people may prefer using the phrase of “mass balance” or “material balance” control (rather than “heat and material balance control”). Their logic is: “in plants, we care about flow rate and quality of the product(s); no one is looking for a specific product with a specific temperature. ” While their logic is true in the majority of cases, it should be noted that to have a product with a specific quality, it may need to go through different steps of operation, which need spe cific operating temperatures. Therefore, “heat con- trol” is still needed. In order to ensure this, theoretically we must have at least one flow control loop and one composition control loop with a manual set point that comes from the H&MB table.15 Plant Process Control

KEY RELEVANCE TO PROCESS PLANT OPERATIONS 39 Example Incident 3.3 – BP Texas City 2005 – ( cont.) Lessons learned in relation to abnormal situation management: Operating procedures: This event happened during a transient operation when starting up the column. The startup procedure was not correctly followed, although the US Chemical Safety and Hazard Investigation Board (CSB) subsequently found that procedural deviations during startup had become common practice. (CSB 2007) Management of Change: Operating practices changed over time to reflect difficult control of startup, but management failed to address the safety im plications of these changes and procedures were not updated. Process Monitoring and Control: o The process was allowed to operate outside normal limits for both the column bottom level and the temperature. Critical alarms were out of commission. A lack of situational awaren ess by the control panel operators resulted in no attempt to bring the column back into safe operating limits during the incident. o Providing suitable informat ion and system interfaces for front line staff enables them to reliably detect, diagnose, and respond to potential incidents. The HMI should be designed so that the operator has multiple ways of viewing plant status, improving situational awareness that would make it easier to identify level gauge failure. Organizational roles and work processes: Personnel response to the incident was inadequate and lacked supervisory intervention. Shift handover documentation was insufficient to help incoming operators understand the gravity of the situation. Safety culture was poor. Work Environment: Inadequate definition of appropriate workload, staffing levels, and working conditions for front line personnel.

Table 21-2 continued Human performance tool (HPT) Description Usage Dynamic Risk Assessment Stop-Think-Act-Review (STAR) assessment is often part of a Dynamic Risk assessment. Following is a description of the S-T-A-R steps: Stop • Look for hazards • Review hazards • Has the situation changed? Think • Evaluate the situation • Evaluate options Act • Apply safety measures • Recommence the work Review • Complete an after Action Review • Reassess the system of work • Record lessons learned for sharing When operating in skill-based and rule- based performance modes. Particularly effective for repetitive tasks. STOP Act now

11. Inherent Safety & the Elements of a RBPS Program Chapter 11 will describe the relation ship between the four main IS strategies, i.e., Substitution , Minimization , Moderation , and Simplification and each element of a PSM/RBPS pr ogram. The possible use of the IS strategies in carrying-out the activities of the elements will be described. The interrelationship between inhere nt safety and PSM programs has been previously described by Kletz and Amyotte (Ref 11.16 Kletz 2010) and Amyotte, Hendershot, et al (Ref 11.5 Amyotte) and their conclusions are summarized herein. Amyotte & Klet z argue that inherent safety is implicitly part of PSM/RBPS programs, that it not be included as a discrete element, but as part of th e thought process for the design of these programs and how they are a pplied and implemented. In their book and in this chapter more explicit applications of the IS strategies are proposed (Ref 11.16 Kletz 2010). For some PSM/RBPS elements the relationship with the IS strategies is direct and relatively strong, e.g., PHA, whereas for other elements the rela tionship is more indirect, e.g., Measurement and Metrics. The four IS strategies are usually applied in the context of the chemicals and the hazards/ risks that they present, e.g., substituting one chemical with another that presents less of a process safety hazard, or minimizing the inventories of the chemicals that can create process safety hazards. Traditionally, IS has b een regarded as an intrinsic part of the design of the process, i.e., part of the “hardware” aspect of the process or as part of the properties of the process with respect to the hazards presented by the hazardous ma terials used in the process. In this chapter, the application of the same thought processes of IS can be applied to aspects of the process that are largely programmatic, i.e., the design and execution of the PSM/RBPS program policies, practices, and procedures. This is a new applicat ion of IS concepts. For example, Minimization has traditionally been asso ciated with reducing the inventory of hazardous materials, wh ich is a physical aspect of the process. However, we can pose the question: Can/should shift schedules 268 (VJEFMJOFT
GPS
*OIFSFOUMZ
4BGFS
$IFNJDBM
1SPDFTTFT
"
-JGF
$ZDMF
"QQSPBDI #Z
$$14 ¥

5IF
"NFSJDBO
*OTUJUVUF
PG
$IFNJDBM
&OHJOFFST

294 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION The experiment is repeated for different do ses and Gaussian curves are drawn for each dose. The mean response and standard deviation are determined at each dose. A complete dose-response curve is produced by plotting th e cumulative mean response at each dose. For convenience, the response is plotted versus the logarithm of the dose. The logarithm form arises because in most organisms some subjects can tolerate high levels of exposure while others are sensitive. Simpler forms of dose response are useful for emergency response planning and these use a standard time of exposure (30 or 60 minutes) and provide different levels of impact from mild impact to serious injury (e.g. ERPG discussed in Section 13.7.2). For most engineering computations, an equati on is more useful than the dose-response curve. For single exposures, the probit (proba bility unit) method provides a transformation method to convert the dose-response curve into a straight line. Probit equations are available for a variety of exposures, including exposures to toxic materials, heat, pressure, radiation, and impact. For toxic exposures, the causative variable is based on the concentration; for explosions, the causative variable is based on the explosive overpressure or impulse, , depending on the type of injury or damage. For fire exposure, the causative variable is based on the duration and intensity of the radiative exposure. The probit method is generally the preferred method of choice for consequence analys is studies. A limitation is the restricted set of chemicals for which probit coefficients are published. Toxic Outcomes Toxic impact models are employed to assess the consequences to human health as a result of exposure to a known concentration of toxic ga s for a known period of time. For consequence analysis, the toxic effects are due to short-term exposures, primarily due to vapors. Chronic exposures are not considered here. Many releas es involve several chemical components or multiple effects. The cumulative effects of simu ltaneous exposure to more than one material are not well understood. Predictions of gas cloud concentrations and du rations at specific locations are available from neutral and dense gas dispersion models. Toxic concentration criteria and methods incl ude the following which were discussed in Chapter 6. Emergency Response Planning Guidelines fo r Air Contaminants (ERPGs) issued by the American Industrial Hygiene Association (AIHA) Acute Exposure Guideline Levels (AEGL) ma intained by the U.S. EPA in cooperation with the National Academies. (EPA AEGL) Immediately Dangerous to Life or Health (IDLH) levels established by the National Institute for Occupational Safety and Health (NIOSH) Emergency Exposure Guidance Levels (EEGLS) and Short-Term Public Emergency Guidance Levels (SPEGLs) issued by th e National Academy of Sciences/National Research Council (NAS) Probit Functions

378 INVESTIGATING PROCESS SAFETY INCIDENTS was set below the current normal batch level. The batch size had been changed from a 70% level to an 85% level but the alarm was still set for 80%; as a result, the batch level alarm was going off each time Kettle No. 3 was filled to the new normal batch level. (Management of Change, Conduct of Operations) viii Absence of Redundant Protection There was no redundant back-up protec tion (second level or monitoring of the pump) to shut down the pump in case it was blocked in. The hazard i d e n t i f i c a t i o n & r i s k a n a l y s i s ( H I R A ) for the raw material storage, catalyst preparation, and catalyst storage areas was up for renewal this year. The prior HIRA was not as thorough as expected by today’s standards. The corporation has now established criter ia for HIRA leaders and has an approved list of resources. (Hazard Identification & Risk Analysis, Engineering Design) N ote: The isopentane feed valve is designed to fail closed on power failure, to prevent reverse flow from th e kettles to the raw material storage tanks. This is the appropriate failure position for this valve. Other Causal Factors: Other causal factors were related to possible improvements in emergency planning and response, as follows: ix Fire Brigade Procedures No fire brigade member reported to the fire pump house when the fire alarm sounded. Interviews sugges t personnel were confused about whose responsibility it was to go to the fire pump house. This may be a training or drill issue. (Emergency Management) x Understanding of Fire Hazards The fire brigade approached the cata lyst preparation area to attempt rescue of a victim while firewater was unavailable. The small amount of water available on the fire engine wa s enough to protect the rescuers from the radiant heat from the fire, bu t was no protection against metal fragments. While the fire brigade did not recognize the potential hazards of this incident, further invest igation is needed to determine if the emergency responders received insufficient training or if the emergency response plan is deficient in this area. (Emergency Management) xi Personnel Headcount Procedures The presence of the contractor (wor king in the instrument house) was not known to unit personnel. The contractor works in the area routinely, sometimes in the instrument house and sometimes in the rack. Because the instrument house is a general purpose area, a permit is not required

Piping and Instrumentation Diagram Development 180 All the systems on the minimum flow protection pipe could be replaced with a device named an automatic recirculation valve or ARC. It looks very attempting to use a single device such as ARC instead of the more complicated, utility dependent control system. However, ARCs have some shortcomings. First of all they are very prone to plugging if the service liquid is not clean. The other limitation is that ARCs are not available in large sizes, possibly larger than 6″. Inside ARCs are a bypass valve and a spring‐loaded check valve. The P&ID symbol for an ARC is very similar to a check valve but the differ - ence is that the ARC symbol has two outgoing lines (Figure 10.12). It is important to know that not all loops around the pumps in the P&ID are “minimum flow protection pipes, ” there could be similar pipes in the P&ID or plants for some other reasons. A loop around the pump could be implemented on a pump for the following reasons: ●As a minimum flow protection pipe. ●To protect the pump from gradual heating during dead head conditions (mainly for flammable liquids). ●To provide fluid moving around the pump to maintain the flow during electrical outages where the pump is supplied by emergency electricity. ●To achieve an acceptable efficiency in very small cen-trifugal pumps (less than 10 m 3 h−1). ●To providing a start‐up pipe for positive displacement pumps. If this loop around a pump is implemented for each of above reasons, they should be sized based on different criteria. 10.6.3 C avitation Cavitation is a phenomenon that is related to the genera- tion and collapse of vapor bubbles inside of pumps. When this happens, the pump fails prematurely because of bubbles slamming into the impellor and the internal side of the casing. The collective name for these events is cavitation. The main underlying reason for cavitation is a lack of enough pressure on the suction side of a centrifugal pump. When a centrifugal pump operates, it basically “sucks” the liquid, which can generate bubbles on its suction side, causing cavitation.Have you tried to suck a carbonated drink out of a tall bottle with a narrow straw? You might notice a bunch of bubbles coming into your mouth. However, luckily cavitation won’t happen in your mouth because bubbles don’t have much speed! If the liquid on the suction side of a centrifugal pump has “enough” pressure, it won’t release gas and cavitation won’t happen. The problem of a “lack of enough pressure on the suction side of a centrifugal pump” can be stated more technically as: a “lack of enough NPSH. ” NPSH or “net positive suction head” is basically the total “effective” pressure of a liquid at the suction flange of a centrifugal pump in “head” units (e.g. meters or feet). Each centrifugal pump has a minimum acceptable NPSH, which is reported by the pump manufacturer and is termed the “required NPSH, ” or net positive suction head required (NPSH R). A typical NPSH R for a centrifugal pump could be a  value anywhere from less than a meter, up to more than 10 m. After buying a centrifugal pump, the process and the control system should be designed in order to ensure that the liquid has enough pressure at the suction flange to prevent cavitation. This pressure, which is provided by the system (rather than the pump), is termed “avail-able NPSH, ” or net positive suction head available (NPSH A), and is reported in the same units as head (e.g. meters or feet). This concept is shown in Table 10.7.NPSH A should be higher than NPSH R by a pre‐selected margin, otherwise the centrifugal pump will most likely cavitate. The service fluid type is part of the “system” too. Where the pumping fluid is hotter or more volatile the pump is more prone to cavitation. There are, however, cases where NPSH A is lower than “NPSH R + margin. ” There different techniques available to solve the problem. To apply some of the solutions the designer should go back to the design stage of project but for some others, some changes during the P&ID devel-opment could solve the problem. Each of the items in Table 10.8 need detailed evalua- tion by process engineers and other stakeholders to check their applicability. A “stand pipe, ” which is a solution to increase NPSH A, is a simple vertical vessel that accumulates the liquid to a higher level for the benefit of the downstream centrifugal pump. The dimension of a standpipe is decided and these rules of thumb can be used. The diameter is prefer - able less than 24″ to be able to use a piece of pipe (seam-less) as the body of the standpipe. The height is primarily defined by the required level of liquid in it to provide enough NPSH A. It is generally preferable to leave the top  of the standpipe open to atmosphere to get rid of Figure 10.12 Minimum flo w protection pipe with an automatic recirculation valve.

34 PROCESS SAFETY IN UPSTREAM OIL & GAS commencing in 1992. This was based extensively on API RP 750. This applies mainly to downstream refining, chemical and petrochemical plants. It also covers some upstream facilities, such as large on shore treatment facilities and gas plants, but not well construction. Coverage is based on threshold amounts of nominated hazardous chemicals. Other in cidents, including Bhopal in 1984 (Less, 2012), led the EPA to issue the Risk Management Plan Regulations, which are similar to PSM, but with a focus on offsite impacts. As we ll as a formal process safety management system, the EPA requires that the facility car ry out an offsite consequence analysis to predict worst case hazard zones and to repo rt on 5-year incident histories. This regulation also applies to larger scale upstream treatment plants based on their maximum inventory, similar to OSHA PSM. In the prescriptive approach, the company and the regulator determine, by a review of documents or by inspection, whether the regulatory requirement has been implemented. This process ensures that go od solutions are adopted. It does require that the regulator or industry body develops and keeps up to date the required solutions. Maintaining current solutions is demanding in industries such as deepwater offshore, which have complex and changing designs. A potential issue with prescriptive requirements is that once the remedy is implemented, then a culture of compliance may develop and not consider other solutions, even if they could potentially provide additional risk reduction. The EU approach for downstream is goal-based. Two downstream major incidents occurred in Europe in the 1970s (Flixborough and Seveso) and these along with the Bhopal incident in 1984 (all described in Lees, 2012 and CCPS, 2008b) resulted in the EU developing the Seveso Directive, a goal-based process safety regulation. A summary of the current requirements is provided on the UK HSE website (see references). This requires a formal process safety management system and a risk assessment as part of a safety case document. In the goal-based approach the process safety objective is defined, but the actual solution is left to the company. The safety objective is often ALARP – As Low As Reasonably Practicable. For ALARP, additional safety measures to reduce risk should be assessed and implemented so long as the measure is practical, considering trouble, time, and money. The company must demonstrate to th e regulator that the solution adopted is adequate for the potential risk. In the upstream domain, the Piper Alpha di saster in the UK sector of the North Sea in 1988 resulted in 167 fatalities (IChemE, 2018 and Oil & Gas UK, 2013). The subsequent Cullen Inquiry (HSE, 1990) recommended, amongst other things, a safety case approach for offshore. This re quires a safety case or for floating drill rigs, an IADC-style HSE Case. The safety case includes a facility description, major hazards identification and risk assessment, the process safety and environmental management system, the technical solutions (addressing key barriers – their performance standards, how they are maintained, and verification), and the emergency response actions/procedures. The UK regulations are implemented in a family of regulations led by safety case which is goal-based but including more

131 At one plant, operators were required to monitor a bulk solids railcar unloading operation. The pneumatic blower and hydraulic vibrator used for the task created a very high noise area around the railcar, requirin g the operator to wear both earplugs and earmuffs as they monitored the unloading operation. It became common practice to “monitor” from farther and farther away from the railcar (to escape the noise), at a d i s t a n c e w e l l i n e x c e s s o f t h e regulatory requirements. An operator’s shed was installed with very effective sound insulation, which (along with the necessary air conditioner!) allowed the operators to closely monitor the process in comfort and safety. A facility which manufactures rock et propellant designed their processing building (in which the propellant was formulated and mixed) with large earthen berms surrounding the building, to absorb the force of any explosions. 7.5 MORE ROBUST PROCESS EQUIPMENT AND DESIGN In parallel with the concept of inherently robust equipment design introduced in Chapter 6, in whic h losses of containment were made virtually impossible, applying the substitution strategy, by using more robust materials of construction or more robust designs can reduce the likelihood of system or component failure, consequently reducing the likelihood of losses of containment, toxic releases, fires, and explosions. In previous examples, higher alloy materials of construction have been shown to provide better corrosion resistance but can also provide increased toughness and fatigue resistance. Improved casing designs for rotating equipment can reduce the likelihood of losses of primary containment in the event of rotating element failure. Pump designs with double mechanical seals are more robu st than single se al pumps. Seal- less pumps greatly reduce the risk of a process fluid leak, but they also introduce new hazards and concerns, such as overheating and internal leakage, which may be very rapi d. Newer seal-less pump designs incorporate advanced thermal protecti on shutdown devices to prevent overheating. Other more robust pump designs, depending on the application, include diaphragm pu mps, jet pumps, and eductors. Using smart transmitters instead of older analog transmitters in a given application make the contro l function more reliable.

262 Human Factors Handbook 21.3 Practical situation awareness tools and tactics Given the importance of situation awareness in many safety critical tasks, advice can be offered on how to minimize the risks of reduced situation awareness. Human performance tools are useful and can be used to reduce human error and lead to various positive outcomes, such as: • Heightened sense of situation awareness concerning safety, presence of error precursors and error traps, task s to be performed, conditions, and surroundings. • More accurate estimates of risk level of activities. • Higher level of self-awareness, including biases, vulnerabilities, deficiencies, and limitations. • The most commonly used human performance tools are shown in Table 21-2. These tools can also help foster agile thinking. For example, the Dynamic Risk Assessment includes reviewing the effect of actions and adapting these.

100 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS 4.2.6 Environmental Health, Safe ty and Security (EHSS) Personnel EHSS personnel are more likely to act as responders to the consequences of abnormal situations that have developed into a loss of primary containment. Events that are more significant typically require a response by emergency teams with outside in dustry assistance and external responding agencies. 4.2.7 Technical Experts Technical experts, also called “subject matter experts” (SMEs) are often located remote from the facility but could be called upon to help to diagnose or respond to an abnormal situation (as discussed in Example Incident 4.5). While it is not possible to consider the whole range of technical expertise that may be involved in such a situation, these individuals could include: Instrument engineer, to explain how a level gauge may not respond to a change in level. Rotating equipment specialist, to help diagnose why a compressor starts to vibrate under abnormal conditions or the rotating equipment vibration profile has changed over time. Quality control engineers/specialists to help troubleshoot contamination and composition of raw and process chemicals. Materials engineer, to help understand the significance of a minor crack that appears in a flange on a pressure vessel. Piping engineer to evaluate the risk associated with the movement of a pipe support. Technology Licensors to provide sp ecific expertise related to the process.

Piping and Instrumentation Diagram Development 376 User User UserUserFigure 17.22 Distribution and c ollection network of a utility. User User UserUserFigure 17.23 Connec ting pipe between distribution and collection networks. DPC Figure 17.24 Connec tion between distribution and collection networks. Steam NPSsCond. line Steam line T NPSc CondensateTFigure 17.25 Connec tion between steam and condensate networks.

PROCESS SAFETY AND MANAGEMENT OF ABNORMAL SITUATIONS 21 Example Incident 2.3 – Texaco Refinery, Milford Haven ( cont. ) Operators missed key information, such as the buildup of liquid in the KO drum, which could have prevented the explosion. However, a previous modification that involved the removal of the automatic pump-out system limited their options. For several hours after the light ning struck, operators were overwhelmed by a flood of 2040 alar ms (at an estimated rate of one every two to three seconds), all of which were designated ‘high’ priority. Many were only informative, but the existing alarm design did not separately prioritize and display safety-critical alarms. Alarms were annunciate d faster than the operators could recognize, acknowledge, and respond to them, including 275 alarms in the final 11 minutes prior to the explosion. Lessons Learned in relation to abnormal situations: 1) For management / engineers: Relief systems should be designed to handle worst-case abnormal situations. Safety critical alarms should be distinguishable from other operational alarms. The number of alarms should be limited to allow effective monitoring by operators. A “fir st-out” alarm system can be a useful design feature in these circumstances. See Chapter 5 on tools and methods. Control panel graphics should include a process overview screen to help with troubleshooting. Modifications (removal of pump-out system) should consider abnormal situations. A plant safety system must be robust enough to handle situations where a human response to an alarm is not enough to mitigate the issue.

2.10 Lear n to Assess and Advance the Culture |69 that constantly refreshes itself. The following paragraphs suggest som e ways to learn and advance the culture. B e adaptable Adaptable organizations continually adopt new and improved ways to do work, and the different units or groups in these organizations often cooperate to create change. Adaptable organizations also demonstrate a strong user/customer focus. This focus on change and improvem ent can also be directed to improving culture. However, without careful review of proposed changes, being adaptable can lead to norm alization of deviance. Therefore, culture im provement efforts should be accompanied with careful review. The organization should define and understand its boundaries of acceptable process safety perform ance. Any variation should keep within these boundaries (Ref 2.3). B e competent Each person whose job addresses process safety in some way should possess the required knowledge and skills to perform this position (Ref 2.33). Clearly, process safety experts need to be competent in their disciplines, but competence does not end there. Process safety com petence applies to everyone in the facility. Design engineers should be com petent in applicable standards. Decision-m akers should understand how to interpret risk assessment data. Operators and m echanics should be able to perform all required tasks and understand the importance of following procedures. Leaders should understand and be able to build process safety culture. B e aware All facility personnel should support the process safety program, whether they have a formal process safety role or not. Each person on the site should be thoroughly

12.8 Embed and Refresh | 167 12.8 Embed and Refresh Three years had passed since Frederik, Pamela, Alexandre, and Reed came up with a plan to address the minor tank overflows. At Reed’s retirement party, Jan said, “First, I want to thank Reed for all his years of service with us. He has been such an important part of our work family. We might not have resolved the overflow incidents a few years ago without his diligence and hard work. We will most certainly miss him.” “Hear, hear!” the crowd cheered. Jan continued, “Second, I’m pleased to announce that we will embark on a new adventure, one that will modernize our terminals. We will put into motion a project to upgrade our tank measuring system with radar technology. Everybody will be critical to the success of this project, and that includes all the new folks, as well as the more seasoned ones, like myself.” Frederik, Pamela, Alexandre, and Reed looked at each and smiled. The plan they had put together three years ago was finally coming to fruition. Alexandre, who Pamela considered her protégé, was going to lead the project. “When I retire, I’ll be leaving the department in Alexandre’s capable hands,” Pamela thought with satisfaction. 12.9 References If so indicated when each incident described in this section was introduced, the incident has been included in the Index of Publicly Evaluated Incidents, presented in the Appendix. Other references are listed below. 12.1 Ilyushina, M. (2020). Putin declares emergency over 20,000-ton diesel spill. CNN (4 June 2020).

36 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Figure 3.1. Lac-Megantic tank cars with breaches to their shells (TSB 2013) Detailed Description The MMA-002 train was traveling from North Dako ta to a refinery in New Brunswick, Canada. The train was made up of 72 cars carrying 7. 7 million liters (2 million gallons) of crude oil (UN1267). Just before midnight on July 5, 2013, the train was parked in Nantes, Quebec, Canada. The 1,433 m (4,700 ft) long train that contained 72 tank cars loaded with crude oil from the Bakken fields in North Dakota. (NTSB 2015) The cars were DOT-111 design. With the fracked crude from primarily Texas and North Dakota, th e U.S. was producing more crude oil than it had in 30 years. Transportation of crude oil by rail had increased significantly to move the crude to refineries for processing. Carloads carrying oil in 2014 rose by more than 5000 percent when compared with 2008 numbers (NCSL 2015). The fracked crude oils from fo rmations such as the Bakken tend to be lighter than other crudes. They are of a lower density, flow fr eely at room temperature, and have a higher proportion of light hydrocarbon fractions result ing in higher API gravities (between 37° and 42°). A Sandia report stated that “No single pa rameter defines the degree of flammability of a fuel; rather, multiple parameters are relevant .” (Sandia 2015) The attention following this incident is continuing to prompt discussion on the safe transport of various classifications of crude oils. The locomotive engineer stopped the train on a downhill grade on the main track. He used the pneumatic brakes, applied the brakes on the locomotive and the buffer car, and began to apply the hand brakes to some cars (but fe wer than recommended in company procedures), and shut down the trailing locomotives. He test ed the hand brake by releasing the locomotive automatic brakes but did not release the locomotive independent brakes. He communicated with the rail traffic controller noting mechanical difficulties he had experienced including excess smoke and a loss of power in the lead engine. They decided to address these issues in the morning. The locomoti ve engineer went off-duty to stay in a Lac-

GLOSSARY 419 qualitative techniques to pinpoi nt weaknesses in the design and operation of facilities that could lead to incidents. Hazard Identification and Risk Analysis (HIRA) — A collective term that encompasses all activities involved in identifying hazard s and evaluating risk at facilities, throughout their life cycle, to make certain that risks to employees, the public, or the enviro nment are consistently controlled within the organization's risk tolerance. High Potential Incident —An event that, under different circumstances, might easily have resulted in a catastrophic loss. Historic Incident Data —Data collected and recorded from past incidents. Human Error —Intended or unintended human action or in action that produces an inappropriate result . Includes actions by designers, operators, engineers, or managers that may contribute to or result in accidents. Human Factors —A discipline concerned wi th designing machines, operations, and work environmen ts so that they match human capabilities, limitations, and needs . Includes any technical work (engineering, procedure writing, worker training, worker selection, etc.) related to the human factor in operator-machine systems. Human Reliability Analysis —A method used to evaluate whether system- required human-actions, tasks, or jo bs will be completed successfully within a required time period. Also used to determine the probability that no extraneous human actions detrimental to the system will be performed. Hypothesis —A supposition or proposed explanation made on the basis of limited evidence as a starting point for further investigation. Impact —A measure of the ultimate loss and harm of a loss event. Impact may be expressed in terms of numbers of injuries and/or fatalities, extent of environmental damage and/or magnitude of losses such as property damage, material loss, lost producti on, market share loss, and recovery costs. Incident —An unusual, unplanned, or unex pected occurrence that either resulted in, or had the potential to result in a process upset with potential process condition excursions beyond operating limits, release of energy or materials, challenges to a protective barrier, or loss of stakeholder confidence in a company’s reputation. Incident Investigation —A systematic approach fo r determining the causes of an incident and developing recomme ndations that a ddress the causes

APPLICATION OF PROCESS SAFETY TO ONSHORE PRODUCTION 91 5.2 ONSHORE PRODUCTION FACI LITIES: RISKS AND KEY PROCESS SAFETY MEASURES The identification processes for risks are generally one of the first activities carried out. The actual processes used are summarized later in Section 5.3.2. 5.2.1 Leak from Production Facilities Risks There are multiple potential leak sources at upstream gas plant facilities, where most of the production is carried out at modera tely high pressures that typically exceed 1000 psi (69 bar). This means that even small holes can cause large leak rates. Typical leak sources include pipes, flan ges, small bore connections, compressor and pump seals, and vessels. Corrosion is often a contributory cause as well fluids contain water and acid gases. Other causes include vibration from compression activity, erosion from sand in the oil, impacts from work activities, or design or construction defects (e.g., in correct gasket materials). A leak of flammable materials can lead to jet fires and/or pool fires. If a vapor cloud forms and is subsequently ignited, a flash fire occurs. If sufficient congestion exists, the flash fire flame can accelerate and cause a vapor cloud explosion. Means to estimate potential outcome hazard zones are provided in the Guidelines for Chemical Process Quantitative Risk Analysis (CCPS, 1999). The actual outcome for a leak depends on the initial condition and how the event progresses. This is shown in Figure 5-2 from IChemE (1996). This is complex and not easy to execute using manual techniques and most companies us e specialized consequence software. As mentioned previously, some onshore production facilities are located inside large buildings, for example on the North Slope of Alaska or North Africa, which allows operators and maintenance personnel to work in a temperate environment which enhances process sa fety. A downside to locating large facilities inside buildings is that leak events, which safely disperse with the wind if outside, can more easily create flammable clouds inside . This increases the potential for flash fires or vapor cloud explosions in conge sted spaces (see box Vapor Cloud Explosion – Short Primer in Section 5.3.2). This h azard is managed by preventing leaks (e.g., mechanical integrity program) or mitigating them if they occur through the provision of gas leak detection with ESD and a more extensive fire detection and suppression system than is typically used in outside locations. Sour gas fields contain H 2S and organic sulfur compounds. In the US, H 2S concentrations are usually less than 4%, but are much higher in wells in several states (e.g., East Texas, Colorado, and Wyoming). In the Middle East, some fields are 50% H 2S. A leak of produced gas at 4% (40,000 ppm) H 2S is a serious toxic hazard as the ERPG-3 is 100 ppm (the conc entration that most people can survive a 1 hour exposure) and concentrations above 500-1000 ppm can be rapidly fatal. This is mitigated to some degree if the vapor produced is buoyant (i.e., dominated by methane content) that tends to disperse any leak upwards, away from ground level. However, natural gas with significant C 3+ and H 2S can be a dense gas and not

xxviii PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION RP Recommended Practice SACHE Safety and Chemical Engineering Education SCAI Safety Controls Alarms and Interlocks SDS Safety Data Sheet SHIB Safety Hazard Information Bulletin SIF Safety Instrumented Function SIL Safety Integrity Level (a s per IEC 61508 / 61511 standards) SIS Safety Instrumented System SOL Safe Operating Limits TEEL Temporary Emergency Exposure Limit THERP Technique for Human Error Rate Prediction TQ Threshold Quantity UFL Upper Flammable Limit U.K. United Kingdom U.S. United States UST Underground Storage Tank VCE Vapor Cloud Explosion

Figure 15-5: Guidelines on shift design

TOOLS AND METHODS FOR MANAGING ABNORMAL SITUATIONS 149 Where possible, learning should be embedded into the operation through hardware, software, or procedur al changes. It is not always easy to identify the occurrence of an abno rmal event, but this can be helped by having a good reporting culture, perhaps assisted by a system of automated metrics reporting as described in Chapter 6. The tools for investigating and lear ning from incidents are available in recognized and available publicatio ns. Therefore, rather than discuss them in this book, some suggested CCPS published reference books on Incident Investigations, Metrics, and Bow Ties are: Guidelines for Investigating Process Safety Incidents , 3rd edition 2019 (CCPS 2019) Guidelines for Integrating Manageme nt Systems and Metrics to Improve Process Safety Performance (CCPS 2016c) Bow Ties in Risk Management (CCPS 2018a) 5.9 CHANGE MANAGEMENT Changes to processes and organizati ons occur for several reasons such as equipment upgrades or failures, process optimizations, design changes, new products, and busin ess impacts. These changes may be very positive when managed thorough ly but may have just the opposite result if not actively managed and fu lly evaluated. Table 5.10 references the management of change procedure for processes as well as organizations.

•Simplification sometimes involves a tradeoff between the complexity of an overall plan t and complexity within one particular piece of equipment. For example, a reactive distillation process for producing methyl acetate requires only three columns and the associated support equipment. The older process required a reactor, an extractor, and eight other columns, along with the associat ed support equipment. The new process is simpler, safer, an d more economical, but the successful operation of the reac tive distillation component is itself more complex and knowledge-intensive (KA). •Redesigned latex reactor cleani ng equipment eliminates the potential for incorrect installation that could result in unwanted reactivity or cross contamination (OK). •Dow eliminated the use of hoses in several hazardous services in favor of hard piped connections (OK). 15.9 ADDITIONAL LITERATURE GIVING EXAMPLES OF INHERENTLY SAFER OPERATIONS •Amyotte, P.R., Goraya, A.U., Hendershot, D.C., and Khan, F.I. (2007). Incorporation of Inherent Safety principles in process safety management. Process Safe ty Progress, 26 (4), 333-346. •CCPS (1993). “Inherently Safer Plants.” Guidelines for Engineering Design for Process Safety (Chapter 2). New York: American Institute of Chemical Engineers. •Commission of the European Comm unity (1997). INSIDE Project and INSET Toolkit. Available for download at www.aeat-safety-and-risk.com/html/inset.html. •Kletz, T.A. (1998). Process Plants - A Handbook for Inherently Safer Design, London, UK: Taylor and Francis. •“Layer of Protection Analysis and Inherently Safer Processes." Process Safety Progress, 18, (4), 214-220, Winter 1999. •'Inherently Safer Approaches to Plant Design', DP Mansfield, 430

Chapter No.: 1 Title Name: <TITLENAME> c09.indd Comp. by: <USER> Date: 25 Feb 2019 Time: 12:23:39 PM Stage: <STAGE> WorkFlow: <WORKFLOW> Page Number: 143 143 Piping and Instrumentation Diagram Development, First Edition. Moe Toghraei. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.Companion website: www.wiley.com/go/Toghraei_PID 9.1 Introduction One important duty of items in process plant is holding and storing process material. Process materials  –  for the purpose of holding or storing  – can be divided into four groups: flowable solids, non‐volatile liquids, volatile liquids, and gases/vapors. Each of these materials may need a specific type of container. 9.2 Selection of Containers The volume of a stored material is also arbitrarily divided into four classes: low volume, medium volume, high volume, and very high volume. As can be seen in Table 9.1, for each type of material, in each class of storage, several options are available. Flowable solids can be simply stored in bags if they are of low volume. If the flowable solid is of medium volume, it could be stored in jumbo bag. And if the volume is high it can be stored in a silo. If the solid is not easily flowable and/or it is of huge volume, it can simply be stored in an open pit or on an open pad in the form of a stockpile. Storing in open pits or an open pad can be done only if the weather doesn’t destroy the solid. For non‐volatile liquids (like water), if they are of low volume they could be stored in tote tanks. Tote tanks are plastic or metallic tanks with a volume of about 1 m 3. Some tote tanks have a volume of 1 m3, some others 1.5 or 1.8 m3. If a liquid of medium volume needs to be stored it could be stored in tanks or vessels. If a non‐volatile liquid of high volume needs to be stored the only available option is a tank. Storing large volume of volatile liquids is done only through using floating roof tanks. A non‐volatile liquid can be stored in ponds or reservoirs, as long as it is safe to do so. For storing volatile liquids, if their volume is low a vessel is available. For medium volume there are two available options of tanks or vessels, and if the volume is high the only available option is a tank. For volatile liquids, if they are supposed to be stored in tanks, it could be a floating roof tank or a fixed tank with a blanketing gas system. To store gas, vapor, or highly volatile liquids in low volume a capsule is the available choice. Capsules are a type of vessel that are ended with spherical heads. If gas, vapor, or highly volatile liquid is of medium volume it can be stored in a vessel and if the volume is high a spheri-cal tank is the only option. Gas in very high volumes can be stored only in underground natural reservoirs. In this chapter we mainly focus on tanks and vessels. The word container can be used for all man‐made enclo-sures in classes of low, medium and high volume. Even though the terms “tank” and “vessel” are sometimes used interchangeably, they are in fact different. Generally, when we are talking about “vessels” it is a container with a high design pressure, while a “tank” refers to containers with a low design pressure, approximately atmospheric pressure. The distinction between low pressure and high pressure is subjective, but usually the division is 15 psig, meaning that if the design pressure of the container is higher than 15 psig it is considered a vessel (or better, a “pressure vessel”), and if the design pressure of a container is below 15 psig, then it is called a tank (or better, an “atmospheric tank”). However, generally the design pressure of tanks is less than 3 psig (approximately 20 kPag or 0.2 barg). From the above explanation it can be realized that usu- ally containers that have medium or high volume cannot be in the form of vessels, because vessels have a higher designed pressure and the thickness of their body is high. Therefore vessels cannot be built with high volumes economically. If a high volume container is built, it is more attractive from an economic point of view for it to be a tank, which has lower design pressure. To store very high volumes of non‐volatile liquids, ponds, or reservoirs can be used. The depth of ponds is generally limited to 4–5 m (Figure 9.1). A reservoir can be considered as larger and deeper versions of ponds. Reservoirs are common for storing water and it is best to use natural terrain as a reservoir so that minimum change and effort is required.9 Containers

RISK MITIGATION 339 Detailed Description The Celanese plant was built in 1952 and produced acetic acid. The unit involved was a liquid phase oxidation (LPO) reactor in which butane was oxidized in the presence of air and a catalyst to make acetic acid and byproducts. This was an ex othermic reaction. The reactor product was sent to several downstream units in the Pampa plant to make products that included acetic acid, acetic anhydride, and me thyl ethyl ketone. The reactor operated at a relatively high temperature and pressure. Fi gure 15.3 is a schematic of the reactor. On November 14, 1987, the reactor was prepared to start up following a shut down the previous day due to a problem in the steam sy stem. Following the normal start-up process, the operators began heating the reactor cont ents. As the reactor approached start-up temperature, an explosion occurred in the ai r sparger inside the reactor. The explosion ruptured the 200 mm (8 in) diameter air piping at two places external to the reactor and one failure occurred internal to the reactor. The fl ammable reactor contents rapidly vaporized to the atmosphere. About 25 to 30 seconds after th e initial explosion, a vapor cloud explosion occurred. The ignition source for the vapor clou d was thought to be the gas boilers that were immediately across the road from the reactor. Extensive property damage occurred in the immediate area and severe damage occurred throughout the plant. Figure 15.4, shows the ca lculated extent of the flammable vapor cloud, extending to the boiler area. Figure 15.3. Schematic of oxidation reactor (Celanese)

4 | 1 Introduction Describing groundbreaking CCPS work in 2005, J ones and Kadri (Ref 1.8) adapted these published definitions to process safety and recognized the link of culture to management: “For process safety management purposes, we propose the following definition for process safety culture: The combination of group values and behaviors that determine the way process safety is managed .” (emphasis added) In the wake of its investigation of a refinery explosion in Texas City, TX, USA, the US Chemical Safety Board (CSB) leveraged the CCPS work J ones and Kadri described (Ref 1.9). CSB recommended that the com pany conduct an independent assessm ent of process safety culture at their five U.S. Refineries and at the Corporate level. The resulting Baker Panel report (Ref 1.10 identified num erous culture gaps and improvement opportunities. They then went on to say, “We are under no illusion that deficiencies in process safety culture, management, or corporate oversight are limited to the company.” This statement proved to motivate many process safety culture improvem ents in refining and chemical com panies globally. Additional study led CCPS to define process safety culture based on the critical role of leadership and management. CCPS’s Vision 20/20 (Ref 1.11) CCPS stated that a committed culture consists of: 1. Felt leadership from senior executives. Felt leadership m eans more than a periodic mention of process safety in speeches and town hall meetings. It means that executives feel a deep personal commitm ent and remain personally involved in process safety activities. 2. Maintaining a sense of vulnerability. 3. Operational discipline, the performance of all tasks correctly every time.

248 INVESTIGATING PROCESS SAFETY INCIDENTS appropriate root cause. ( N ote: In some circumstances, the facts may not allow root causes to be identified with out further investigation.) 3. All branches and sub-branches sh ould be considered because an individual causal factor can ha ve more than one root cause. 4. As each branch is considered, the investigator should ask if there are other root causes associated with that category that are not listed on the tree. The team shou ld ask, “Are there any other causes that anyone has in mind that have not been identified?” (Predefined trees are designed to capture most, but not necessarily all, root causes.) 5. The procedure (steps 2 through 4) is then repeated for each causal factor, in turn. 6. When all the root causes have be en identified from the tree, the investigator should ask why to each one in turn as a test to ensure that they are really under lying root causes. If it is possible to identify a lower level cause, this lower-level ca use should be recorded as the root cause. ( N ote: This is analogous to applying the 5 Whys.) 7. Finally, the investigator should cons ider other generic causes of the incident that are not identified by the predefined tree categories. For example, the investigator should consider the plant operating history. Other incidents may indicate repetitive failures that may indicate generic management system problems. Predefined trees are relatively easy to use and generally require less training and effort to conduct root cause analysis than logic trees. 10.8.2 Example—Environmental Incident The following is an example of the use of a predefined tree to analyze an environmental incident. While the st ructure (number of branches and levels) and terminology of predefined trees vary, this exampl e demonstrates the basic method. During a normal night shift at a process plant, a temporary water treatment unit, operated by cont ract personnel, overheated and released hot, low pH water to one of the plant’s outfalls. This

xxxviii Process Safety Information (PSI) Physical, chemical, and toxicological in formation related to the chemicals, process, and equipment. It is used to document the configuration of a process, its characteristics, its limitations, and as data for process hazard analyses. Process Safety Management (PSM) A management system that is focused on prevention of, preparedness for, mitigation of, response to, and restoration from catastrophic releases of chemicals or energy from a pr ocess associated with a facility. Process Safety Management Systems Comprehensive sets of policies, proc edures, and practices designed to ensure that barriers to episodic incidents are in place, in use, and effective. Protective Action Criteria (PAC) Protective Action Criteria are esse ntial components for planning and response to uncontrolled releases of hazardous chemicals. These criteria, combined with estimates of exposure, provide the information necessary to evaluate chemical release events fo r the purpose of taking appropriate protective actions. Protective Action Criteria includes AEGL, ERPG, and TEEL and is available in 3 levels for over 3100 chemicals. Pool fire The combustion of material evaporating from a layer of liquid at the base of the fire. Reactive chemical A substance that can pose a chemical reactivity hazard by readily oxidizing in air without an ignition source (spontaneously combustible or peroxide forming), initiating or prom oting combustion in other materials (oxidizer), reacting with water, or self-reacting (polymerizing, decomposing or rearranging). Initiation of the reaction can be spontaneous, by energy input such as thermal or mechanical energy, or by catalytic action increasing the reaction rate. Recognized and Generally Accepted Good Engineering Practice (RAGAGEP) A term originally used by OSHA, stem s from the selection and application of appropriate engineering, operating, and maintenance knowledge when designing, operating and maintaining chemical facilities with the purpose of ensuring safety and pr eventing process safety incidents. It involves the application of engineering, operating or maintenance activities derived from engineering knowledge and industry experience based upon the evaluation and analyses of appropriate internal and external standards, applicable codes, technical reports, guidance, or recommended practices or documents of a similar nature. RAGAGEP can be derived from singular or multiple sources and will vary based upon individual facility processes, materi als, service, and other engineering considerations. PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION

A.4 Report References | 225 NPO Association for the Study of Failure (ASF) of Japan Incident Database (Continued) (For incident reports J1–J163: see www.shippai.org/fkd/en/lisen/cat102.html) Code Investigation J112 Explosion in the Piping Caused Due to a Back Flow of Ammonia in the Amination of 5-Chloro-1,2,3-Thiadiazole (1980) J113 Fire Caused By Explosion of P-Nitrophenol Sodium Salt Due to Friction During Transfer With a Conveyor (1979) J114 Explosion and Fire Caused By a Runaway Reaction On Start-Up of the Preparation of an Adhesive Manufacturing Plant (1978) J115 Explosion of DMTP Due to Improper Temperature Control at a Manufacturing Plant of Pesticide Intermediates (1977) J116 Leakage of Toxic Substances at a Chemical Plant (1976) J117 Explosion Due to Heat of Adsorption of an Adsorption Tower Used to Deodorize a Methyl Acrylate Tank (1976) J118 Explosion at a Reactor Caused By a Temperature Rise in a Recovery Process for Hydroxylamine Sulfate at a Pharmaceutical Production Plant (1974) J119 Disaster of Chemical Plant at Flixborough (1974) J120 Spouting of High-Temperature Liquid from the Reactor Due to a Hot Spot Formed On Stopping the Agitator (1974) J121 Rupture of a Vacuum Distillation Drum for 4-Chloro-2-Methylaniline Caused By Air Leakage and Misjudgment (1973) J122 Leakage and Explosion of a Vinyl Chloride Monomer Due to Valve Damage at Distillation Column Feed Piping at a Plant Manufacturing Vinyl Chloride Monomers (1973) J123 Fire Caused By Stopping an Agitator in a Liquid Seal-Type Reactor During a Temporary Shutdown Procedure at an Ethylidene Norbornene Plant (1973) J124 Explosion of 2-Chloropyridine-N-Oxide Left As a Distillation Residue (1973) J125 Explosion Due to Delayed Start of Agitation at the Start-Up of Reaction of O-Nitrochlorobenzene (1973) J126 Runaway Reaction During Manufacturing Pesticide Due to a Decomposition Reaction of Tarry Waste (1973) J127 Explosion Due to Condensation from Miss-Charge of Toluidine Into a Vessel of Diketene (1972)

3.4 Consequences of Not Learning from Incidents | 35 institute systems to transfer the experts’ knowledge into the corporate memory. Assessing Blame Rather than Correcting Root Causes Across industries, we continue to struggle to perform in-depth incident investigations that identify root causes that can be corrected. Investigators often stop once they identify the individual who made the ultimate error and correct the error by punishing the individual. Although punishment could potentially prevent the individual from making the same error, this step doesn’t address systematic process or design problems or work conditions that led to the error. That makes the error likely to happen again. Worse, it may drive others in the organization to hide incidents and near-misses, preventing the company from being able to fix the root cause. The desire to avoid blame may also lead managers to limit the scope of an incident investigation or improvement recommendations. This action may be driven by legal counsel seeking to avoid fines or lawsuits or by other personnel fearing a negative performance review. Misplaced Conservatism One of the ironies of the human condition is that as much as we seek to continuously improve, innovate, and advance technology, we also resist change, even when change is well justified. Both extremes—progressivism and conservatism—can create problems in process safety. In plants and companies with a culture of doing business as usual, opportunities to improve from learning may be unnecessarily blocked. Most of the factors described in this section can lead to gradual degradation of culture and corporate memory. This outcome can only be combatted by seeking continuous improvement. This will help ensure that any institutional knowledge that may have briefly been forgotten will become reinstated within the culture. 3.4 Consequences of Not Learning from Incidents When we don’t learn from incidents, we run a higher risk of repeating them. Learning from incidents is important because human lives are at stake. Incidents can result in fatalities and/or injuries to employees or to the public. They can result in property damage to the plant and the surrounding area.

OVERVIEW OF RISK BASED PROCESS SAFETY 39 3.2.1 Pillar: Commit to Process Safety The aim for the first pillar is to ensure that the foundation for process safety is in place and embedded through out the organization. RBPS Element 1: Process Safety Culture This element describes a positive environment where employees at all levels are committed to process safety. It starts at th e highest levels of the organization and is shared by all. Process safety leaders nurtur e this process. Process safety culture is differentiated from occupational safety culture as it addresses less frequent major incident prevention cultures as well as occupational safety. This element highlights the necessary role of leadership engagement to drive the process. Safety culture should not be thought of as a passive outcome of a specific work environment; rather it is something that is managed and improved. Example Incident: Deepwater Horizon The Deepwater Horizon National Commission identified poor process safety culture as a leading cause of that incident, while simultaneously having a positive occupational safety culture and achievement of excellent performance with traditional safety indicators. High drilling co sts, delays, and a desire to move onto the next task led to poor decision making and discounting of danger signs. Similar issues have been apparent in the downstream industry as well. The US National Academies (2016) reviewed how to strengthen safety culture in the offshore oil and gas industry and they endorsed the BSEE nine characteristics of a positive safety culture. The UK HSE has also addresse d this topic in multiple publications, including Reducing Error and Influencing Behaviour (HSE, 1999). RBPS Application Process Safety Culture provides an overview and suggested means on how to improve process safety culture. RBPS Element 2: Compliance with Standards Organizations should comply with applicable regulations, standards, codes, and other requirements issued by regulators and consensus standards organizations. These requirements may need interpretation and implementation guidance. The element also includes proactive development activities for corporate, consensus, and governmental standards. RBPS Element 3: Process Safety Competency This element addresses skills and resources that a company should have in the right places to manage its process safety hazards . It includes verification that the company collectively has these skills and resources and that this information is applied in succession planning and management of organizational change.

32 The layers of protection can be expensive to build and maintain throughout the life of the process . Initial capital expense, and the costs of operation, training, and maintenance, along with diversion of scarce and valuable technical resources into maintenance and operation of the ac tive layers of protection can add significantly to plant operatin g budgets. In particular, SISs are difficult and expensive to design, procure, and maintain. Also, to maintain the calculated le vel of reliability of SISs, i.e., the safety integrity level (SIL), interval testing schedules and procedures must be diligently followed. For these reasons, it is more advantageous to design or modify the process to be able to inherently withstand the hazard(s) if possible, rather than, as the old adage says, add more “bel ts and suspenders.” For example, it is inherently safer to design the maximum allowable working pressure in a process to be higher than the maximum pressure that can be achieved in the process rather than to add additional active overpressure protection devices s uch as relief valves (although relief valves may be required by applicab le regulations and/or engineering codes and standards). Figure 2.4 shows the relationship between inherently safer designs and layers of protection. This figure depicts the concept of layers of protection (inc luding the inherently safer solutions in the “design” layer) representing how the layers of an onion build upon each other to form the whole (adapted from Ref 2.29 CCPS Metrics) 2.8 INTEGRATING INHERENT SAFETY IN PROCESS RISK MANAGEMENT SYSTEMS How does inherent safety fit into an overall process risk management program? As discussed above, “risk” is defined as “the likelihood that a defined consequence will occur.” Efforts to reduce the risk arising from the operation of a chemical processing facility can be directed toward reducing the likelihood of incidents (incident frequency), reducing the magnitude of the loss, injury or damage should an incident occur (incident consequences), or eliminating the potential consequences altogether. A key engineering risk tool is a management system appropriate for the risks being addressed (i.e., heal th, occupational safety, process

100 | 3 Leadership for Process Safety Culture Within the Organizational Structure Succession Planning Succession planning is an important way to maintain process safety culture and PSMS performance through leadership changes. A good succession plan, supported by a sound organizational m anagem ent of change process, helps maintain com petency, performance, and culture during organizational changes. A succession plan ensures that qualified and motivated employees are ready to take over when a key person leaves the organization. Whether or not the actual successors are known, a succession plan includes experience and competency requirements for potential replacements. Having a succession plan demonstrates to stakeholders that the organization is com mitted m aintaining consistent functioning at all times, including during times of transition. The HR Council (Ref 3.24) offered an example highlighting what can happen without succession planning: A mid-sized organization relied heavily on the corporate memory, skills and experience of a longtime employee. In her final position, she was responsible for office administration including payroll and budget monitoring. During her career, she held many positions and understood well the organization's operations and history. Her unexpected death was both an emotional blow and a wake-up call to her colleagues. Everything she had known about her job was “in her head.” While management discussed regularly the need to document her knowledge to pass it on to others, this had never happened. Ultimately, the organization did regroup and survive the transition, but employees experienced high stress as they struggled to determine what needed to happen when. A great deal of time and effort was spent recreating systems and processes and even then, some things fell through the cracks resulting in the need to rebuild relationships with supporters.

15.4 PROCESS ROUTE SELECTION – EARLY R&D EXAMPLE Inherent safety has been institut ionalized into procedures that researchers must follow when develo ping a process chemistry and/or a process design. Researchers are required to review hazards and document a process hazards analysis (PHA) for each experimental set- up and/or significant change in that set-up. A checklist is a required part of that PHA effort and inherent safety (IS) questions ar e included in the instructions for completing the checklist. The format (template) for technica l reports on product, chemistry, and process development includes a section on IS, as does the report format for applying for permission to seek a patent. The principal instruction on the template for ea ch is similar to the following: “If the process or chemistry being addressed is aimed at the implementation of a new manufacturing process, or improvements/changes to an existing manufacturing process, then a discussion of the anticipated hazard/risk level at the commercial scale must be included. This discussion should be from an IS perspective and include consideration of the quantity of hazardous materials involved, and the severity of process conditions. The use of a standard “index” sheet—a form that dictates IS be considered—is required.” The index sheet (Ref 15.9 Heikkila) is a chart that gives five levels of definition (from low to high) for toxicity, flammability, and reactivity (i.e., the material factor), for quantity (i .e., the quantity factor), and for reaction severity, pressure, tempera ture, potential for corrosion or erosion, dust content, operability, and experience (i.e., the process f a c t o r ) . T h e r e s e a r c h e r i s a s k e d t o a s s i g n a “ l e v e l o f s e v e r i t y ” t o e a c h factor, and to sum them. The higher the resulting number, the more hazardous the chemistry or process is, from an IS standpoin t. Several alternative chemistries and/or processes must be proposed , and the “index” sheet used for them as well. If competing chemistr ies and/or processes exhibit a lower total “level of severity,” the research er is obligated to defend his choice. No chemistry that exhibits "severe" factors in all categories is accepted. 411

232 | 6 Where do you Start? com pletely in one effort. Try to make improvem ents in each core principle one or two at a time. When selecting core principles to address, it may be helpful to address those at the beginning of the list first. The Nature of Management System s and Documentation Models Section 4.4 discussed the institutionalization of PSM S via centralized and decentralized organizational models. B oth have advantages and disadvantages. If the nature of the PSM S is found to contribute to process safety culture issues, decide whether the m odel needs to be more centralized or more decentralized. The m any types of documentation required by the PSMS serve a critical role in assuring PSM S performance. However, when documentation requirements are redundant, use software that is not user-friendly, or appears to not have a purpose, the stage is set for a check-the-box mentality and the normalization of deviance . Often, carefully designed docum entation systems can m ake documentation easier. Involving the users of the documentation in the design process can also help in building the culture. Com munication Communication break-downs between silos rank high am ong the many comm unication barriers discussed in section 2.4. Considering the multi-functional nature of PSMSs, connecting silos is essential to help inform ation flow more freely between groups and individuals. This also helps reinforce the key point that the process safety culture and PSMS requires full participation and integration. B reaking down silos can be accelerated by getting workers in one group to be interested in and fam iliar with the PSMS elements their co-workers in another group have responsibility for. This can lead to m utual appreciation about each other’s roles and

120 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Figure 7.1. Damage to Concept Sciences Hanover facility Courtesy Tom Volk, The Morning Call (CSB 2002) CSI developed a four step process to make 50% HA: 1. Reaction of HA sulfate and potassium hydroxide to produce a 30 wt.% HA and potassium sulfate aqueous slurry: 2. Filtration of the slurry to remove precipitated potassium sulfate solids. 3. Vacuum distillation of HA from the 30 wt . % solution to separate it from the dissolved potassium sulfate and produce a 50 wt. % HA distillate. 4. Purification of the distillate through ion exchange cylinders. The distillation process is shown in Figure 7.2. The charge tank was a 2,500 gallon (9.5 m3) tank. In the first step of the distillation, a pump circulated 30 wt. % HA to the heating column, which is a vertical shell and tube heat exchanger. The HA was heated under vacuum by 49 °C (120 °F) water. Vapor was drawn off to the condenser and collected in the forerun tank and concentrated HA was returned to the charge tank. When the concentration in the forerun tank reached 10 wt. %, it was then collected in the final product tank. At the end of the first step of the distillation, the HA in the charge tank was at 80 - 90 wt. % HA. At that point, the charge tank was supposed to be rinsed with 30 wt. % HA. The first distillation done by CSI began on Monday afternoon, February 15, 1999. By Tuesday evening, the HA in the charge tank wa s approximately 48 wt. %. At that time, the process was shut down for maintenance after it was discovered that water had leaked into the charge tank through broken tubes in the heater column. The distillation restarted on Thursday

200 Carriers and shippers have worked together to improve transportation safety. The American Association of Railroads has agreed to designate routes that handle 10,000 loads per year or more of chemicals as “Key Routes.” Key Routes receive upgraded track, enhanced equipment to detect flaws in equi pment or trackage, and lower speed limits. “Just-in-time” supply of materi als may affect the mode of transportation and could actually incr ease associated risks. For example, if drums of a chemical are stockpile d near a user, the material is not under the same level of control that could be provided by either the supplier or user if the inventory was maintained in a storage tank at one or the other facility. This type of risk should be included when contemplating “just-in-time” shipments. 8.10.4 Improved Transportation Containers Transportation risk can be reduced by applying inherently safer design principles to transportation cont ainers. Some examples of design improvements are as follows: The shipment of environmentally-sensitive materials in general purpose rail cars has been phas ed out, and DOT specification 105 pressure cars are used instead (Substitution ). Thermal insulation can be used to maintain lower temperatures in the containers and to provide improved protection from fire (Moderation ). Baffles and subdivided internal barriers reduce the amount of chemical that can be released if only one compartment is breached ( Minimization ). Non-brittle containers can be us ed to improve resistance to impact or shock damage. In addition, transportation cont ainers can and have been made more inherently robust and enhanced l ay er s o f p r ote cti on h a ve b e en added to them as follows: Rail car design, particularly for certain cargoes, has been enhanced so that they can with stand more serious and energetic crashes and derailments without breaching. For example, the

118 Human Factors Handbook Competency includes not only the applic ation of technical skills and knowledge, but non-technical skills such as communica tion; this is important for supervisory roles. 10.3 Competency Management Competency Management refers to methods used to categorize and track the development of employee’s competency. Th is allows the organization to track progress and identify training needs. Clear competency performance standards should be developed. People should also receive appropriate training and devel opment opportunities, to maintain their competency over time. Key steps in ac hieving process safety competency are shown in Figure 10-1. The five phases follow: • Phase 1: Establish Requiremen ts – Determine competency Before starting Competency Managemen t, it is important to identify activities that may affect operational, and/or occupational safety. Risk assessments are especially useful here, as they can identify safety critical tasks (see Chapter 6 for advice on Safety Critical Task Analysis). All the requirements for the position should be identified. As a priority, the focus should be on safety critical task and required competency, followed by all other tasks’ competency requirements. The next step is to identify and define operational competency and performance standards for these tasks. • Phase 2: Identify Learning Requirements This phase includes assessment of the gap between individual and team competency, and the competency that individuals need to develop. It is then necessary to identify the learning required to bridge the gaps, including the type of learning most suitable to develop and demonstrate the competency. This includes providing a description of learning objectives, to aid development of the on-the-job learning and training programs in Phase 3. • Phase 3: Develop Competency Relevant learning and coaching programs are designed to target competency gaps and fulfil the learning objectives defined in Phase 2. These programs are provided to selected individuals.

Piping and Instrumentation Diagram Development 198 (wet gas) is higher than for liquids with non‐tolerable gas content. ●As the gas movers are large, their moving elements may have huge momentum and in the cases of a quick trip in the system, they may move for a while even after shutdown. Therefore there should be some systems to take care of gas movers in those situations. Some of the required actions could be implemented as part of the interlock system. Interlock systems will be discussed in Chapter 16. ●If the to‐be‐pressurized gas is flammable, more cau-tion should be taken in design. This is mainly because by pressurizing a gas, it will be hot. ●Compressors may be placed in an enclosed or semi‐enclosed building for different reasons. One reason could be the sound level around the com-pressor. The other reason is to provide a dedicated space for compressors for ease of inspection by operators. The “compressor building” could be a multi‐story structure for the ease of inspection and maintenance. 10.8.3 Gas M over Drives The same drives of centrifugal pumps are available for gas movers too. Electric motors, turbines (steam driven and gas turbines), and internal combustion motors are some of them.10.8.4 Auxiliar y Systems Around Fluid Movers Auxiliary systems around fluid movers could be support - ing systems for different components of the fluid mover. A fluid mover consists of drive section and fluid mover section. A “connection” attaches these components together (Figure 10.40). The “connection” could be shaft, gear box, crankshaft or a combination. Each of the above components may need some type of auxiliary system. Some of these auxiliary systems are listed in Table 10.12. The seal type for the gas movers could be the same type of seal types for liquid movers. The liquid mover seal type was discussed before and is a “wet type sealing. ” For a long time it was assumed that the only type of seal for compressors are the wet type. But later it was recog-nized that even a gas could be used as a sealing fluid. Then a newer type of sealing system, the dry gas seal, came on the scene. In a dry gas sealing system an inert gas like nitrogen gas can be used as the sealing gas. Drive Connection Fluid mo ver Figure 10.40 Fluid mo ver and drive as a pair. Table 10.12 Fluid mo ver auxiliary systems. Components Type Required auxiliary system Drive Electric motor ●Bearings may need cooling ●The majority of motor blocks are generally air cooled and no specific system is required. However some of them need forced air cooling and air should be cooled (by air or water) and recirculated. Some huge electric motors could be oil cooled, then a cooling system is needed Internal combustion engine (diesel, gas) ●Pistons need lubrication ●The motor block needs cooling Air operated drive ●No specific auxiliary system is required. Only plant air should be piped Solenoid operated drive ●No specific auxiliary system is required. Only DC electricity is required Steam turbine ●The block may need cooling Gas turbine ●The block needs cooling Connection Shaft ●The penetration point should be sealed, lubricated, cooled, and flushed Gear ●Gear box may need lubrication Crankshaft ●Crankshaft may need lubrication Fluid moverLiquid mover ●Bearings may need cooling ●(As mentioned) penetration points should be sealed, cooled, and flushed ●In compressors the block should be kept cooled but this is done by the intercooler shown on the main process P&ID and not the auxiliary P&IDGas mover

98 | 7 Keeping Learning Fresh You could also create your own learning exercise keyed to this style of communication using readily available short videos, for example the CSB video about the Illiopolis, IL, USA incident. Show the video (CSB 2007) and then ask participants to break out into groups to answer these questions: • Why did the operator use the emergency air hose without thinking? • Why did the supervisor run upstairs to stop the release rather than evacuating? • For extra credit: Why was a hose clearly labeled “Emergency Air” provided under each reactor? • Based on your answers to these questions, how would you ensure your team acted properly? During the report-back time, steer the conversation toward the underlying issues, including normalization of deviance, conduct of operations, human factors, and where in the PHA this scenario might have been captured. Here’s another way to communicate in a logical-mathematical style about the proper response to person-down incidents. Create a bulletin, poster, video, or presentation showing that the incidence of health conditions like heart disease and diabetes is lower among workers in the process industries than in other industries. If possible, compare your company’s own statistics for these conditions to the rate in the general population. Then ask, “In a facility like ours that handles toxic materials and asphyxiant gases, why would the person have collapsed on the floor? What should you do?” 7.5 Kinesthetic Intelligence People with kinesthetic intelligence learn best when they can have hands-on learning experiences. To support this kind of learning, aspects of some incidents can be simulated under safe conditions: for example, controlled demonstrations of chemical reactivity, vapor cloud, and dust explosions. Simulations are especially effective when observers can feel some heat from the fire or feel and hear the pressure wave from the explosion. Computer simulations can also be effective learning tools, although it is harder to simulate the heat or pressure wave. Role-playing is another a beneficial way to learn, especially if a moderator can supply stimuli and sound effects that help the participants imagine what an explosion or fire feels like. Another aspect of kinesthetic learning is learning while moving. After a major incident occurred in its facility, one company set up a walking path with

EDUCATION FOR MANAGING ABNORMAL SITUATIONS 103 This type of arrangement works we ll for HAZOP studies and has the added benefit of providing feedback to designers and engineers who are not on the front-line and may not have a full appreciation of potential issues with the equipment they are providing. A common criticism offered by operating personnel is that the designers do not have to run the process and therefore sometimes fail to have an appreciation of some of the problems that certain desi gn features can create. Traditional training and learning processes are normally targeted at teaching personnel to operate the process per standard operating procedures (SOPs), step-by-step tasks or checklists, and using the computerized process control system. Th e training tools that are used are typically very structured to ensure consistency in how the training information is communicated and how und erstanding is verified. Although these types of tools have been demonstrated to work well, they may not be the optimum tools for educating for abnormal situations. These matters are discussed further in Chapter 5. No matter the training objective, logistics of the training session including time, setting, frequency, and number of personnel on shift, as well as type of training, (e.g., Classr oom / Computer/ Desktop / One-on-One) should be considered in advance. Management of the training should include a system for enforcement of the training and metrics to measure the effectiveness of the program, as discussed in 6.2. The next section describes some of the traini ng topics and structure. 4.3.2 Structure of Training Topics 4.3.2.1 Basic Process Operations For front line control room operators and field operators, an introductory understanding of the process chemistr y, potential chemical interactions, relationships between temperature, pressure, flow, and level as they relate to process operations are essential to building an operator’s knowledge and the ability to recognize and respond to abnormal situations. Training in these basic relationships can be accomplished using basic process simulation software that typically runs on a desktop personal computer. Additionally, the training may be accomplished in informal training sessions with senior plant personnel or in formal training with chemical process instructors.

208 INVESTIGATING PROCESS SAFETY INCIDENTS concerns had grown to the point that further launches were postponed until an attempt was made to remedy the situation. But these remedies were in effective and did not deal with the causal factors or root causes of the problem joint. Despite all that was known about the O-ring problem, a decision was made to launch the Cha llenger on a cold January morning with deva stating consequences. The Challenger space shuttle disaster is an excellent example of the principle that apparently simple mechanical problems are related to more complex underlying causes rooted in management systems. The recommendations submitted by the presidential commission focused on root causes. These involved changes in management systems that would not only fix the ring joint problem, but also the systems, procedures, and overall appr oaches to identifying, evaluating, resolving, monitoring, and auditing safety-related concerns. 10.3 M ETHODOLOGIES FOR ROOT CAUSE ANALYSIS 10.3.1 5 W hys Technique The 5 Whys is a simple methodology for identifying root caus es that involves repeatedly asking the question “why ?”. The methodology is easy to understand and perform, and the technique adds so me structure to group brainstorming. Large quantities of in formation and data are not necessary (although useful for complex process safety incidents), and therefore the technique is suitable for minor incidents, especially those involving human factors and interactions. The 5 Whys is also widely used as an integral part of Kaizen, Lean Manufacturing, and the Six Sigma methodology [e.g., the Analyze phase of Six Sigma DMAIC (Def ine, Measure, Analyze, Improve, Control) ]. Although called 5 Whys, five is only a rule of thumb, and sometimes the investigation team will ask “why?” more or fewer than five times. The technique requires that th e investigation team asks “why?” a negative event occurred or undesirable condition existed (i.e., causal factors), and then asks “why?” enough times to reach a management system deficiency. The process is repeated until all the causal factors ha ve been considered. In essence, this is analogous to a logic tree approach wi thout actually draw ing the logic tree.

100 PROCESS SAFETY IN UPSTREAM OIL & GAS Smaller onshore oil facilities tend to fo llow API and related standards (listed above), and either RBPS or the OSHA PSM regulation (although they may be exempt formally) for structuring thei r process safety management system. Regulations An overview of regulations was provided in Section 2.8. Specific to onshore production, OSHA and EPA share responsib ility in the US for onshore regulations for facilities exceeding nominated invent ory thresholds. Similarly, PHMSA regulates pipelines. OSHA PSM has a focus on worker safety, while EPA RMP (Risk Management Program) has a focus on offsite / public safety and environmental incidents. OSHA does not require a single document collecting all elements of the PSM program, just that all elements must be present. Conversely, the EPA does require a risk management program document. For facilities in Europe and parts of Austra lia (e.g., Victoria), a safety case must be prepared. The EU regulation is known as the Seveso Directive, originally passed in 1986, and updated twice to address serious incidents that were not well covered. This regulation was covered in Section 2.8. For larger onshore production facilities that exceed regulated thresholds, both the EPA RMP and the safety case (now known as safety report) requirements in the EU Seveso III Directive require that important parts of the major hazard plan be documented. This includes a listing of hazardous materials, the facility description, hazard identification and risk ranking, th e management system, and the emergency response plan. In the US, most states ha ve regulations for production facilities. Both RMP and Safety Reports must address emergency response and this topic is outlined in Section 5.3.4. Company Practices Most companies have documented practices and procedures that describe how the company applies regulations, standards, and lessons learned. Additionally, large companies have their own internal guidelin es for engineering and process safety. 5.3.2 Hazard identification and Risk Analysis The Hazard Identification and Risk Analysis approaches described in Chapter 4 also apply to onshore production. Refer to Chapter 4 for further details. The main methods used include What-If, the What-If Checklist, and HAZOP (Hazard and Operability Study). What-If What-If examines operations for possible deviations to see what the consequences might be and what safeguards are employed. The safeguards listed may be barriers, barrier elements or degradation controls as discussed in the bow tie methodology (CCPS, 2018c). What-If or the What-If Chec klist can be employed early in a process

54 Guidelines for Revalidating a Process Hazard Analysis • Are causes meaningful? A cause such as “opera tor error” for a “High level” deviation is not meaningful to anyone outside of the PHA team, and its specific meaning will probably be forgotten by the PHA team members themselves a few days after the PHA meetings are complete. A more meaningful statement of the cause might read, “operator fails to remain with the tank during the fill step and stop the flow at the ‘full’ mark,” or “operator fails to stop the flow at the ‘full’ mark due to being distracted by another process alarm.” It is easier to understand and revalidate causes that are more specific. • Do the stated consequences pr esume failure of all safeguards? Taking credit for safeguards in the consequence definition is a common PHA flaw. If this error is pe rsistent throughout the PHA, it is a sign the PHA team did not accurately document and discuss worst credible consequences in th e PHA. Consider the possibility of high pressure in a vessel. If “process shutdown” was listed as the consequence, the prior PHA team may have assumed the high- pressure interlock would shut th e process down as designed and disregarded the potential for cata strophic vessel overpressure. If “discharge to the flare” was liste d as the consequences, the prior PHA team likely assumed the pres sure relief valve worked as intended to protect the vessel. PHA teams may be willing to accept the risk of a particular outcome based on some (but not all) of the safeguards working, but if a qualitatively worse outcome went unrecognized, risk may not be managed as intended. In the case of vessel overpressure, the unacknowledged consequences are that the vessel might rupture (if the maximum allowable working pressure [MAWP] exceedance is great enough), and the prior PHA team did not document its implicit acceptance of that risk. • If a consequence of interest might result from successful operation of a safeguard, is it addressed as well? The March 2005 explosion at a Texas City refinery [34] did not involve failure of all safeguards. As the pressure in a distillation column rose, the pressure safety valves functioned as intended, discharging flammable liquid above its normal boiling point to the atmospheric blowdown stack and releasing a large flammable vapor cloud. Ignition of the cloud led to an explosion that killed 15 and injured 180 at the refinery. Similarly, the fire protection systems worked as designed to suppress a 1986 agrichemical warehouse fire in Switzerland. The firewater runoff carried highly hazardous chemicals into the nearby Rhine River and poisoned aquatic life downstream to the North Sea [35]. Thus, it is important that the prior PHA team also considered scenarios, particularly those

PROCESS SAFETY AND MANAGEMENT OF ABNORMAL SITUATIONS 19 Example Incident 2.2 – Bayer Crop Science Plant, ( cont. ) Lessons Learned in relation to abnormal situations: 1) For management / engineers: Overly complex operating procedures Inadequate operator training on the newly installed DCS Temporary changes not evaluated Malfunctioning or missing equipment (faulty new relay caused both centrifuges to trip, solven t drip line was missing a valve) Insufficient technical expertise available in the control room during the restart Operational readiness review was inadequate 2) For supervisors / operators / technicians: Deviation from written operating procedure (several required steps were not completed) Safety-critical equipment bypassed (interlocks on residue treater) or not operable (toxic gas monitoring system not in service) Misaligned valves Example Incident 2.1 an d Example Incident 2.2 – (Broadribb, CSB) involved transient operations of startup and shutdown. However, abnormal situations can also occur during normal operation, as Example Incident 2.3 (HSE 1997) illustrates . More details of Example Incident 2.3 are also provided in Chapter 7.2.

Piping and Instrumentation Diagram Development 230 12.15 Deciding on an Emergency Release C ollecting Network One important issue regarding P/VRDs is the final destina- tion of the released fluid. An emergency release collecting network could be used to collect the instantaneous releases from PSDs and direct them toward the disposal system. Figure 12.24 shows a simple schematic of an emergenc y release collection network. The emergency network has several requirements, which are shown in Figure 12.25. As is visible from Figure 12.25, the main header should be sloped toward the disposal system. If there is needed PGor “pressure switch to control room Burst sensor PAH SET @ 146 PSIg(d)(c)(b)(a) 123 Figure 12.19 Combina tion of safety valves and rupture disks. Figure 12.20 Combina tion of safety valves and rupture disks.PG Figure 12.21 A combined solution t o deal with leakage of the rupture disk upstream of the PSV. Utility water Flush ring Figure 12.22 A PSV with a flush ring . Figure 12.23 Rupture disk upstr eam and downstream of a PSV on the outlet of a PD pump.

7.2 Sustainability of Process Safety Culture |241 The last point may seem circular, but reflects a key tenet of sustainability. Sustainability must be intentional. That is why “Learn and advance the culture ” is one of the core principles of process safety culture. More generally, each of the core principles is required for a strong process safety culture to endure. 7.2 SUSTAIN ABILITY OF PROCESS SAFETY CULTURE Process safety culture, like any culture can degrade quickly without comm itted effort to sustain it. Almost any event, good or bad, can create conditions that unravel previous efforts. The following exam ples describe events that can degrade culture, how this could happen, and what leaders can do to sustain the culture. Serious process safety incidents Process safety incidents with severe consequences can represent a crossroads event in the life of an organization. In a strong or improving culture, leaders take the opportunity to re- exam ine the process safety culture and PSMS, learn and apply the lessons-learned broadly across the company, and re-commit to process safety. However, in a weaker or degrading culture, managem ent m ay turn the investigation to finger-pointing and a search for a scapegoat. In response to regulatory and public pressure, the com pany may seek a legal settlement. While this is a norm al practice, a weaker culture will treat the settlem ent as evidence that the causes of the incident have been resolved. Such a settlement would not deter a stronger culture from seeking improvement. Incidents can som etimes be caused or contributed to by an individual who takes actions that are forbidden by company policy (e.g., violating lock-out/tag-out). The investigation team should determ ine if it was only the one person breaking the rules, or part of a pattern where policy violations are common. If the policy violation was an unusual event, the com pany should not be

9.2 Seek Learnings | 123 first start-up following the replacement of spent catalysts with an improved catalyst that the plant had not previously used. Before introducing MPK, the catalyst bed was heated by circulating ethylbenzene through the reactor and an external heat exchanger. The company did not know that the new catalyst could react with ethylbenzene at temperatures that could typically be reached during start-up due to normal temperature fluctuations. Due to the reactor fluctuations, the liquid level in the vapor-liquid separation tank on the flare line also fluctuated widely. Every time the tank reached high level, an interlock would close the valve between the tank and the flare and had to be manually re-opened by an operator when the liquid level dropped. Just before the runaway, an operator had neglected to re-open the valve to the flare. A thermal runaway occurred. With the valve to the flare left closed, the resulting pressure could not be relieved, leading to the explosion. Although Jason was not able to find other public cases of runaway reactions involving catalyst pellets, he did find several other cases where unexpected runaway reactions occurred: Oita, Japan, 1996 During a trial batch of a new pesticide, an intermediate was held at high temperature during a process delay due to a pump failure. The intermediate self-reacted exothermically, leading to an explosion that injured an operator and damaged the production building. Kitakyushu, Japan, 1996 A contaminant present when raw material was fed to a resin intermediate process led to an explosive decomposition of the raw material. The contaminant back flowed from the vapor treatment system. While no one was injured, the plant was destroyed and did not restart. The contaminant had entered the vapor treatment system from another part of the process. See Appendix index entry J30 See Appendix index entry J35

256 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Table 12.5 continued Process Phase Example Objectives Hazard Analysis Technique Pilot plant Identify ways for hazardous materials to be released to the environment. Identify ways to deactivate the catalyst. Identify potentially hazardous operator interfaces. Identify ways to minimize hazardous wastes. Checklist Preliminary Hazard Identification (HAZID) Analysis What-If What-If / Checklist Hazard and Operability Study Failure Modes and Effects Analysis Fault Tree Analysis Event Tree Analysis Detailed engineering Identify ways for a flammable mixture to form inside process equipment. Identify how a loss of containment might occur. Identify which process control malfunctions will cause runaway reactions. Identify ways to reduce hazardous material inventories. Evaluate whether designed safeguards are adequate to control process risks to tolerable, required or as low as reasonably practical (ALARP) level. Identify safety-criti cal equipment that must be regularly tested, inspected, or maintained. What-If What-If / Checklist Hazard and Operability Study Failure Modes and Effects Analysis Fault Tree Analysis Event Tree Analysis Construction and start-up Identify error-likely situations in the start-up and oper ating procedures. Verify that all issues from previous hazard evaluations were resolved satisfactorily and that no new issues were introduced. Identify hazards that adjacent units may create for construction and maintenance workers. Identify hazards associated with vessel cleaning procedures. Identify any discrepancies between as- built equipment and the design drawings. Checklist What-If What-If / Checklist Critical Task Analysis

PROJECT DESIGN BASICS 173 Figure 10.9 Inherently safer design principles (CCPS 2019 b) The “simplify” principle addresses human fact ors which will be discussed in Chapter 16. Humans make mistakes, even with the best kn owledge and skills. Incorporating aspects in an engineering design to enable successful human performance will support good process safety performance. These four techniques make ISD sound simple, bu t in reality, it is a bit more complex. “A technology can only be described as inherently saf er when compared to a different technology…A technology may be inherently safe r than another with respect to some hazards but inherently less safe with respect to othe rs…” (DHS 2010) For example, minimizing the quantity of a hazardous material stored at a fa cility may reduce the fac ility risk; however, it may necessitate increased transp ortation of that hazardous ma terial to supply the facility resulting in an increased risk along the transportation route. ISD is applicable through the life cycle stages and includes manufacturing, transportation, storage, processing, and decommissioning. The greatest opportunity to implement ISD is during the design phases where the implementa tion of an ISD approach only involves a changing ink on an engineering drawing as opposed to modifying steel in a facility. ISD is iterative during engineering design de velopment. During conc eptual phases, major decisions are made such as use of an inhere ntly safer technology or choosing to export product by ship versus truck. As the design details are progressed, the size of vessels is determined, and the control systems are develo ped. In an operating facility, operating procedures can be provided in a clear and concis e way with visual cues that makes them easy to understand and use. Hierarchy of Controls Hierarchy of controls is similar, in concept, to ISD and is used in many industries. It depicts hazard controls in order of decreasing effect iveness as shown in Figure 10.10. The first two levels in the figure are ISD principles and ar e typically more easily implemented during the early stages of a project prior to the detailed equipment design. The remaining levels reflect that “It is unlikely that any technology will be ‘i nherently safer’ with respect to all hazards, and other approaches will be required to manage the full range of hazards and risks.” (DHS 2010) When the engineering design is finalized or the project is completed, administrative controls are easier to implement.

164 Guidelines for Revalidating a Process Hazard Analysis scenarios for initial and refresher training activities. Companies often keep current electronic copies of their PHAs on their intranet to make this information conveniently accessible to all employees . When the PHA revalidation report is issued, an electronic link can be easily sent to everyone who works in that process area, those who might be affected by its hazards, and anyone else who needs to know the revised PHA is available. For the benefit of the next revalidation team, the PHA files could include a record, such as an end-of-PHA checklist, identifying what information was used, how it was used, and where it is stored (if not with the PHA itself). If a third-party facilitator was used, a copy of their digita l files should also be retained in the PHA files. Finally, it is important that this valu able information be protected. Company policies or procedures often establish retention, redundancy, and diversity requirements for the PHA-related record s. Regulations may impose additional requirements on some facilities. For ex ample, in the United States, companies with processes covered by the OSHA PSM or EPA RMP regulations must keep the initial PHA and all subsequent PHA revalidation reports, as well as documentation of the resolution of re commendations from these reports, for the life of the process. Covered facilities should ensure that archival copies of all versions of the PHA documentation are maintained, regardless of the documentation option used for prior PHA reports. The “life of the process” may span de cades, so the archive copies should resist both physical loss (e.g., flood, fire, theft) and technological obsolescence (e.g., outdated software, inaccessible storage media). When a process unit is dismantled or demolished, its productive “life” has ended, and the PHA records can be discarded or retained in accordan ce with company policy. However, if the process is merely shut down, decommissio ned, or “mothballed,” (even if they have no plans to resume production), it is prudent to retain the PHA records. Circumstances change, and years later there may unexpectedly be a need to resume production. If the PHA records were kept, revalidation would be an option; if not, an entirely new PHA would have to be completed before production could resume. Even in ca ses where the process equipment is ultimately dismantled and removed, access to PHAs could be of value to anyone evaluating the hazards of remediation effo rts, which might be needed to allow new construction at the site.

342 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Safeguards, Barriers, IPLs, and Other Layers of Protection The concept of layers of protection was discussed in Section 14.7, Layer of Protection Analysis, and illustrated in Figure 14.10. Over the years, there have been many terms used to describe these layers including safeguards, barriers, and independent protection layers. Although these terms sound similar, they do have different a ttributes as defined below and in Figure 15.5. Layers of protection - Independent devices, systems, or actions which reduce the likelihood and severity of an undesired event. (CCPS Glossary) Safeguard - Any device, system, or action that interrupts the chain of events following an initiating event or that mitigates the consequences. (CCPS Glossary) Barrier - A control measure or grouping of control elements that on its own can prevent a threat developing into a top event (prevention barrier) or can mitigate the consequences of a top event once it has occurred (mitigation barrier). A barrier must be effective, independent, and auditable. (CCPS Glossary) Independent Protection Layer (IPL) - A device, system, or action that is capable of preventing a scenario from proceeding to the undesired consequence without being adversely affected by the initiating event or the action of any other protection layer associated with the scenario. Note: There are specific functional cr iteria for protection layers that are designated as "independent." A protection layer meets the requirements of being an IPL when it is designed and managed to achieve the following seven core attributes: Independent; Functional; Integrity; Reliable; Va lidated, Maintained and Audited; Access Security; and Management of Change. (CCPS Glossary) The term safeguard is a generic, all-encompassi ng term that is frequently used in HAZOPs to capture all the potential layers of protection . The term barrier is more specific in that a barrier prevents the incident or mitigates the co nsequences on its own. The term is used in Bow Tie Analysis and this specificity supports the integrity of Bow Ties. The term IPL was introduced in Section 14.7. It is even more prec ise which is required to support the integrity of a Layer of Protection Analysis. Many devices, systems, and actions are used to reduce process safety risk and can be described by these terms. These are discussed in Section 15.4.

182 | 5 Aligning Culture with PSMS Elements com pany. Regulations are also considered by this element. While not strictly standards, process safety regulations tend to be developed and written as if they are standards, and some are even referred to by that nam e. Espousing a strong system of internal and industry standards provides a framework for correct design, proper installation, and effective inspection, testing and m aintenance. This is indicative of positive safety cultures. The absence of standards, or inconsistent application of standards is conversely an indicator of a weak culture. If equipment is not designed to a form al code or standard, how will you know it will m eet the process dem ands or how to properly insect and test the equipment? B y determining which standards are applicable and understanding how they work, com panies can leverage experience in designing and operating processes and m anagement systems, while complying with applicable regulations. Companies should keep abreast of new developments in standards and regulations, and address changes appropriately in its technology and m anagement systems. Additionally, standards developed for different industry sectors or even countries should be considered if they address challenges in the facility’s sector. The American Petroleum Institute’s RP-755 addressing fatigue management is an excellent exam ple of a standard with broad usefulness outside the US petroleum sector. Establishing and maintaining internal corporate standards is an effective way to keep abreast with developments in standards and interpret these standards applicable to com pany technology and culture. Some companies will implement internal standards at the facility level to explicitly address local and national standards. Others will strive to have one set of corporate standards that applies regardless of location. This is largely a m atter of preference. The im portant thing is that it is considered part of the imperative for process safety to identify, understand, and im plement the applicable standards.

495 Table E.3 continued Notes: 1. It is recognized that threshold quantities given in kg and lb or in lb and bbl are not exactly equivalent. Companies should select one of the pair and use it consistently for all recordkeeping activities. 2. Refer to guidance on selecting the correct Threshol d Release Category and the use of Material Hazard Classification Option 1 and Option 2. Table E.4. Examples fo r material categories Category Example Material 1 Br, HCN, Phosgene 2 BF 3, Chlorine, H 2S 3 HCl, HF, SO 2 4 Ammonia, CO, EO 5 Acetylene, Ethylene, Vinyl Acetate Monomer, Aluminum Alkyls 6 Vinyl Acetate Monomer, Benzene, Cyclohexane 7 Diesel, Mineral Oil, Muriatic Acid Nonflammable/nonpoisonous gases E.4 Classifying PSE Tier 1 and Tier 2 Events The flowchart in Figure E.1 can be used to classify an LOPC as a PSE Tier 1 or Tier 2. APPENDIX E - CLASSIFYING PROCESS SAFETY EVENTS USING API RP 754 3RD EDITION

202 | Appendix: Index of Publicly Evaluated Incidents Section 1. RBPS Elements (Continued) MOC—Primary Findings A1, A2, A4, A6, A7, A10 C22, C22, C25, C31, C36, C47, C50, C51, C61, C63, C70, C72, C76 D9 HB5 J12, J24, J45, J55, J72, J86, J90, J94, J102, J106, J119, J132, J139, J167, J214, J240, J255, J260, J264 S4, S7, S16, S17 MOC—Secondary Findings C3, C7, C8, C62 J10, J14, J16, J26, J91, J111, J154, J205, J238, J239, J246, J250 S13, S14 Operational Readiness/PSSR—Primary Findings C7, C11, C70, C76 HA10 J26, J40, J69, J90, J111, J146, J171, J174, J177, J178, J179, J183, J184, J188, J189, J192, J213, J243, J246, J249, J263, J269 Operational Readiness/PSSR—Secondary Findings D9 HA3, HA6, HB4 J23, J62, J103, J124, J147, J170, J185, J210, J212, J216, J217, J251 S5 Conduct of Operations and Operational Discipline—Primary Findings A2, A5, A10 C3, C11, C12, C18, C26, C43, C50, C57, C58 D9 J2, J19, J28, J38, J49, J50, J51, J52, J53, J54, J55, J56, J57, J58, J61, J63, J67, J70, J72, J73, J114, J127, J130, J147, J151, J165, J171, J174, J178, J180, J182, J183, J188, J190, J192, J208, J209, J211, J217, J243, J247, J248, J259, J262, J270, J271 S3, S4, S5, S13, S14 Conduct of Operations and Operational Discipline—Secondary Findings A6, A7 C13, C15, C20, C24, C27, C28, C60, C76 D7, D19 J21, J22, J24, J25, J32, J35, J40, J64, J65, J75, J76, J91, J108, J109, J116, J119, J128, J129, J131, J133, J162, J163, J170, J176, J181, J184, J185, J186, J212, J237, J253, J261 S1, S10, S12, S15

APPLICATION OF PROCESS SAFETY TO ENGINEERING DESIGN, CONSTRUCTION AND INSTALLATION 135 Inherently Safer Design Trevor Kletz was an early proponent of inherent safety and CCPS formalized his papers into books, initially in 1996, ag ain in 2009, and mo st recently into a Guideline CCPS (2019a). These updates did not change the concepts but refined the implementation examples. Some key words used to describe inherent safety include: ●Minimize : Reduce inventories of hazardous materials ●Substitute : Replace hazardous material or pr ocess with a more benign one ●Moderate : Use less hazardous processi ng or storage conditions ●Simplify : Eliminate unneeded complexity a nd make designs error tolerant Figure 7-2. Concept for includi ng inherently safer design (Broadribb, 2010) This figure incorporates ISD in a hierar chy of safety measures – with eliminate hazards being the preferred option but fo llowed by reduced severity and reduced likelihood options. Segregation uses la yout to separate hazards and reduce escalation. Passive and active safeguards improve safety for people, but do not address the inherent hazards of the process or the materials, instead mitigating their potential consequences. The final measure, procedural safeguards, which is least reliable, is the lowest category and it relies on personnel actions to reduce risk. 7.2.3 FEL-3 This stage takes the basic design from FEL-2 and refines it further. Earlier HIRA studies (including PHA, ISD and CRA) ar e updated to reflect the greater design detail available to ensure hazards have been identified, inherent safety principles

8 • Emergency Shutdowns 152 8.5.2 After the shut-down to different end state However, if the emergency shut-down procedure results in an abnormal equipment end state, such as rapid discharge of a reactor’s contents to a temporary tank or other safe location, then the discharged contents should be safe ly handled and removed from the affected process equipment or seco ndary containment before restart. If the equipment’s condition is n ot evaluated before restart, and as noted earlier, it is essential that everyone understands the final condition the equipment to ensure that it is prepared and ready before resuming start-up. If a part of the facility goes in to a circulation or a standby mode while other parts are shut-down, the hybrid operational state can be defined as a different end state , as well. Lessons Learned from incidents that occurred when trou ble-shooting other possibly shut- down parts of the process with oth er equipment placed into a standby mode were illustrated in Chapter 7 (e.g., C7.6.1-1 and C7.6.2-1). 8.5.3 Addressing damaged equipment and processes after a significant incident If emergency shut-down procedur e or Emergency Shutdown Device (ESD) causes damage to the equipment it is protecting due the consequences of not shutting the equipment down in the normal manner, there should be a proced ure for addressing the equipment damage before restart. Thorough tes ts and inspections, especially to check for potential (or known) internal damage, should be performed before the equipment is restarted. Th us, it is essential that everyone understands the final condition the eq uipment, that the equipment is assessed for potential damage, and that the equipment is repaired, if needed, before resuming start-up.

Piping and Instrumentation Diagram Development 164 and it is coned because of the higher weight of the tank walls, as was mentioned before. The selection of floor cone‐down angle is a process decision to facilitate mate-rial movement toward the center. The floor cone‐down angle could be limited because of construction implica-tions. Generally speaking wherever the tank has a larger diameter, it is more expensive to have a larger floor cone‐down angle. For large diameter tanks (possibly more than 5–10 m diameter) the floor cone‐down angle is generally kept below 5°. A rule of thumb that can give a guideline for the maximum allowable floor cone‐down angle – from a constructability viewpoint – is as follows: SDmax. 30 where: Smax. is the maximum allowable floor cone‐down angle in degrees (°) and D is the tank diameter in meters (m). However, there are some cases where the tank has a large diameter and for process reasons it should be high angled. There are some solutions for that. For example in some clarifiers or thickeners the floor should be high to help directing the settled solids (sludge) toward the center and finally removing it. If the clarifier or thickener is a large diameter one, the slope is possibly limited to 5° or a value around that. A rake system can be implemented to sweep the settled sludge toward the center of the tank. 9.18 Container Arrangement Containers could be in a series and/or parallel arrange-ment when they are used for unit operation or process units. As tanks are less likely used for unit operation or as process units, vessels are more commonly seen on P&IDs in series or parallel arrangements. If a vessel is supposed to be used for holding (short time) it is usually a single vessel. If tank(s) are supposed to be used for storing material (long term) it could be single in the majority of cases. However, they could be in multiple arrangements. This multiple arrangement could be neither parallel nor in series. The tanks could be connected to each other with a specific piping arrangement to provide flexibility for the operators to use them in series or parallel, depending on the case. Figure 9.34 shows such an arrangement. 9.19 Merging Containers Sometimes containers are merged together to save money. There are at least three ways to merge contain-ers: complete merge, merge with volume dedication, and merge while keeping physical boundaries. Complete merge is using one container for more than one purpose. This can be done by replacing two or more containers with only one of the same or larger volume. “Merge with volume dedication” has the same concept as “complete merge” with only one difference. In this type of merge the volume of the container is “somehow” dedicated to each user or purpose. For example in Figure 9.35 a single tank is used for fire water AND plant water. By a specific internal arrangement the plant water is allowed only to use the T ank “A”T ank “B”Exit from tank “B” Exit from tank “A” when it is in parallel Figure 9.34 Tanks in parallel and series arrangements. Figure 9.35 Dedication of tank water to fire water and to plant water.

Piping and Instrumentation Diagram Development 222 12.7 PRD Structure The earliest type of pressure device was the one used for  steam engines in the 1900s. It was basically a plug that  exerted force using a hanging weight to close the pressure‐releasing hole (orifice) and keep it close to the point that the pressure exceeds a specific value. The value of the pressure at which the device is intended to start to opening, or the “set pressure, ” could be adjusted by sliding the weight along the lever (Figure 12.4). Principally speaking, every pressure/vacuum relief device comprises three elements: 1) A pre ssure‐sensing element 2) A log ic 3) An opening e lement. The pressure‐sensing element senses and monitors the pressure of the enclosure and reports to the logic. The logic decides if/when the pressure exceeds a pre‐set pressure, or set pressure, and sends orders to the open-ing element to open the device to enable the release of the excessively high pressure. In case of a vacuum, it can also “suck” from the outside of the enclosure to decrease (“break”) the vacuum in the enclosure (Figure 12.5). Based on the fundamental concept of a PRD, all of these elements must be mechanical and their communi-cation must be through mechanical links. Electrical, pneumatic, or hydraulic signals are not acceptable as communication routes in pressure/vacuum relief devices. The above concept can be seen in the operation of a pressure relief valve (PRV) in Figure 12.6. 12.8 Six Steps to Providing a P rotective Layer There are six elements that should be considered during the design stage to make sure an optimal protective layer is provided. These elements are: 1) Lo cating the PRD 2) Po sitioning the PRD 3) Sp ecifying the PRD 4) Se lecting the right type of PRD 5) Se lecting the right type of PRD arrangement 6) Che cking the functionality of the PRD We will discuss all of the above items except for items 3 and 4. We briefly introduce “specifying PRD” because we need to understand the technical information of PRDs Table 12.7 Codes in the pr essure relief device industry. Code Code basisEnforcement agent (approval authority) US (general)NB‐501,boiler and pressure vessel codeASME B&PV code NBBI (National Board of Boiler and Pressure Vessel Inspectors) Canada (general)Safety codes CSA‐B51, ASME B&PV code A safety authority under the government: TSSA, ABSA, etc. Figure 12.4 Early t ype of pressure relief valve. Pressure sensing element LogicFully mechanical Acting (opening) element Figure 12.5 Fundamen tals of relief device operation. Inlet connectionOutlet connection Inlet connectionOutlet connection Closed position Open position Figure 12.6 PRV schema tic and operation.