Datasets:

hndc

/

AI4Hazard-small

like 0

7 • Unscheduled Shutdowns 132 LS3) Staged operator and craftsmen presence by starting the units up in sequence : fired steam boilers; rolled power of the process gas and refrigeration compressors; ethane furnaces. Then mana ged C4 storage, heavy ends processing and compositions (Note: water build -up, an issue not recognized during the restart, has been added to the start -up procedures). RBPS Element Strengths Hazards Identification and Risk Analysis Operating Procedures Asset Integrity Risk management system start-up weaknesses: LL1) Inadvertent miss of the boiler feed water line up during start- up, resulting in a Pressure Safety Valve (PSV) relieving water to the oily water sewer separator that also received benzene- containing streams. The separator overfilled, rele asing benzene to the atmosphere at levels requiring cordoning off of the separator area due to exceeded exposure limits . The sewer system overfilled, as well, requiring road area cordoning off due to the presence of hydrocarbons and benzene. The control room air conditioners ingested the vapors had to be turned off due to the odors. In addition , the investigation identified little use of the respirators after detection of the hydrocarbons and benzene. Relevant RBPS Elements Operating Procedures Training and Performance Assurance Emergency Management

Chapter No.: 1 Title Name: <TITLENAME> c08.indd Comp. by: <USER> Date: 25 Feb 2019 Time: 12:22:51 PM Stage: <STAGE> WorkFlow: <WORKFLOW> Page Number: 129 129 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 8.1 Introduction P&ID development for the purpose of inspection and maintenance was briefly discussed in Chapter  5. This concept, however, is so important it needs a chapter devoted to it. This chapter is about the required provisions that should be considered in P&IDs to facilitate an inspection and/or maintenance operation. These provisions are used for the purpose of facilitating inspection and maintenance, and also sometimes as part of a shutdown system. The instrumentation requirements of safe shutdown are mentioned in Chapter 15 but here the process requirements are mentioned. In this chapter we discuss the requirements of three elements of each process plant: equipment, utilities and instruments regarding ease of maintenance and inspection. The inspection/maintenance of instruments and util- ity systems is also mentioned here. 8.2 Different Types of Equipment Care From a maintenance viewpoint, “equipment care” can be categorized based on the location of applying care and the timing of the care. The care could be done in workshop, or “in‐workshop care, ” or could be done when the piece of equipment is not dislocated, “in‐place care. ” The in‐place care can be categorized further into “in‐line” or “off‐line” operation. The in‐line care is performed during the operation of the unit of interest while in off‐line care the unit of interest has to be pulled out of operation for the required care. We love equipment and units that need in‐line care rather than off‐line care. One attempt of equipment fabricators is to develop new equipment with less need for off‐line care and more in‐line care. For example in upstream oil extraction there is a piece of equipment called an FWKO drum or “free water knock out drum. ” This vessel may see heavy settlement of sands in it. In older days, at specific intervals the FWKO drum would be taken out of service for “de‐sanding operation” but these days there is equipment with automatic de‐sanding systems that can remove the sand from them during the normal operation of FWKO drums. Although it could be said that major maintenances occurs in workshops and only minor repair can be done in-place but it is not the cases always. There are some huge equipment that all of their maintenance-small or big- should be done in field, or “in-place care” . Tanks are a famous example of them. Each of these two different types of equipment care forces us to provide different types of provisions for equipment on P&IDs. Table 8.1 shows the responsibility of different groups in process plants for different types of equipment care. 8.3 In‐place In‐line Equipment Care “In‐place in‐line care” means any activity that takes care of a process item without any disturbance to the normal operation of the element or with just mild disturbance. There are mainly two types of activities in this group: activities related to the rejuvenation of the equipment and activities related to monitoring the equipment. Rejuvenation activities are needed for equipment that is inherently in an intermittent or semi‐continuous mode of operation. One example of rejuvenation activities has already been stated as de‐sanding in FWKO drums. The rejuvenation activities could be triggered by one or a combination of these parameters. They can be initi-ated by receiving a specific sensor signal (event based), or they could be initiated after a specific time interval (time based), or by the decision of the operator. From the type of operation viewpoint, a rejuvenation process can be done manually or automatically. In manual 8 Provisions for Ease of Maintenance

392 INVESTIGATING PROCESS SAFETY INCIDENTS Logic Tree (4 of 9)

4 • Process Shutdowns 71 Relevant RBPS Elements: Hazard Identification and Risk Analysis Operating Procedures Management of Change LL2) Equipment isolation for de- inventorying or decontamination activities prior to maintenance should use a double -block-and- bleed valve design and not drain contents directly into sewers. Relevant RBPS Element: Process Knowledge Management 4.7.2 Incidents during start-up after planned project-related shutdowns C4.7.2 -1 – Maintenance Replacement [2, pp. 50-52] Incident Year : Not noted; published incident in 2015 Cause of incident : Similar, available, and not-per-design - specifications check valve installed during routine seal maintenance on a large, offline pump Incident impact : Release of 18,000–23,000 kg (20- 25 tons) of flammable light hydrocarbon which ignited and burned for more than 5 days, resulting in a costly three -month business interruption Risk management system weaknesses : LL1) The replacement valve was the only check valve in the warehouse that could be fitted in th e 0.3 m (12 inch) 21 bar (300 psi) discharge line of the large petroleum pump. Although the replacement check valve matched the piping’s diameter and pressure rating, it required a smal l spool piece and additional gasket to bridge the gap in the piping. The failed gasket material was rubber suitable for fire water serv ice, not hot hydrocarbon service. Recommendations from the incident review included upgrades to the site’s change program, especially “small changes,” and

384 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Figure 18.1. Flixborough Reactors (CCPS 2005) On March 27, 1974, a crack was detected on Reactor No.5. A Maintenance Engineer recommended complete closure for 3 weeks. Th e Maintenance Manager, whose job had been filled for several months by the head of the labo ratory while awaiting a reorganization of the company, proposed dismantling Reactor No. 5 and connecting numbers 4 and 6 together by a 500 mm (20 inch) diameter temporary connection. To support the piping, the proposal was to use a structure made from conventional construc tion industry scaffolding, Figure 18.2. This resulted in a piece of bent piping with two bellows, one at either end. During start-up activities this piping was subjected to asymmetrical lo ads and failed as the double bellows design allowed the pipe connection between the bello ws to be unsupported and unrestrained. A single bellows would have been effective to abso rb thermal expansions with structural rigidity coming from fixed connections at either end and only a short length of piping. The temporary connection was not adequate fo r the forces and temperatures involved, and failed, releasing 30 metric tons of cyclohex ane in 30 seconds. Of the 28 fatalities, 18 were in the control room. The fire lasted over thr ee days with 40,000 m² ( 10 acres) affected. See Figure 18.3. (CCPS 2008).

328 | Appendix E Process Safety Culture Case Histories retired. The new Vice President was challenged by the business shift and may have been distracted from the predecessor’s process safety action plan. After a few more years and m ore staff changes, the com pany experienced a cluster of major and minor incidents involving injuries and fatalities that would have been unheard of just a few years later. Over the next few years, incidents began cropping up at m any sites, and to many, the com pany’s process safety culture appeared to have collapsed. Could the company’s culture have been toppled like a Jenga tower by the removal of one strategic piece? Or, continuing the analogy, can culture survive m any sm all weaknesses until critical failure? In other words, was the departure of the Vice President the critical factor, or had the culture become so weakened that it would have collapsed even with that Vice President’s leadership? What was the true state of culture when the metrics trend first started to go negative? Could the culture have already collapsed, and could the absence of frank and open communication have prevented the well-m eaning Vice President from knowing about it until it was too late? Did the com pany truly have an imperative for safety , or was its reputation built on a few very visible safety cham pions? Could the imperative for safety been focused on occupational safety and not enough on process safety? Could the Vice President’s successors have talked-the-talk about process safety, but not exhibited “Felt Leadership?” Establish an Imperative for Safety, Provide Strong Leadership, Maintain a Sense of Vulnerability, Understand and Act Upon Hazards/Risks, Ensure Open and Frank Communications. E.38 Failure of Imagination? In February 1967, an electrical fire within the crew capsule of the Apollo 1 spacecraft killed all three astronauts as they conducted a simulated launch Actual Case History

388 INVESTIGATING PROCESS SAFETY INCIDENTS Sequence of Events (cont.) DATE TIM E EVENT Aug 1 ~ 11:20 A.M. Fire brigade uses limited w ater supply on engine to shield tw o members of team and attempts rescue of lead operator. Another explosion occurs and four fire brigade members are injured by metal fragments. ~11:22 A.M. Local fire department arrives. After 11:22 A.M. Spread of fire is slow ed using w ater from fire department trucks. ~11:30 A.M. Maintenance completes move of batteries from No. 1 diesel fire pump to No. 2 diesel fire pump. No. 2 diesel fire pump is started. After 11:30 A.M. Automatic deluge sprinkler system found to be severely damaged by fire / explosions an d is now valved into OFF position. Three fixed fire monitors directed on fire at full flow . Tw o hose streams from hydrants directed on fire also. ~11:58 A.M. Fire deemed under control. ~12:10 A.M. Final extinguishment of fire. Aug 2 Lead operator dies from burn complications. Aug 3 No. 1 diesel fire pump repaired.

Pumps and Compressors 193 A typical P&ID representation of rotary pumps is shown in Figure 10.32. 10.7.2 PD Pump A rrangements PD pumps could be placed in parallel to provide a spare pump. However, having two (or more) PD pumps func - tioning at the same time is not very common. Less com-mon is using PD pumps in series. When two (or more) PD pumps are placed in a parallel arrangement they may need additional considerations during P&ID development. Some companies put another PSV for each PD pump in parallel to protect the suction side of them from being over‐pressurized by the operat - ing PD pump. If the discharge isolation valve of the spare PD pumps is negligently left open, the operating PD pump pressurizes the spare pump, both on the discharge side and suction side. Another PSV around the suction isolation valve protects the suction side of the PD pump (Figure 10.33).10.7.3 Mer ging PD Pumps Similar to centrifugal pumps, PD pumps may also be used in multi‐service applications. However such appli-cations are much less common than centrifugal pumps. It is generally recommended to use one PD pump for one suction source and one discharge destination. 10.7.4 Tying Together Dissimilar Pumps Here we are talking about specific requirements when dissimilar pumps are tied together. Such practice is not a very good practice but sometimes can be done after con-sidering all the requirements. For a parallel arrangement of dissimilar pumps (Figure 10.34) there could be different reasons for paral-leling, similar to centrifugal pumps. However, it is not very common to see such arrangements in process plants. One application could be when the viscosity of the service liquid is changing a lot and two different pumps should be ready to be able to handle liquids with different viscosities through two different pumps. The other cases could be when the main pump is the centrif - ugal pump and a small PD pump is used just to keep the system pressurized when the centrifugal pump is off. For a series arrangement of dissimilar pumps one rule of thumb can be memorized: “using a PF pump as the booster pump of another pump is not good practice. ” Different cases of this rule of thumb and its applica- bility can be seen in Figure 10.35. 10.7.5 PD Pump D rives The most common type of drives for PD pumps is elec - tric motors. This is similar to centrifugal pumps. For reciprocating type PD pumps there are, however, some other options available. One option is air‐operated drives. In air‐operated drives a source of air, like plant air, is used to generate a reciprocating movement. This reciprocating shaft is connected to the pump shaft to operate it. Such pumps may be named “air‐driven pumps. ” This type of drive is very common for diaphragm pumps. It should be PSV If you can not trust inte rnal PSVNo dampener Check va lve No min. flow Figure 10.32 P&ID arr angement of a rotary pump. To protect pump Accidentally close OperatingMT MT SpareTo protect low pressure suction side Accidentally open Figure 10.33 Additional PSV s for PD pumps in parallel. Figure 10.34 Dissimilar pumps in parallel .

119 Figure 6.6: The same process as Fi gure 6.5 in a series of simpler reactors (Ref 6.7 Hendershot) 6.10 LIMITATION OF AVAILABLE ENERGY Adding energy to a chemical proces s is often a necessary component of product manufacture. However, in some cases, the energy addition method is not carefully matched to the amount of energy needed. The method of energy addition used can result in excess energy being added because its design does not incorporate inherently safer design principles. Examples of proper matc hing of required energy are:

340 | Appendix E Process Safety Culture Case Histories powder and solvent that had a 24 hour “time to maximum rate” a few degrees below the drying tem perature. The drying had not always been run at a warm tem perature. However, when drying was done at a cooler temperature, an unacceptable amount of solvent remained in the filtered powder. To obtain a drier cake, improving the occupational safety during pack-out, the plant increased the drying tem perature. The MOC review concluded that the change was necessary to improve safety in packing out the cake. However, in conducting the M OC review, the original reactivity data were not fully considered. Instead, a new set of thermal tests on solvent-free powder were conducted, which was found to be quite stable. When the investigation team reported its findings, a m anager com mented, “This is the last place I thought we’d have an accident.” He explained that the site ran m any highly energetic reactions, handled highly toxic chem icals, and distilled m any volatile and flam mable solvents. Surely if there was an incident on the site, he said, it would not happen in a filter. What culture questions m ight the investigation team have considered? Did an unbalanced imperative for safety lead to disregarding the original reactivity study (process safety hazard) so that the plant could address the solvent exposure issues (potential occupational safety hazard) during pack-out? The technical team responsible for the process was unaware of the potential for reaction between powder and solvent. What barriers to open communication could have existed between the technical team and the owner of the process safety information? What other com munication roadblocks m ight there have been? Did the plant feel less sense of vulnerability for that operation than they should have because it was “m erely” a filter, and if so, could that have contributed to the incident? How should a site

2.4 Ensure Open and Frank Communications |39 outside organizations in a variety of situations that span from routine to crisis. These outsiders include labor, emergency responders, law enforcement, media, regulators, interest groups, and comm unity groups, among others. Open communication channels with outside groups must be available for use when needed, sometimes on an em ergency basis. B y establishing personal relationships between the parties, the foundation of trust can be established. With that relationship, it can be easier to com municate clearly in their language. In emergency situations, being able to provide the critical information needed to responders and coordinate efforts is essential. Comm unicating with the media during a crisis is particularly im portant and very sensitive (see section 4.4). The closure of once-open communications channels can place individuals in an ethical dilemm a when they believe that an important process safety message is being suppressed or failing to reach decision makers. This was a contributor to the Challenger incident (Ref 2.17) and raises ethical questions (see section 4.3). Workplace comm unications can be written or verbal, official or unofficial, and form al or inform al. Each has advantages and disadvantages, but all should be addressed in culture improvement efforts. Written communication includes the m any form s of electronic com munication. Well-written communications, edited to im prove clarity, can be less subject to misinterpretation. Verbal com munication has the potential advantage of being able to fully convey feeling. If that feeling shows the speakers positive emotional commitment to process safety, the communication can be strongly enforced. Of course, pitfalls exist with both written and verbal com munication. Communications written poorly can be easily m isinterpreted or simply ignored. Verbal communication can be

NOTIFICATION , CLASSIFICATION & INVESTIGATION 91 These internal notifications can be trigger mechanisms for starting specific portions of the incident in vestigation managemen t system and for other associated decision-making (see Section 5.4 below). The initial notifications are ofte n based on the actual severity classification of the incident. Sometime s there may be a delay in determining the potential severity until more information on the incident is available. In this case, if the potential severity is greater than the actual severity, a new notification may be required. Some co mpanies require notification of ‘high potential’ incidents (e.g., CCPS/API Tier 1 ≥ single fatality) even if the actual severity was less severe (e.g., lost-time injury or near-miss). 5.3.2 Agency N otification Depending on the jurisdic tion, the regulatory agen cy(s) may require verbal and/or written communication that an incident of a certain severity has occurred. A timeframe for this communica tion is often specified and typically varies, for example, 8 hours by US OSHA for a fatality and 24 hours for hospitalization or severe injury. In a few cases a longer timeframe is permitted. For example, in the UK under the RIDDOR regulations (HM Government, 2013), an accident resulting in a fatality or hospitalization of non-workers is required to be reported within 10 days. In instances where a verbal notification by telephone is sufficient, it is advised that the individual reporting the incident make a written record of this verbal communication, noting th e time, person involved, extent of information disclosed, and any special instructions or r equests made by the agency at the time of no tice. Some jurisdictions may use a recorded line, and the individual reporting the incident should keep a similar written record. In some jurisdictions, the initial noti fication may also be the basis for the company to satisfy regulatory requirements on the timely initiation of the investigation process. 5.3.3 Other Stakeholder N otification Initial notifications to other stakehol ders may be appropriate depending on circumstances. These notifications may be managed through the relevant corporate and/or site management syst em, and include, but are not limited to: Family members Neighboring facilities Neighboring community

176 | 13 REAL Model Scenario: Internalizing a High-Profile Incident First, they would need a general policy, owned by Samir and enforced by Prasad, clearly stating their process safety expectations. The policy would be the cornerstone of a more formal operating culture which would include better operating discipline. Supervisors would drive operating discipline and adherence to the culture rather than over-stressing production. The four struggled with this last point, worrying that if they didn’t push production, productivity would suffer. They expressed this concern to Samir when he stopped by to check on their progress. Samir complimented them on the work they’d already done, and then addressed the question. “How much productivity do we waste looking for temporary fixes?” he asked. “We should more than make up the difference if we manage our operations in an orderly fashion.” 13.7 Implement Abishek’s engineering team developed the new standard equipment designs, worked with procurement to designate approved vendors, and implemented the new designs during a 30-day turnaround. During the turnaround, Samir and Rakesh led workshops and discussions with the supervisors to coach them on the new policy and the desired culture, especially the increased focus on operational discipline. Then they watched and provided support as the supervisors relayed the message to the workers. At first, the new focus on operational discipline felt uncomfortable, especially for the long-time employees. But Samir was insistent, walking through the plant regularly to encourage good behavior, correct risky behavior, and answer questions. After a few weeks, workers began to see the benefits of the new approach. They came to appreciate Samir’s commitment, and they began to get excited. Toward the end of the turnaround, the mechanics and operators received specialized training on the modified processes and equipment. As start-up approached, Chana Oil Seed felt like a completely new company. 13.8 Embed and Refresh After the usual start-up headaches, Chana settled into the new routine of production and maintenance. The first challenge arose three months later. Rakesh was in the plant monitoring for hydrogen leaks when he noticed a group of workers looking toward the ceiling, arguing and gesturing wildly. He

Pumps and Compressors 183 10.6.6.1 Centrifugal Pumps in Parallel For parallel arrangement of pumps there are different arrangements including the operating‐spare arrange-ment, the lead–lag arrangement, and simultaneously operating pumps. Centrifugal pumps can be arranged in parallel forms for different reasons. They are: ●If the flow rate that should be handled by the pump is huge and there is no available single pump to be used, two or more centrifugal pump can be placed in parallel instead of one single pump. ●When a higher reliability is needed: this is the case that we need to put a spare pump in parallel with the main pump. However, in this case there is no time that all parallel pumps work together. ●When enough NPSH A is not available: sometimes the suction side of a centrifugal pump doesn’t provide enough NPSHA and it is close to or less than the NPSH R one solution is using two or more smaller centrifugal pumps. This can solve the problem because smaller pumps have smaller NPSH R. Figure 10.14 shows a schematic of pumps in parallel. However, the designer should be careful about recogniz-ing centrifugal pumps in parallel. They are not always easy to recognize. If there is a single pump pumping a liquid from point A to destination B and somewhere in the middle of discharge pipe the discharge of another centrifugal pump is merged to the pipe, these two pipes are considered as parallel pumps. The importance of recognizing parallel pumps is that they usually are through one single data-sheet. When two centrifugal pumps operating as parallel pumps they should be completely identical. Because of that they are usually placed in one single datasheet to make sure they are from one single pump vendor and they are identical. 10.6.6.1.1 Minimum Flow Pipe for Parallel Pumps If there are parallel pumps but only one of them works at the time and the rest are spare pumps there is no concern about the minimum flow protection pipe. In this case, the minimum flow protection pipe can be branched‐off from the main discharge header of the pumps. However, when there are operating parallel pumps the concept is a little bit more complicated. There are at least three avail-able options (Figure 10.15). Option 1: one minimum flow protection pipe as a shared minimum flow protection pipe for all the parallel operating pumps. Although this option is acceptable when only one parallel pump is operating at a time it is not a good practice for simultaneously parallel operating pumps. The reason that this is not a good idea is that this option doesn’t necessarily compensate the starving pump. As the parallel operating pumps don’t have dedicated minimum flow protection pipes if one of the pumps operating with the flow rate below the minimum flow rate and the other pump operating well, this system cannot work properly. This option could be improved with a good control system. The control system for this option will be discussed in Chapter 14. Option 2: Dedicated minimum flow protection pipe for each parallel pump but with a merged pipe after the control valve. In this option the better approach of using a dedicated pipe and a dedicated control valve is used; however, to save some money all the dedicated Figure 10.14 Par allel pumps.Figure 10.15 Minimum flo w pipe for parallel operating pumps.

Piping and Instrumentation Diagram Development 306 15.6.2 Controlling Multiple Pipes In t his section we talk about the control of multiple con- nected pipes. These arrangements can be classifies into two main arrangements of flow merging and flow splitting. 15.6.2.1 Flow Merging Stream merging could be for one of the following purposes: ●For gathering ●To provide a carrier ●For blending ●For make‐up or back‐up purposes (can be referred to as the “preferred source” concept). “Gathering” is the operation where we “unify” several pipes to send the fluid to a single destination. Generally, in these cases, all the streams are the same fluid, or are compatible with each other. In this application, we are not looking for a specific flow rate on any stream; we only want to merge them together. Stream merging for the purpose of providing a car - rier is when we add one stream to another to make it easier to transfer, or to make it easier for some other final purpose. For example, adding water to lime in the powder form and making lime slurry in the majority of cases is done simply to convert the powder, which is hard to move, into a liquid, which is easy to move. In such applications, generally only the flow rate of one stream is important, and only one stream is controlled. In the lime slurry example, we need to add enough water to the lime powder slurry to make it moveable; however, if the plan is to send lime to a reactor for a reaction, then the concentration of the lime needs to be much lower than the adequate lime concentration for conveying. As the reaction happens mainly in the liquid phase (rather than the solid phase), the lime needs to be in the liquid phase, and with diluting it we will have a better reaction with less unconsumed lime leaving the reactor. The “flow” of lime powder may be controlled upstream of its merging point, since it is a reactant in the down-stream reactor. In blending operations, several streams are merged together to achieve specific properties in the final mixed product. For example, in the final stages of refineries, they mix different proportions of light and heavy products with each other to make gasoline with specific proper - ties, kerosene with specific properties and so on. In factories that make orange juice from concentrate, at the start of the process, they make concentrates from oranges (by evaporation), and then they add enough water to make orange juice with a specific brix number (sweetness) and concentration. These are all examples of blending. In blending, the flow rates of all streams are important. In make‐up or back‐up applications, one (or more) stream(s) work(s) to compensate for the shortage or lack of other stream(s). Generally, all streams have a unique composition in make‐up or back‐up applications. The best way to illustrate the different ways of control- ling flow merging is to look at examples. Figure 15.19 shows a stream‐merging example where only one stream is flow controlled. In this example, stream A could be the fluid of importance and stream B could be the carrier. Figure 15.20 shows a stream‐merging example where stream A is the solution of importance and stream B is the carrier. Since the concentration of the active ingredi-ent in stream A is fluctuating, to make sure that we always send a specific mass of the active ingredient, we put a composition loop control on stream A. Figure 15.21 shows an example in which the concen- tration of the final product is important. This could be an example of a blending operation. The only issue with this control system could be its slow response. As we know, the majority of composition analyzers are slow. However, it could be acceptable if the process analyzer is not very slow, which could be the case for pH or conductivity analyzers. FC FT Stream A Stream B Figure 15.19 Flo w merging: route of least resistance. Stream AATAC Stream B Figure 15.20 Flo w merging with composition loop.

C Human Factors Competency Matrix Table C-1 Human Factors Competency Matrix HF Competency Performance/ Knowledge Criteria Level 1 - Operator Level 2 - Supervisor* Level 3 - Manager** Overarching concepts, principles, and knowledge Human Factors Understand what is meant by Human Factors Understands the definition of Human Factors as it applies to process operations Understands the concept of Human Factors and its component disciplines Understands the contribution of Human Factors to improvement of process operations Human Factors concepts Can explain and apply Human Factors concepts to process operations Understands Human Factors concepts (systems approach to human performance, error is a symptom, error traps, no blame, etc.) Understands implications of Human Factors concepts (systems approach to human performance, error is a symptom, error traps, no blame etc.) for their supervisory role Can instruct on Human Factors concepts (systems approach to human performance, error is a symptom, error traps, no blame, etc.) Human Factors Handbook For Process Plant Operations: Improving Process Safety and System Performance CCPS. © 2022 CCPS. Published 2022 The American Institute of Chemical Engineers.

148 INVESTIGATING PROCESS SAFETY INCIDENTS • Pressure containing equipment • Gaskets and flanges • Seized parts • Misaligned or misassembled parts • Control or indicating devices in the wrong position • Use of incorrect components • Samples from all relevant vessels and piping including: - Raw materials - Intermediate products - Completed products and chemicals - Pools of residues of chemicals or materials - Waste products (solids, liquids, gases) - Scales and deposits - Quality control samples - Any “new” chemicals present • Foreign objects • Portable and temporary equipment (including tools, containers and vehicles) • Undamaged areas and equipment • Pressure relief device components • Metallurgical samples • Conductivity measurements • Explosion fragments • Data recorders • Sensors • Process controls • Electrical switch gear • Missing physical data such as plant and equipment, stains, oxidation, etc. Not everything in the incident zone will be significant, although it is often important to identify equ ipment, structures, and pipework that are not damaged. The key is to quickly identify what may be irrele vant, while causing minimum disturbance to what could prove to be relevant. This judgment is based on team members’ experience and expertise. Ke y physical items should be photographed and tagged before any movement, if possible. A guide rule for the decisio n on what to keep is— too much is better than too little . Any known or anticipate d dismantling, disassembly, or opening of equipment should be plan ned and coordinated with the appropriate groups using a test plan or written protocol. This is important to ensure the activity is conducted in a safe manner while not inadvertently damaging evidence.

40 reported in plants that have been operating for many years (Ref 2.8 CCPS GED, Ref 2.28 Wade, Ref 2.7 Carrithers). 2.9 SUMMARY Inherent safety represents a fundamentally different way of thinking about the design and operation of chemical processes and plants. It focuses on the eliminatio n or reduction of the hazards, rather than on risk management and control via addi tional layers of protection. The demarcation line between an IS measure and a layer of protection can be defined as follows: IS measures eliminate or reduce the hazard(s), whereas layers of protection do no t eliminate the hazard(s) but reduce their likelihood of occurrence. Someti mes the reduction in likelihood is substantial, but there is still a residu al (and measurable) likelihood that the hazard can occur, i.e., the risk of the incident of interest has not been reduced to zero. This approach will result in safer and more robust processes, and it is likely that these inherently sa fer processes will also be more economical in the long run (Ref 2.20 Kletz, 1984, Ref 2.21 Kletz, 2010). However, the cost of changing an existing design to a more inherently safer technology may be unjustified or difficult to justify from a strictly investment standpoint. For this reason, the options must be holistically weighed, and the total life cycl e costs and risks analyzed for completeness. Indirect costs of implementing inherently safer technologies, such as in advertent risk transfer, must also be weighed before a final decision is made to implement them. Eliminating or reducing the haza rd through the application of inherently safer design is the fi rst element in the process risk management hierarchy. Other strategies—passive, active or procedural—constitute layers of protection for the hazard. 2.10 REFERENCES 2.1 The American Heritage® Dictionary of the English Language, Fourth Edition, Houghton Mifflin Company, 2000.

200 INVESTIGATING PROCESS SAFETY INCIDENTS provided the method helps the inve stigator and others understand the incident. 9.6.3 Developing a Causal Factor Chart The first step in developing a causal fa ctor chart is to define the end of the incident sequence. Construction of th e chart should start early from the end point and work backward to reco nstruct what happened before the incident by identifying the most imm ediate contributing events. Starting at the end point, it is th en necessary to convert the collected evidence into statements of either fa ct or supposition. By taking a small step backward in time, the investigator asks, “W hat happened just before this event?” It is important to clearly distingu ish any assumptions as supposition. Then the investigator writes a statement for what happened and enters the fact (or supposition) as an event block or condition oval on the causal factor chart at the appropriate locati on on the timeline. Statements that caused an event to occur should be tr eated like conditions and added in an oval. The investigator tests this new event (or condition) for sufficiency by asking, for example, questions such as: “Does this block always lead to the next block (in this case, the endpoint)?” “Are there any layers of protection that should have prevented this sequence?” The process is repeated slowly working backward in time. The entire causal factor chart is th en reviewed to identify any omissions or gaps in the chronology. Additi onal effort is required to gather further evidence to close these gaps. If new data are inserted into the timeline, the sequence should be retested fo r sufficiency. Some gaps may remain even after this additional effort. The causal factor chart review should also identify and eliminate any facts that are not necessary to describe the incident. Detailed rules for causal fa ctor charting are shown in Figure 9.3. Charting the events and conditions on a causal factor chart assists the investigator in thinking logically through the incident. However, the investigator must exercise care to avoid locking into a preconceived hypothesis. It is important to keep an open mind and objectively analyze all possible hypotheses for the events and conditions lead ing up to the incident. Initial assumptions can change dramatically during the course of an

270 INVESTIGATING PROCESS SAFETY INCIDENTS maintaining a specific competency and qualification level or effective operational discipline • Employee previously rewarded for deviating from the procedure, due to a culture of rewarding speed over quality, resulting from and reflecting a defective quality-assurance management system and a defective operational discipline system • Employee following personal example set by his supervisor, due to a defective system for establishi ng and maintaining supervisory performance standards or operational discipline • There are multiple accepted practice s (daytime versus weekends for example), due to the presence of dual standards, due to defects in the supervisory or auditing management systems or operational discipline system • Employee is experiencing temporary task overload, due to defects in the scheduling and task allocation system, and/or due to ineffective implementation of downsizing • Employee has physical/mental/emotio nal reason(s) that causes him or her to deviate from the established procedure, due to defects in the fitness-for-duty management system • Employee believed he was using the correct version of the procedure, but due to defects in the document management system, he was using an out-of-date edition • Employee was improperly trained due to defects in the training system • Management’s expectations for pr ocedure use. Depending on the complexity of the process and th e activity, a procedure could be written with the intent that it be followed at a detailed level or it could be written for training and reference. Consequently, the procedure type/style could also be a defect. In some instances, the failure to follow established procedure may be due to inadequate knowledge. The classi c recommendation that accompanies this symptom is to provide training (or refresher training) to ensure the person understands how to follow the established pr ocedure. An example of a typical recommendation associated with this mistake is “review the procedure with the employee to ensu re that he understands the proper action expected.” The traini ng activity may be beneficial to the person(s) who receive it, but in most cases, the training fails to identify and address the underlying cause(s) of the defic iency in the knowledge/competency system

xx PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Figure 14.10. Hong Kong societal risk criteria ............................................................................. 3 26 Figure 14.11. Types of ALARP demonstration .............................................................................. 327 Figure 14.12. Typical layers of protection .................................................................................... . 328 Figure 14.13. Frequency of hole sizes ......................................................................................... .. 333 Figure 14.14. Risk Matrix ..................................................................................................... ........... 333 Figure 15.1. Oxidation reactor after explosion ............................................................................ 338 Figure 15.2. One of several units impacted by explosion ........................................................... 338 Figure 15.3. Schematic of oxidation reactor ................................................................................ 33 9 Figure 15.4. Predicted flammable vapor cl oud from reactor explosion ................................... 340 Figure 15.5. Terminology describing la yers of protection .......................................................... 343 Figure 15.6. Swiss cheese model ............................................................................................... .... 343 Figure 15.7. Example bow tie model ............................................................................................ . 345 Figure 15.8. Steps in constructing a bow tie model .................................................................... 345 Figure 15.9 a. The left side (threat legs) of a bow tie for loss of containment ......................... 346 Figure 15.9 b. The right side (consequence legs ) of a bow tie for loss of containment .......... 347 Figure 16.1. Smoke plumes from Formosa plant ........................................................................ 356 Figure 16.2. Formosa reactor building elevation view ................................................................ 356 Figure 16.3. Cutaway of the Formosa reactor building ............................................................... 357 Figure 16.4. Model for human factors .......................................................................................... 359 Figure 16.5. Human factors topics related to the human individual ........................................ 361 Figure 16.6. Human failure types .............................................................................................. .... 363 Figure 16.7. Decision making continuum ..................................................................................... 36 4 Figure 16.8. Factors impacting work team performance ........................................................... 365 Figure 16.9. Task Improvement Plan steps .................................................................................. 368 Figure 17.1. Piper Alpha platform ............................................................................................. .... 374 Figure 17.2. Piper Alpha platform after the incident .................................................................. 374 Figure 17.3. Schematic of Piper Alpha platform .......................................................................... 375 Figure 17.4. An operational readiness review should follow these activities .......................... 378 Figure 18.1. Flixborough Reactors ............................................................................................. .... 384 Figure 18.2. Schematic of Flixborough piping replacement ....................................................... 385 Figure 18.3. Flixborough site after explosion ............................................................................... 3 85 Figure 18.4. MOC system flowchart ............................................................................................. . 387 Figure 19.1. Exxon Valdez tanker leaking oil ................................................................................ 3 94 Figure 19.2. Shoreline cleanup operations in Northwest Bay, west arm, June 1989 .............. 396 Figure 19.3. Conduct of operations model ................................................................................... 40 1 Figure 19.4. Car seal on a valve handle ....................................................................................... . 404 Figure 19.5. Example valve line-up ............................................................................................ .... 406 Figure 19.6. OSHA QuickCard on permit-required confined spaces ......................................... 410 Figure 20.1. Video stills of WFC fire and explosion ...................................................................... 413 Figure 20.2. Fertilizer building overview ..................................................................................... .. 415 Figure 20.3. WFC and community growth (left - 1970; right - 2010) .......................................... 415 Figure 20.4. Overview of damaged WFC ....................................................................................... 41 6 Figure 20.5. Apartment complex damage .................................................................................... 416 Figure 20.6. Coffeyville Refinery 2007 flood ................................................................................. 420 Figure 20.7. Incident command structure .................................................................................... 42 2 Figure 21.1. Fire on Deepwater Horizon ....................................................................................... 428

49 3.4 TUBULAR REACTORS Tubular reactors often offer the greatest potential among reaction devices for inventory reduction. They are usually extremely simple in design, containing no moving parts and a minimum number of joints and connections. A relatively slow reac tion can be completed in a long tubular reactor if mixing is adequa te. There are many devices available for providing mixing in tubular reac tors, including jet mixers, eductors, and static mixers. It is generally desirable to minimize the diameter of a tubular reactor because the leak rate in case of a tube failure is proportional to its cross-sectional area. For exotherm ic reactions, heat transfer will also be more efficient with a smaller tubular reactor. However, these advantages must be balanced against the higher-pressure drop due to flow through smaller reactor tubes. 3.5 LOOP REACTORS A loop reactor is a continuous tube or pipe that connects the outlet of a circulation pump to its inlet (Figure 3. 1). Reactants are fed into the loop, where the reaction occurs, and product is then withdrawn. Loop reactors have been used in place of batch stirred tank reactors in a variety of applications, including chlorination , ethoxylation, hydrogenation, and polymerization. A loop reactor is typically much smaller than a batch reactor producing the same amount of product. Wilkinson and Geddes (Ref 3.23 Wilkinson) describe a 50-liter loop reactor for a polymerization process that has a capacity equal to that of a 5000-liter batch reactor. Mass transfer is often the rate limiting step in gas-liquid reactions, and a loop reactor design increases mass tr ansfer, while reduci ng reactor size and improving process yields. As an example, an organic material was originally chlorinated in a glass-lined batch stirred tank reactor, with chlorine fed through a dip pipe. Replac ing the stirred tank reactor with a loop reactor that fed chlorine to the recirculating liquid stream through an eductor reduced reactor size, increased productivity, and reduced chlorine usage. These results are summarized in Table 3.1 (Ref. 3.2 CCPS).

376 Human Factors Handbook A.2.3 Making Compliance Easier The Energy Institute’s ‘Hearts and Mind’ have issued extensive guidance on “Making Compliance Easier” [120]. The guide states that: “The new way of thinking sees non-co mpliance as a natural consequence of work situations that are far from ideal. It is therefore up to organizations to ensure that their systems and processes do not give rise to situations that make mistakes and non-compliances more lik ely... World class organizations …make rules and procedures clear, helpful and ea sy to follow, and are always open to ideas for improving them.” (p4) “The human performance principles we re designed to help organizations consider how to keep everyone safe, healthy and productive. They acknowledge that everyone makes mist akes, and that performance may be compromised by factors like complexity of a task, distraction and repetition.” (p2) Two figures from the guide are shown in Figure A-1 and Figure A-2.

4. Supporting human capabilities 39 Abnormal or emergency operations requir e a high level of psychological skills, including decision-making under stress and maintaining situation awareness. Process operators often require a combinat ion of knowledge, procedural skills and psychomotor skills. Psychomotor skills are physical skills such as applying pressure to an accelerator. Chapter 24 explores the Human Factors of helping to prepare people to handle abnormal and emergency operations. 4.6 Cognitive heuristics/biases 4.6.1 Cognitive heuristics and performance In order to help process large amounts of complex information and make decisions faster, people tend to use ment al shortcuts known as “rules of thumb” or “cognitive heuristics”. Mental shortcuts are a pattern of thinking, like a “go to” way of making decisions. The brain uses “s hortcuts” to make fast decisions. These are often subconscious and habitual ways of thinking. They are often correct. These rules of thumb are often based on past similar experience. An example of a “rule of thumb” is that “rust thickness is more or less 10 times the amount of steel lost (that has now been transformed into rust)”. These “short cuts” help people simplify the process of making sense of complex information. Cognitive heuristics are relevant to rule- based and knowledge-based human performa nce, especially when making sense of events, forming judgments and making decisions. In addition, if a person must complete a complicated task in a short period of time, they may carry out the ta sk by paying attention to enough but not all of the information available and consider enough but not all of the possibilities. This allows a “satisfactory” or “good enough” deci sion to be made in the time available. As previously noted, people may not be able to pay attention to all of the information, so selective attention may be necessary. Although cognitive heuristics are useful and important ways to make decisions, these “rules of thumb” can also cause mist akes. By taking a short cut when making a judgment, some information will be missed, which means the right option may not be considered. This may be even more likely if a person feels they are under time pressure or are feeling fatigued, demotivated or bored. However, they can have a negative impact on judgment, thinking and decision- making. In a safety critical setting, it is vital to recognize the potential for these mental shortcuts becoming inaccurate bias and to correct this. Cognitive heuristics help people to make quick and sensible judgments and decisions, even when information is missing or not available. But they can also cause mistakes.

223 include changes in the process itself could be eval uated to create a more robust, effective and individualistic facility security program. The IS knowledgeable chemical en gineer has an opportunity to positively influence security, as well as safety, through the process life cycle. As is true for safety, securi ty issues can be addressed in the concept and design phases of a projec t, allowing for mo re cost-effective considerations that could eliminate or greatly reduce security risks. As described in Chapter 5, inherent safe ty concepts may be applicable to both new and existing/modified proc esses by answering questions such as: 1.First, could a potential conseque nce be eliminated altogether? (First order IS measure) 2.Second, could the magnitude or severity of the potential consequence be reduced? (2nd order IS measure) 3.Finally, could the remaining vuln erabilities be reduced through consideration of the hierarchy of passive, active, and procedural measures? (Layers of Security) Consequence . As described earlier, chem ical security consequences typically result from one of five security issues: Release of a toxic, flammable or explosive chemical Theft or diversion of a chemical that could be used as or used to produce a chemical weapon Sabotage or contamination of a shipment to cause a toxic release during transportation Loss of a government or corporate mission critical chemical Loss of a corporate or nationally important economically critical chemical or economically critical facility Inherent safety (IS) principles are more applicable to potential consequences from some of these security issues than others and may be most readily apparent in the contex t of security release scenarios. If feasible, substituting a less hazardou s chemical, reducing the amount of hazardous chemicals used, or operat ing under more mo derate process

INTRODUCTION 3 It is important to use standard te rminology when referring to incident investigation so that those investigating an occurrence all share a common language that efficiently and accurately supports their investigation objectives. Some investigators may define the terms presented below slightly differently or use other descriptive term s that have the same meaning. Some organizations may desire to further su b-divide these terms into different levels. Within the scope of this book, the following definitions for key terms will apply throughout: 1.2.1.1 Incident —an unusual, unplanned, or unexpected occurrence that either resulted in, or had the potential to result in harm to people, damage to the environment, or asset/business losses, or loss of public trust or stakeholder confidence in a company’s reputation. Some examples are: • process upset with potential process excursions beyond operating limits, • release of energy or materials, • challenges to a protective barrier, • loss of product quality control, • etc. 1.2.1.1(a) Accident —an incident that sequence involving: • human impact, • detrimental impact on the community or environment, • property damage, material loss, • disruption of a company’s ability to continue doing business or achieve its business goals, (e.g. loss of operating license, operational interruption, product co ntamination, etc.). 1.2.1.1(b) N ear-miss —an incident in which an adverse consequence could potentially have resulted if circumstances (weather conditions, process safeguard response, adherence to procedure, etc.) had been slightly different. For most occurrences, protective ba rriers prevent a resultant adverse consequence. Such occurrences are often referred to as near-hits, near- misses, or close calls. For every inciden t labeled a near-miss, more subtle precursors exist that, if investigated and understood, coul d provide valuable insights into factors that could be app lied to mitigating or preventing other incidents. results in a significant con

Anatomy of a P&ID Sheet 17 For companies that have a Holds area on a P&ID, the reasons for the hold can be stated here. The importance of having a Holds area is that it prevents confusion. For example, the following conversation highlights this. The Notes area is where all the information that cannot be presented as symbols in the main body of a P&ID are stated. Because a P&ID is a visual document and tool, the Notes should be eliminated or at least the number of notes should be minimized. The Notes can be classified in different ways: main body notes and side notes; specific notes or general notes; and design notes or operation notes. Where placing a note on a P&ID is inevitable, the best location for a note should be determined, whether in the main body or the Notes block of the P&ID. If a note is specific to an item and is very short, it can be placed on the main body near the item. But if a note is general or long, it is better to be placed in a Notes block. Notes should not be more than three or four words in the main body of the P&ID. For a practical approach, main body notes are more preferable than side notes because they can be easily recognized compared with side notes that will be more likely overlooked. Notes in on the side could be specific or general. Specific notes are those that refer to specific points of the P&ID. Such notes should have the corresponding phrase “Note X” (X represents the note number in the Notes block) in the main body of P&ID. This concept is shown in Figure 3.7. One common problem in P&IDs are “widowed” notes, which are the notes in Notes area with no corresponding phrase in main body of the P&ID. General notes refer to any specific area or item on the P&ID. These do not have any corresponding Note X in the main body of P&ID. Notes can be classified as design notes or operation notes depending on their applicability. If a note is placed for the designers (piping, instrument, etc.) and for the design duration of a project, it is called a design note. If a note gives an information to operators or plant managers during the operation of the plant, it is called an operator note (Figure 3.8). Engineer 1: Oh, the pump flanges is still missing? Put the pump flanges sizes pleaseEngineer 2: Two month ago I received an email from the vendor and they said the flange sizes are tentatively 6” and 4” , but they never confirmed.Engineer 1: ok, put them here on P&ID for now. Later during pump installation it was found that the flange size by the vendor are not correct and they wanted to firm it up, which never happened. This problem could be prevented by putting tentative flange sizes, cloud it, and in hold area put a comment as: ”To be confirmed by vendor”0 No.D ateR evision22/10/20 10 Issued for appr oval RP BY ENG.LP PE APPR. APPR.CLFigure 3.5 Typical R evision block. Notes Holds1. Vents and drains shall be provided at the suitable locations during the construction. 2. Deleted3. Deleted4. PVRVs complete with bird screen. 1. The sizes of all relief valves are preliminary and to be confirmed during detailed engineering stage. 2. Centrifuge tie-point for vendor confirmation. Figure 3.6 Typical C omment block. Notes: Note 21. XXXXXX 2. Min. suction length Figure 3.7 Typical Not e in note side area.

THE UPSTREAM INDUSTRY 29 2.Protect all potential future commercial production zones 3.Prevent in perpetuity leaks from or into the well 4.Cut pipe to an agreed level below the surface and remove all surface equipment When an installation has reached the end of its life, process equipment must also be safely decommissioned and removed. Onshore this is the same as decommissioning a downstream plant. Offshore, most jurisdictions now require jackets to be recovered and taken to shore. Floating installations are towed to dismantling facilities, similar to oil tankers. An early decommissioning example was the Brent Spar rig in the North Sea in 1991 which originally was to be disposed by dumping in deep water. This caused controversy as there were public concerns about potential hazardous materials remaining onboard. While these were exaggerated, the operator later decided to dismantle the rig onshore and onshore disposal is now standard in the North Sea. Process Safety Issues Plugging and abandoning a well can be difficult as there may have been a change of ownership and lack of accurate records of the original design and subsequent changes to the well. A process safety issue observed for onshore wells has been small leaks in the well column that lead to sustained annular casing pressure and possible groundwater contamination. In the US onshore, decommissioning rules are usually based on State regulations and offshore, BSEE provides rules for plugging wells in a Notice to Lessees (NTL 2010-G05). Vrålstad et al (201 9) set out issues for offshore and common solutions in good detail. Production tubing and tools are removed and cement plugs, sometimes with supplemental mechanical plugs, are inserted at several locations up the casing. Generally, multiple plugs are required to address both the inside casing and annular spaces. 2.7 DEFINING “BARRIERS” A major topic for process safety management covering all aspects of upstream activity is barrier definition and their ongoing management. Barriers must maintain their effectiveness through life, and this is a challenge as many are rarely used – just ready to operate if a safety demand occurs . Since the barrier topic is relevant for Chapters 4, 5, 6 and 7, it is included here as an introduction. Specific barriers are discussed later in each chapter. Historically there have been differences in terminology regarding barriers between downstream and upstream. The term inology became more consistent with the widespread adoption of the LOPA method. A barrier, in LOPA – Layers of Protection Analysis (CCPS, 2011), is an independent protection layer (IPL) and must have specific attributes to meet this definition. Those parts of upstream similar to downstream (e.g., large gas plants and LNG facilities) also use LOPA and tend to follow the same terminology. Barriers are usually identified in a hazard

APPENDIX D – EXAM PLE CASE STUDY 395 Logic Tree (7 of 9)

8.5 Buncefield, Hertfordshire, UK, 2005 | 113 8.5 Buncefield, Hertfordshire, UK, 2005 The Buncefield oil storage depot experienced the severe consequences of overflowing a large fuel tank. Fortunately, there were no fatalities. The tank had two forms of level control: • A gauge that enabled employees to monitor the filling operation. • An independent high-level switch (IHLS) designed to shut down operations automatically if the tank level exceeded the level setpoint. At the time of the incident, the first gauge was stuck. Although site management and contractors were aware that the gauge stuck frequently (14 times from August to December), they did not attempt to resolve its unreliability. There was a general lack of understanding of how to use the IHLS, and as a result it was bypassed. There was no way to alert the control room staff that the tank was filling to dangerous levels. Eventually, the fuel overflowed from the top of the tank. A vapor cloud formed and ignited, causing a massive explosion and a fire that lasted five days. Prior to Buncefield, some experts believed that vapor clouds resulting from spilled gasoline could ignite but would not explode because they are unconfined. However, an explosion did happen. This has led to many theories and a significant amount of research. Some of the factors being considered include gasoline aerosolized and well-mixed with air by cascading over the wind rib; an unusually long vapor cloud; and even trees, shrubs, and parked vehicles near the spill providing confinement (Davis 2017 and Oran 2020). This remains an area of ongoing research. In the event of a fuel overflow, the site relied on a retaining wall around the tank and a system of drains and catchment areas to prevent liquids from being released to the environment. These containment systems were inadequately designed and maintained. Both forms of containment failed. Pollutants from fuel and firefighting liquids leaked from the bund, flowed off site, and entered the groundwater. 8.6 West, TX, USA, 2013 Echoing similar trends around the world, population growth around an industrial site was an issue in the West, TX, USA, disaster. Although the facility was initially built away from the center of population, the See Appendix index entry S3 See Appendix index entry C74

xxviii | Nomenclature References to process safety culture core principles: Throughout the book the names of the core principles of process safety culture are typeset in italics . Italics are also used when the context requires use of a different syntax, including the negative forms, such as “They allowed deviance to be normalized , leading to… ” Should vs. m ust and shall: The term should , used throughout the book, refers to actions or guidance that are recommended or presented as options, but not mandatory. The pursuit of process safety culture is very personal, and therefore a single approach cannot be mandated. The term s must and shall , commonly used in voluntary consensus standards and regulations, appear in this book only when quoting other sources. Quotes are offered only to provide perspective, and their use in this book does not m ean that the authors consider the quoted text to be mandatory.

54 INVESTIGATING PROCESS SAFETY INCIDENTS developers prepare their team by researching the basic incident investigation principles and priorities. This book is a good resource for orienting a development team. Developers can provide leadership to help the team determine which investigation methodologies best fit the particular culture of their organization. 4.1.4 Integration with Other Functions and Teams An active incident investigation will touch other functions within the organization. Preplanning for this inte raction begins during the development stage of the management system by identifying known areas of mutual interaction. The management system developers should review other existing management syst ems such as those liste d below to identify opportunities for integration and communication. • Crisis Management • Emergency response • Environmental protection • Employee safety • • Security • Regulatory compliance • Insurance interactions • External media communications • Corporate legal policies and procedures • Engineering design and risk re views (such as process hazard analyses or management of change reviews) • Accounting and purchasing practices • Quality assurance • Hazard Identification an d Risk Analysis (HIRA) One approach is to mesh all investigation and root cause analysis activities under one incident investigation management system. Such an integrated system should address all four business drivers: (1) process and personnel safety, (2) environmental responsibility, (3) quality, and (4) stakeholder interests. This approach works well since techniques used for data collection, causal factor analysis, and root cause analysis, can be the same regardless of the type of incident or business sector (i.e. not just petroleum or chemicals). Many companies realize that causal factors and root causes of a product quality or business continuity, etc. incident may also share a commonality with occupational or process safety incidents.

INCIDENT IN VESTIGATION TEAM 97 one involving a highly complex system. Team selection is dependent on the circumstances, complexity, and severity (actual or potential) of the incident. Although highly experienced investigators may not be needed for every investigation, seasoned investigators can still support an investigation through consulting, quality assurance, and peer review. It may not be necessary for the lead investigator to be part of the line management team; however, it is important th at the leader of the inve stigation be provided with adequate training, coaching, support and authority by management. 6.2 ADVANTAGES OF THE TEAM APPROACH There are several advantages to usin g a team approach when performing incident investigations. 1. Multiple technical perspectives assist in analyzing the findings — A structured analysis process is used to reach conclusions. Individuals with diverse skills and perspectives best support this approach. 2. Diverse personal viewpoints enhance objectivity — In comparison to a single investigator, a team is less likely to be subjective or biased in its conclusions. A team’s conclusions are more likely to be accepted by the organization than the conclusions of a single investigator. 3. Internal peer reviews can enhance quality —Team members with relevant knowledge of the analysis process are better prepared to review each other’s work and provide constructive critique. 4. Additional resources are available —A formal investigation can involve a great deal of work that may exceed the capabilities of one person. Quality may be compromised if one person is expected to do most of the work. 5. Scheduling requirements are easier to meet —Deadlines set by management, outside parties, or the team leader may require several activities to be performed in parallel. This demands a team approach. 6. Regulatory authority may require a team approach— Specific regulations, such as OSHA’s Process Safety Management regulations in the US, call for a team approach. Management needs to be aware of whether a facility falls under such regulations. 7. W orkforce involvement – Participation in an incident investigation team provides an opportunity to engage with workers, for them to learn, as well as to contribute, and to build support for the recommendations with peer workforce members.

xxii GUIDELINES FOR MANAGING ABNORMAL SITUATIONS Boiling Liquid Expanding Vapor Explosion (BLEVE) A type of rapid phase tran sition in which a liquid contained above its atmospheric boiling point is rapidly depressurized, causing a nearly instantaneous transition fr om liquid to vapor with a corresponding energy release. A BLEVE of flammable material is often accompanied by a large aerosol fireball, since an external fire impinging on the vapor space of a pressure vessel is a common cause. However, it is not necessary for the liquid to be flammable to have a BLEVE occur. Bow Tie Model A risk diagram showing how various threats can lead to a loss of control of a hazard and allow this unsafe condition to develop into a number of undesired consequences. The diagram can also show all the barriers and degradation controls deployed. Conduct of Operations The embodiment of an organization's values and principles in management systems that are developed, implemented, and maintained to (1) structure operational tas ks in a manner consistent with the organization's risk tolerance, (2) ensure that every task is performed deliberately and correctly, and (3) minimize variations in performance.

E.22 Stop Work Authority/Initiating an Emergency Shutdown |307 day, in the same group as employees, and doing sim ilar jobs. The only difference was that their paychecks and benefits came from their own employer, not the host facility’s company. B ecause of this strict policy, the resident contractors in the instrument shop did not attend the daily toolbox meeting and did not receive som e key process safety related inform ation. Consequently, a resident contractor instrument technician m ade an error perform ing a proof test, and a m inor incident resulted. The root cause analysis revealed that facility instrument technicians received the specific knowledge that was given to the at the toolbox meeting, but contract instrument technicians did not. The com pany expected that this inform ation would be relayed through the contractors’ employer, but it was not. How can leaders work with the Human Resources function to assure contractors receive needed process safety inform ation? Ensure Open and Frank Communications. E.22 Stop Work Authority/Initiating an Em ergency Shutdown A high-risk facility has clearly written procedures giving on-duty operators the authority to initiate an em ergency shutdown when the conditions warrant, without obtaining any other approval. A review of operator training m aterials shows that this authority is clearly and explicitly stated. Operators have stated in interviews that the procedures were known and understood and confirmed that the training was given as it appears. Despite this policy, an operator on duty in the control room did not initiate an em ergency shutdown during a significant transient event in which the process pressure and tem perature rose rapidly due to a runaway reaction. This incident resulted in a significant release of flam mable m aterials and a vapor cloud B ased on Real Situations

RISK ASSESSMENT 335 HSE c, “Review of human reliability assessment me thods”. U.K. Health and Safety Executive, Health and Safety Laboratory, 2009, https://www.hse.gov.uk/research/rrpdf/rr679.pdf HSE Failure Rate, “Failure Rate and Event Data for use within Risk Assessments”, U.K. Health and Safety Executive, Chemicals, Explosives and Microbiological Hazardous Division 5, https://www.hse.gov.uk/landus eplanning/failure-rates.pdf. NUREG, U.S. Nuclear Regulatory Commission, "Handbook of Human Reliability Analysis with Emphasis on Nuclear Power Plant Applic ations - Final Report", CR-1278, https://www.nrc.gov/docs/ML0712/ML071210299.html. IOGP 2019, “Risk assessment data directory - Process release frequencies” , The International Association of Oil and Gas Producers Report 434-01. First, Risk matrix, personal correspondence, 2021. OREDA, Offshore and Onsh ore REliability DAtabase, https://www.oreda.com . OSHA 1990, “Phillips 66 Company Houston Chemical Complex Explosion and Fire: A Report to the President”, Occupational Health and Safety Administration, U.S. Department of Labor, Washington, D.C. 1990. Primatech, LOPAWorks, https://www.primatech.com/software/lopaworks Rasmussen 1975,"Reactor safety study. An assessment of accident risks in U. S. commercial nuclear power plants. Executive Summary", WASH-1400 ( NUREG75/014), U.S. Nuclear Regulatory Commission. RPA 1982, “Risk Analysis of Six Potentially Hazard ous Industrial Objects in the Rijnmond Area, A Pilot Study”, Rijnmond Public Authority, Springer. Spouge/IChemE 2006, “Leak Frequencies from the Hydrocarbon Release Database”, Symposium Series 151, Institute of Chemical Engineers, Rugby, England. Williams 1985, “HEART – A Proposed Method for Achieving High Reliability in Process Operation by means of Human Factors Engineering Technology”, Proceedings of a Symposium on the Achievement of Reliability in Operating Plant, Safety and Reliability Society, 16 September 1985, Southport, England. Williams 1986, “A proposed Method for Assessi ng and Reducing Human Error”, Proceedings of the 9th Advance in Reliability Technology Symposium, University of Bradford, England. Williams 1988, “A Data-based method for asse ssing and reducing Human Error to improve operational experience”, Proceedings of Institut e of Electrical and Electronics Engineers 4th Conference on Human Factors in power Plants, Monterey, California. Williams1992, “Toward an Improved Evaluati on Analysis Tool for Users of HEART”, Proceedings of the International Conference on Hazard identification and Risk Analysis, Human Factors and Human Reliability in Process Safety, Orlando, Florida.

15. Fatigue and staffing levels 173 Figure 15-6: Signs and symptoms of fatigue People may be screened for signs of fati gue at the start of shifts and monitored throughout a shift. Tasks should allow for rest breaks throughout the shift. Fatigue detection technology can monitor ey e closures and head posture. 15.3.4 Task scheduling As noted in section 15.2.3, people’s sleep/wake cycle means that complex tasks may best be scheduled for the start of th e day and the start of the work block, when people are most alert and rested. Work planning should avoid scheduling complex tasks for night shifts, after lunch (13.00 to 15.00), or during early starts. There are differences in people’s fatigue risk. Those at greatest risk can include older people, especially over 50 years of age, people with a challenging home sleep environment (e.g., with young children) , and those with long commutes. The allocation of work and the scheduling of tasks may take account of individual needs. For example, the tasks requiring th e highest level of concentration should be allocated to individuals at lower risk from fatigue.

138 | 4 Applying the Core Pr inciples of Process Safety Culture com ments above regarding these groups apply to the media as well. When incidents happen, however, the media has the potential to have im pacts that go beyond their role as a com munication vehicle. Media m ay sensationalize bad news out of traditional journalistic practices or to promote readership or viewership. Media frequently over-sim plifies stories, om itting facts that that could have placed the facility in a better light. They may also repeatedly run stark images of the worst of the incident with dram atic voice-over, reinforcing a negative im age in the m inds of the reader or viewer (Ref 4.10). They may do this for innocent reasons such as to meet a deadline, to fit space available, to make the story broadly understandable, or because they are waiting for more information. However, in recent years, numerous publishing, broadcast, and social m edia outlets have em erged that intentionally slant the news. Som e of the slanted news could be m ore favorable to the facility, while others could be less favorable. And today, it is not unusual for anyone with a cellphone to capture video and broadcast their view of what they are seeing. Generally, the facility should engage with the media and not try to avoid it or stonewall. Trying to fight or avoid the media will typically backfire, resulting in the negative im pact of the sim ple phrase “The facility refused to comment.” Facility leaders should be aware that m edia m ay attempt to contact employees, hoping to get an inside story about an incident or a sound bite that supports the story. It is im practical and not even desirable to provide all employees with media training. However, if the process safety culture is strong, the em ployees’ com mitment to process safety should come through.

Piping and Instrumentation Diagram Development 228 The sample technical information of a PSV is shown in Figure 12.13 and for a rupture disk in Figure 12.14. 12.14 Selecting the Right Type of PRD Arr angement The PSDs could be installed in different arrangements. The simplest arrangement is a single PSD. Multiple PSDs could be placed in some cases. They could be is series or in parallel. When they are in series they could be the same type or different types. In the series arrangement it could be a PSV and a rupture disk or two rupture disks in series. When they are in parallel arrangement they could be all functional (such as a 2 × 50% or 3 × 33% ar rangement) or a functional‐spare arrangement (such as the 2 × 100% spar e philosophy). The parallel arrangement can also be classified based on the similarity or dissimilarity in types. There could be mul-tiple PSVs in parallel, or multiple rupture disks in parallel, or few PSVs and few rupture disks in a parallel arrangement. The different arrangements of PRDs can be seen in Figure 12.15.The single PSV is the default choice to protect process enclosures (Figure  12.16). The other option is using a single rupture disk. However if there is a need to inspect or maintain the PRD while the rest of the plant is operating, the PRD needs an isolation system (isolation valve and drain/vent valves) and some other provisions. The provisions are the systems to allow pulling the PRD out of operation and to allow an operator to func - tion as a “PRD” for the time the PRD is out of operation. There should be a connection to connect a portable pressure gauge (or an already installed pressure gauge in place). This system should be completed with a bypass pipe and a manual throttling valve on it. During the time the PRD is isolated from the to‐be‐protected system, an operator needs to watch the pressure gauge and also be ready to open the manual throttling valve to release the pressure if the pressure goes higher than allowable. Such a “manual safety system” is shown in Figure 12.17.However not in all cases such “simple” system can be used. The other systems to support inline care for PRD’s will be discussed later here. There are some cases that a single PSD doesn’t satisfy the requirement of safety. In such cases PSDs in parallel or in series are used. PSDs can be placed in parallel for different reasons.A non‐exhaustive list of cases in which we may need to use parallel PRD’s is: ●When the required PRD is bigger than the maximum available PRD in the market and several smaller PRDs in parallel and all functioning are required. ●When the required PRD is big and for economical rea-sons it is better to use several smaller PRDs in parallel and all functioning. ●When there is a difference of more than 20–30% of release flow in different credible scenarios, two (or more) PSE 1234 Governing casePSV pops at “a little” higher than 200 KPagBurst pressure: 200 KPagHolder size: 2”Case: Fire Figure 12.14 Rupture disks on P&IDs . • Single PSD • Multiple PSD• In parallelAll functioning Spare The same type Different type The same type Different type• In series Figure 12.15 Differ ent PRD arrangements. Figure 12.16 A single simple PRD.

8 • Emergency Shutdowns 142 There has been a loss event requiring a safe emergency response; There has been an unpredictab le natural hazard event (e.g., an earthquake, lightning strike). These shut-downs were illustrated in Figure 6.2, the transient operating modes associated with emergency operations. Depending on the facility's emergency response resources and on the extent of the loss event, the Emergency Response Team (ERT) and the Emergency Response Plan (ERP) may have to be activated. Similar to an unscheduled shut-down, thes e shut-downs may have special procedures, checklists, and decision aids to address other potentially hazardous conditions that may occur during the shut-down. If there is a loss of containment event, there might be additional PPE required during the emergency response when using the special emergency response procedures, activating any Emergency Shutdown Devices (ESD), and activating the ERT. 8.3 Safely responding to an incident The goal of an emergency response during and after an emergency shutdown is to ensure that everyone is safe, that the injured, if any, are reached and cared for quickly, an d that the loss event causing the shut-down is contained as quickly an d as safely as possible. All types of emergency responses to shut-downs are based on thoroughly written and implemented emerge ncy response procedures. The preplanning procedures that help en sure everyone’s safety during a loss event includes thoroughly written and implemented plans, such as: An Emergency Action Plan (EAP ) which describes the actions of everyone, including contrac tors, when an incident is occurring at the facility, and

Evaluating Operating Experience Since the Prior PHA 69 gained since the last PHA. The discussion explains how experience gained since the prior PHA can affirm that the Update approach will be sufficient, or if a Redo to address certain errors or deficiencies would be more prudent. If significant deficiencies are found in the operatio nal experience records, management should be advised of the issues so they can be addressed outside of the revalidation process. 4.2.1 MOC and PSSR Records When considering operational history, the number and extent of changes since the prior PHA will have the biggest effect on the revalidation effort. The study leader’s first task is to select a “fr eeze” date for the revalidation. Changes commissioned before the freeze date are included in the revalidation; those occurring afterward will be included in th e next revalidation cycle. Typically, the freeze date is about a month before the start of the revalidation meeting(s), so the team will have reasonable time to prepare for the revalidation without constantly revising documentation for the meeting. Sometimes a natural break i n t h e f l o w o f M O C s m a k e s i t e a s y t o s e l e c t a f r e e z e p o i n t . F o r e x a m p l e , t h e revalidation might include all changes through the October shutdown; any subsequent changes would be included in the next revalidation. The freeze date should be documented in the revalidated PHA, so the next revalidation team will know where to start its review of operating experience. In general, the more changes and the more complex the changes that have occurred before the freeze date, the more time PHA revalidation team meetings will require and the more likely a Redo will be advantageous. However, the complexity of MOCs cannot be considered in isolation – the robustness of the MOC program is equally important. Sp ecific items to consider include: Terminology: Management of Change , Pre-Startup Safety Review, and Operational Readiness The MOC program should have a record of all process changes. There is usually a companion record kept by the PSSR program of inspections performed to ensure that any physical changes are ready to be put in service. The RBPS model expands PSSR to Operational Readiness (OR), which ensures that processes are safe to restart, even when no permanent changes were intended. Thus, the needed operational history records may be found in MOC, PSSR, and/or OR documentation.

26. Learning from error and human performance 337 Preventing reoccurrence of errors requires people to: • Fully understand why an error occurred (the causal factors). • Determine trends/patterns in errors. This is done by analyzing whether the cause of an error is unique to that particular task and set of circumstances, or whether it occurs in different activities and situations. In addition, depending on the type of error (such as skill-based versus knowledge-based errors), the cause and solution will differ, as shown next: • Lapses (forgetting a step/a ction) can be caused by fatigue or distractions. • Slips (completing the action incorrectly) can be due to poor layout of controls. It is only possible to identify solutions on how to avoid error, by understanding the factors that contributed to the error. For example, if an error occurred because of fatigue due to lengthy working hours, the shift system will require review and adjustment. This includes ensuring that there are more than two individuals on safety critic al tasks and adding in more frequent breaks (to maintain situation awareness). Reoccurrence of error and repetition of accidents, incidents, and near misses is also linked to: • Failure to perform adequate root cause analysis. • Not applying identified improvements. • Applying ineffective improvements be cause of poorly conducted root cause analysis. 26.2.2 Learning process Incidents and errors offer valuable lear ning opportunities and lessons. If these lessons are acted upon, they will help pr event reoccurrence of errors, and also enable improvement in the way risks are managed. Several steps are required to achieve learning. These steps are provided in Figure 26-1. See Chapters 2 and 3 for more information on errors, types of error, and the SRK (Skills, Rule, Knowledge) model. More information on error solutions (improvements and solutions) is provided in section 26.5.4.

APPENDIX D – EXAM PLE CASE STUDY 373 Sequence of Events and Description of the Incident On August 1 at 10:30 A.M., a control room operator remotely started the feeds to Kettle No. 3 in the catalyst preparation area. The normal procedure was to fill the kettle to approxim ately 80%, but Kettle No. 3 was apparently completely filled this time. The level indicator showed a high level, but the alarm did not sound. (The alarm was later found to be bypassed.) A high-pressure alarm for this vess el was acknowledged at 11:03 A.M. by the control room operator. At 11:00 A.M., a severe thunderstorm had started and within 5 minutes caused a power outage throughout the immediate vicinity. The ambient temperature wa s about 83 °F and winds were from the northwest at about 3 mph. With an available diesel emergency gen erator supplying power to critical pumps, the control room operators initiated shutdown procedures for the two reactor areas. An uninterrupti ble power supply (UPS) kept power to the DCS screens and instruments; however, the DCS system cl osed all catalyst preparation and reactor feed valves on loss of power as designed. Outside operators were sent to manually block in reactor feeds. At 11:09 A.M., a high-LEL detector in th e catalyst preparation area sounded on the DCS. The lead outside operator was contacted by radio communications to investigate the problem. He said he was just leaving the Reactor No. 1 area and would go righ t to the catalyst preparation area. The thunderstorm had pa ssed overhead and the rain was diminishing. At about 11:10 A.M., a “whooshing” noise (now believed to be the fireball) was heard by many and the heat detector for the automatic water-spray sprinkler coverage in this area alarmed in the control room. The lead outside operator did not respond when called on the radio. The plant fire brigade and the lo cal volunteer fire department were notified by the supervisor of the ca talyst preparation area by 11:12 A.M. On their arrival to the scen e of the fire at 11:15 A.M., the plant fire brigade saw the lead outside operator down abou t 40 feet from the fire, in between the catalyst preparation area and reacto r building No. 1. They also found a seriously burned unknown person abou t 120 feet from the fire, near the finishing building. (This person was ev entually determined to be a service contractor who entered the premises at 10:30 A.M. to calibrate equipment in the instrument house for Reactor No. 1.) The fire had engulfed most of th e catalyst prepar ation area. The automatic deluge sprinkler coverage for this area had actuated, but water did not flow. The fire brigade tri ed to activate a fixed monitor, but again got no water flow. With the limited water supply from the plant fire engine available as a shield, the fire brigade members felt they could reach the lead outside operator. Meanwhile, the commander of the plant fire brigade

Part 4: Operational competence Human Factors Handbook For Process Plant Operations: Improving Process Safety and System Performance CCPS. © 2022 CCPS. Published 2022 The American Institute of Chemical Engineers.

8 • Emergency Shutdowns 153 8.5.4 Rebuilding equipment and processes after a major incident After a major fire, explosion, flood or hurricane, there will be significant property damage that will need to assessed, with equipment repairs or replacement required before the facility can resume operations. If the affected processes will be rebuilt, the post- incident projects may have consid erable business pressure to re- establish production due to co mmitments to supply products to customers safely and quickly. Comp anies may decide to duplicate the original design specifications for the rebuild. However, if whatever materials are available on short delivery are procured, even if the specifications are not identical to th e original process equipment, the facility needs to have an effectiv e management of change system to ensure that the process safety risk s, as well as the demolition and construction risks, associated with the changes can be effectively managed. An additional issue that may need to be addressed includes designing and constructing the replaced equipment to meet the current design standards since older equipment may have been build years ago. Additional discussion for effectively managing engineering projects, especially when significant equipment is being rebuilt, were described briefly in Chapter 4. 8.6 Incidents and lessons learned Details of some emergency shutdown -related incidents are included in this section. The incident su mmary is provided in the Appendix. 8.6.1 Incidents during an emergency shut-down General note: Although the emergency shut-down resulted in an incident, the severity of the incide nt was reduced due to the mitigative

CASE STUDIES/LESSONS LEARNED 175 shows the PF (First Officer “F/O”) in put from the stick in blue and the actual pitch of the aircraft in brown (BEA 2012). Figure 7.2 Aircraft Pitch Co mmands and Pitch Attitude from 02:10:05 to 02:10:26 At 02:11:33, the PF stated, “I don’t have control of the airplane any more now.” At about 02:11:42, the Captain re-e ntered the cockpit, 1½ minutes after the autopilot disconnect. By this time, the aircraft was in a rapid descent and at 02:11:43, the PNF stated, “What’s happening? I don’t know I don’t know what’s happening” . Two seconds later, the PF stated, “We’re losing control of the aeroplane there” . After 2:11:45, the stall warning tri ggered another ten times, two of which coincided with a pitch-up input by the PF. At about 02:12:00, as the aircraft dropped through 31,500 feet, the angle of attack was around 40 degrees (nose up) and the altitude was dropping at a rate of 10,000 ft/min. In the period 02:12:00 to 02:14:07, the aircraft dropped to 4,000 feet, with the pilots pitching the aircraft up and attempting to regain control. From 2:14:17, the Ground Proximity Warning System (GPWS) “sink rate” and then “pull up” warnings sounded, and the recordings stopped at 2:14:28. The final data from the f light recorder showed a descent rate of 10,912 ft/min, a ground speed of 107 knots and a pitch attitude of 16.2 degrees nose-up. No emergency mess age was transmitted by the crew.

206 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Figure 11.22. Schematic of baghouse (Donaldson-Torit) Figure 11.23. Dust collector explosion venting (Fike)

25 Loss Event : Point of time in an ab normal situation when an irreversible physical event occurs that has the potential for loss and harm impacts. Examples in clude release of a hazardous material, ignition of flammable vapo rs or an ignitable dust cloud, and over pressurization rupture of a tank or vessel. An incident might involve more than one lo ss event, such as a flammable liquid spill (first loss event) followed by ignition of a flash fire and pool fire (second loss event) that heats up an adjacent vessel and its contents to the point of rupture (third loss event). 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 environmen tal damage, and/or magnitude of losses such as property damage, material loss, lost production, market share loss, and recovery costs. The most effective strategies will e liminate a hazard. Inherently safer design can also reduce the potent ial for propagating an incident sequence before there are major im pacts on people, property, or the environment. Inherently safer concepts should be an essential aspect of any process safety program. If the hazard s can be eliminated or reduced, the extensive layers of protection to control those hazards may not be required or may be less robust. Howeve r, inherent safety is not the only process risk management strategy av ailable and may not always be the most effective solution. For example, capital expenditure limitations may preclude the wholesale adop tion of a new, inherently safer process in an established, depreciated chemical plant. Targeted application of inherently safer design techniques to parts of the existing process, coupled with additional layers of protection may be the only economically feasible option. A system of strategies that includes both inherently safer design and addition al layers of protection may be needed to reduce risks to an acceptable level. The application of inherent safety co ncepts in a hierarchy of controls is shown in Figure 2.3. The steps of managing risk should ideally be executed in a hierarchical manner and iteratively. The ability to apply IS strategies following the logic of the hierarchy of controls may be restrained.

3.2 The Impact of Company Culture | 33 events. This can lure leaders and employees alike into thinking that process safety risks are lower than they really are, or that the process safety problem has been solved. This misperception can be further enhanced if the company is improving performance in occupational safety. Despite their common use of the word safety (and a few areas of overlap), process safety and occupational safety differ significantly in their general competencies, technologies, and management structures. Yet improved performance in occupational safety can lead to intentional or unintentional de-prioritization of process safety, de- emphasizing continuous learning. Risk misperception can also occur when companies have previously had negative HSE experiences with one chemical—or even in a single HSE area. As a result of past experiences, the perceived risk seems higher than it might have otherwise. A company with past landfill liability may view managing its landfills as riskier than any process safety hazard and therefore divert process safety resources. Similarly, a company that has had many workers get cancer from occupational exposure to a chemical may manage that chemical as toxic—and only toxic— although it is also flammable and reactive. Ultimately, if a practice is considered less important, it’s easy to just forget about it. Make it your goal to ensure that the importance of maintaining and continuing to improve process safety becomes deeply embedded in the corporate culture—so that these misperceptions can never occur. Compliance Mindset and Anticipation of Litigation CCPS (CCPS 2019b) has described the perils of relying on the compliance- only mindset: • Regulations are designed to protect society, not to protect the company. • Regulations don’t cover everything you need to do. • Regulators inspect only rarely. • Regulators don’t know your process as well as you do. A finding of compliance by a regulatory agency is no guarantee that you are managing your process safety risks adequately. However, such a finding may fool people into thinking there is no need for improvement. In this kind of environment, there is no incentive for learning. A compliance-only mindset often goes together with a mindset of continuously expecting litigation. The legal outcomes of both criminal and civil

302 | Appendix E Process Safety Culture Case Histories This exam ple shows both good and bad examples of the role of leadership in process safety culture. What are they? Combat the Normalization of Deviance, Provide Strong Leadership, Maintain a Sense of Vulnerability. E.15 Check-the-Box Process Safety M anagem ent System s A corporate process safety audit found that the documentation for key process safety activities at a facility was extremely sparse. Previous internal audit reports consisted of 2-page memos. PHA reports of major process units contained 10 pages of worksheets and these contained m any blanks. Incident investigation reports contained root cause analyses that were described in a brief paragraph. Further interviews revealed that these documents were created as the result of activities intended m ainly to get activity off the facility’s to-do list. The auditors pointed out that such practices and the thin docum entation did not reflect typical industry practices for those PSMS elem ents. The Facility M anager and members of his management team reacted angrily. They stated forcefully that the facility had never suffered a process safety incident and that their docum entation m et the minim um regulatory. This, they said, was proof enough that no additional effort was required or needed. What other symptom s of weak process safety culture do you believe existed at this facility? Combat the Normalization of Deviance, Understand and Act Upon Hazards/Risks, Establish and Imperative for Safety, Provide Strong Leadership. E.16 There’s N o Energy for That Here During PHAs at a facility, team leaders typically screened the recommendations made by the team B ased on Real Situations Actual Case History

APPENDIX D – EXAM PLE CASE STUDY 391 Logic Tree (3 of 9)

144 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION FM Global 2019, Property Loss Prevention Data Sheet FM 1-40 “Flood”, FM Global, Hartford, Connecticut. HSE, Health and Safety Executive, https://www.hse.gov.uk/comah/sragtech/casemarathon87.htm. IAEA 2015. “The Fukushima Daiichi Accident”, Technical Volume 1/5 Description and Context of the Accident, ISBN 978–92–0–107015–9 (set), International Atomic Energy Agency, Vienna, August. Reproduced with permission IFW 1982, Industrial Fire World , https://www.industrialfireworld.com/540292/tacoa-venezuela- dec-19-1982 . IWP&D 2010, “Sayano Shushenskaya accident – presenting a possible direct cause”, International Water Power & Dam Construction, 22 December 2010, https://www.waterpowermagazine.com/featur es/featuresayano-shushenskaya-accident- presenting-a-possible-direct-cause. KDA, https://agriculture.ks.gov/divisions-progra ms/dwr/floodplain/resources/historical-flood- signs/lists/historical-flooding. LANL, Los Alamos National Laboratory, Electrical Safety Hazards Handbook , https://www.lanl.gov/safety/elect rical/docs/arc_flash_safety.pdf . MAC, The Mining Association of Canada , https://mining.ca/our-focus/tailings- management/tailings-guide/ Madehow, http://www.madehow.com/Volume-3/Fertilizer.html NAIIC 2012. “The official report of The Fukush ima Nuclear Accident Independent Investigation Commission”, The National Diet of Japan, 2012. NASA, National Aeronautics and Space Administration, https://earthobservatory.nasa.g ov/images/43883/eruption-of-eyjafj allajakull-volcano-iceland. NHC, National Hurricane Center, https://www.nhc.noaa.gov/surge/. NOAA 2019 a, National Oceanic an d Atmospheric Administration, https://noaa.maps.arcgis.com/apps/MapSeri es/index.html?appid=d9ed7904dbec441a9c4dd7 b277935fad&entry=1. NOAA, National Oceanic and Atmospheric Administration (NOAA), https://www.spc.noaa.gov/new/SV Rclimo/climo.php?parm=allTorn. UNECE, Natech, https://unece.org/industrial- accidents-convention-and-natural-disasters- natech. USGS 2018, U.S. Geological Survey, https:// www.usgs.gov/news/post-harvey-report-provides- inundation-maps-and-flood-details-l argest-rainfall-event-recorded.

276 INVESTIGATING PROCESS SAFETY INCIDENTS 11.3 OTHER REFERENCES The CCPS Human Factors Methods for Improving Performance in the Process Industries (CCPS, 2007) provides a basic overview of human factors topics in the process industries. The EI Learning from Incidents, Accidents and Events (EI, 2016) describes the learning from incidents process, from investigating to learning. The UK Civil Aviation Authority Flight- crew human factors handbook (Civil Aviation Authority, 2014) includes a very good theoretical explanation of human processes and behaviors presented in a simplified way. 11.4 SUM M ARY This chapter discussed human factors concepts including human action types and classes of human failures. It also presented human factors models including the facilities/equipment, people and management system model presented by CCPS (CCPS 2007) and the SRK (Rasmu ssen, 1983) mental processes model. The other references noted in Section 11.3 provide greater detail on the topic of human factors. Human factors are important before , during and after an accident investigation. Before the investigation, leadership can set the tone about the importance of learning from incidents (as opposed to placing blame). During the incident digging beyond the causal factors to understand why a human behaved a certain way can lead to un derlying root causes in management systems. Creating recommendations that address thes e underlying root causes will aid in preventing a wide range of similar incidents (and not just prevent the one incident fr om recurring). A good investigation considers the impact of human factors, strives to un derstand the underlying root causes in human factor/management syst em terms, and then makes recommendations aimed at setting the human up for success.

210 | Appendix: Index of Publicly Evaluated Incidents Section 4: Additional Index Terms (Continued) Index Term See: Fatigue Causal Factor Human Factors Flame Arrestors RBPS Element Asset Integrity Freeze Protection RBPS Element HIRA Hazard Communication RBPS Elements Training, Workforce Involvement, or Process Knowledge Management, as appropriate Hiring RBPS Element Management of Change (includes organizational change) Hot Work RBPS Element Safe Work Practices Inherently Safer Design Causal Factor Safe Design, and RBPS Elements Compliance with Standards, Process Knowledge Management, or HIRA, as appropriate Intoxication, Alcohol, and Drugs RBPS Element Training and Performance Assurance Isolation RBPS Element Safe Work Practices Job Change RBPS Element Management of Change (includes organizational change) Job Hazard Analysis, Job Safety Analysis RBPS Element Safe Work Practices Line and Equipment opening RBPS Element Safe Work Practices Lockout/Tagout RBPS Element Safe Work Practices Maintenance Job Planning RBPS Element Asset Integrity Management RBPS Elements Management Review, Continuous Improvement, and Conduct of Operations, and the Culture Core Principles, as appropriate Non-routine Work RBPS Elements Safe Work Practices or Asset Integrity, as appropriate Organizational Change RBPS Element Management of Change Permits to Work RBPS Element Safe Work Practices Positive Material Identification RBPS Element Asset Integrity Process Control RBPS Elements Process Knowledge Management or Management of Change, as appropriate

EVIDEN CE IDEN TIFICATION , COLLECTION & M ANAGEM ENT 151 Table 8.2 Examples of Paper Evidence Examples of Paper Evidence M anagement Policy and Programs • Company safety policy • PSM program and procedures • Contractor records/procedures/policy manuals Site details • Site description • Construction project files • Site map/ plot plan / firewater plan Design/ Hazard Analysis • Material safety data sheets (MSDS) • Operating procedures, ch ecklists, and manuals • Piping and instrumentation drawings • Material and energy balances • Process specification sheets • Equipment installation drawings • Equipment Engineering drawings • Electrical area classification drawings • Process hazard analyses (PHA) • Design calculations and design basis assumptions • Scenarios for the sizing of relief, venting, and emergency equipment • Dispersion calculations • Descriptions of normal and abnormal chemical reactions, including incompatibilities • Consequence analysis study results • Safe operating limits • Alarms and set points for trips • Instrument and electrical drawings • Interlock drawings • Ladder logic drawings • Control system software logic • Engineering standards and codes • Management of change (M OC) records • Prior incident investigation reports • Completion of actions from PHAs, MOC and previous incidents Operating / M aintenance Data • Shift log sheets • Run histories / Batch sheets • Process data records—strip and circular charts • Raw material quality control records • Retained sample documentation • Quality control (QC) lab logs • Work permits • Lockout–tagout procedures and records • OEM manuals • Maintenance procedures Inspection Data • Maintenance and inspection records • Repair records • Corrosion data • Test/inspection procedures Incident Data • Meteorological records • Phone logs • Emergency responder logs • Printed event logs • Gate/building entry/exit logs Personnel • Training manuals and records • Professional qualifications • Job instruction development • Supervisor selection criteria • Supervisor training requirements • Aptitude exams • Physical exams • HR records • Supervisor appraisal • Employment application

198 Example 8.3 A plant produced methyl methacryla te by reacting hydrogen cyanide with acetone to produce acetone cyanohydrin, followed by further processing to produce methyl me thacrylate. The hydrogen cyanide was produced at another site and was transported to the methyl methacrylate plant by rail. A hydrogen cyanide plant was subsequently installed at the meth yl methacrylate plant site to eliminate the need for shipping hydrogen cyanide or acetone cyanohydrin. Example 8.4 A company produced bromine in Arkansas and brominated compounds in New Jersey. A risk assessment resulted in a recommendation to consider the transfer of the bromination processes to the bromine producti on site in Arkansas. Economic considerations and the decrease in risk justified such a transfer, and it was implemented. Although safety was not the only consideration, it was an important factor in this decision. 8.10.2 Shipping Conditions The physical condition and characte ristics of the material shipped should be considered in transportation risk assessments on a case-by- case basis. There may be options ava ilable to reduce the transportation risk by reducing the potential for rele ases or the severity of the effects of releases. A few possible ways of improving safety by modifying conditions, and hence using the IS strategy of Moderation or Minimization are: Refrigerate and ship the material at atmospheric pressure or at reduced pressure. Ship materials in a concentrated state to reduce the number of containers, then dilute the concentrate at the user site. Ship and use the material or a su bstitute in diluted form, i.e., aqueous ammonia instead of a nhydrous ammonia, or bleach instead of chlorine.

138 Figure 8.1: Stages in the Life Cycle of a Chemical Process The notion that there is always a safer chemical to use is not correct; it is not always possible to substitute a less hazardous chemical to achieve the reaction needed to produce the desired molecule(s). Sometimes there is only one known way to produce a desired product. However, if there is an alternative chemistry that involves a different way to produce a given chemical this is the optimum stage to employ it. Sometimes this will require first- principles basic research to be conducted. Such activities have a fair amount of uncertainty because they may not result in the desired pr oduct. Therefore, process engineers

395 15.3.1 An Exothermic Batch Reaction An existing semi-batch process is used to carry out an exothermic reaction: In addition to the highly exothermic nature of this reaction, there is an additional hazard. If Reactant B is significantly overcharged (double charge or more), a side reaction can occur that generates thermally unstable by-products, and a runaway reac tion can result. Similarly, if the batch temperature gets too high, there is a potential for a thermal runaway due to undesired side reactions. A simplified version of the proce ss equipment is shown in Figure 15.3. Reactant A is dissolved in Solvent S in Weigh Tank A, and the solution is then fed to the reactor. Catalyst C is added to the reactor, and cooling is started to the reactor c ooling coils. Reactant B, the limiting reagent, is then gradually fed to th e reactor by gravity addition from Weigh Tank B, which has been pre-charged with the proper amount of reactant. The feed rate of Reacta nt B is controlled by the batch temperature. The process has a safety instrumented system (SIS) that takes it to a safe state when abno rmal operation is detected. The SIS takes action if the reactor agitator fa ils, or cooling fails, or high reactor pressure or temperature is detected . All of the Safety Instrumented Functions (SIFs) stop the feed of th e limiting reagent, Reactant B, by closing the feed flow control va lve and a dedicated block valve. As a result of the findings from a PHA, the system was modified as shown in Figure 15.4. The modified system includes several inherent safety features, which were implemented in this existing plant with a relatively small investment. 1.The Reactant B feed tank has been moved to the floor below the reactor level, and the feed is now controlled by a metering pump. In case of high temperature, the metering pump is turned off and a block valve is closed by the SI S on reactor high temperature. It is less likely that there will be a significant leak through the

66 INVESTIGATING PROCESS SAFETY INCIDENTS Table 4.1 Suggested Training for Effective Implementation Complex Incidents Investigation Team Leader Training M oderate/ M inor Incident Investigation Team Leader Training Incident and Near-miss Reporting/ Notification Awareness Training These leaders will handle the most complex incidents (top 10% or less) These leaders will handle low to moderate complexity incidents (90% or more of the incidents All operations and maintenance staff; appropriate purchasing, accounting, and other staff All staff individuals may fill any role in the system; this is the starting module of training Individuals who are expected to identify and report all incidents, including near-misses. Some of these individuals may become team leaders or members or may be interviewed during an investigation. Training Agenda • Investigation planning • Data protection • Data collection • Causal factor determination • How to fill gaps in data • Root cause identification • Writing recommendations • Using the incident database • Programmatic issues such as reporting, communication, legal issues Training Agenda • Data collection • Causal factor determination • Root cause identification • Writing recommendations • Using the incident database Training Agenda • Near-miss definitions and examples • The learning value of incidents • No blame approach • Root causes are management system failures • Incident reporting system Training Agenda • What is changing in how you approach incidents? • What can each person do to help the system work? • Expected impact to most jobs

2 PROCESS SAFETY AND MANAGEMENT OF ABNORMAL SITUATIONS 2.1 IMPACT ON PROCESS SAFETY Process safety requires a discip lined framework for managing the integrity of operating systems and processes handling hazardous substances by applying good design principles, engineering, and operating practices. Therefore, pr ocess safety management systems focus on the prevention of , preparedness for, miti gation of, response to, and restoration from catastrophic rele ases of chemicals or energy from process facilities. At some companies, these management systems have been in place for many years and ar e generally credited with reducing major accident risk in the process industries. The risks can range from incidents involving toxic and flammabl e material releases resulting in toxic effects, fires, or explosions with potential impacts of harm to people (injuries, fatalities), to in cidents that can cause harm to the environment, property damage, or production losses, and provide a conduit for adverse business publicity. As discussed in Chapter 1, abnormal situations can be characterized as unplanned transitions or events that occur during normal and transient operations. For example, on e or more process upsets that the automated control system (typically a distributed control system (DCS)) cannot correct can be the basis of an abnormal situation. This event may require intervention by operator(s) to augment the control system, to avoid a production/quality impact, or potentially, a serious process safety incident. Figure 2.1 illustrate s the relationship between abnormal situations and the process safety in cident categories in the CCPS and American National Standard Instit ute (ANSI)/American Petroleum (API) process safety metrics documents (CCPS 2018e; ANSI/API 2016).

322 INVESTIGATING PROCESS SAFETY INCIDENTS frequency and potential consequences. Accurate risk assessment is important to conducting a meaningful cost-benefit analysis. Layer of protection analysis (LOPA) (CCPS, 2001) is one tool that may be useful in this evaluation. Another challenge to effective re commendation resolution occurs when the action intended by the in vestigation team is not clearly or completely stated. This avoidable mist ake can lead to misunderstandings on the part of management decision-make rs. A common example of obscure wording is ineffective use of the te rms “consider” or “review.” If the investigation team believes that a particular system defect exists and should be corrected, then the team should st ate this finding very clearly and recommend a specific measurable task (e.g. Task (1) “Study… ”, and Task (2) “Implement the findings of Task (1)… ”. Any attempt to designate a recommendation as implemented and thus designated as “Closed” upon reachi ng an intermediate or temporary milestone should be discouraged. Typically such atte mpts stem from poorly worded recommendations and are based upon the promise of future actions. For example, “issue a project request for… ” with a status of “Project Request Approved” is not verification th at the recommendation has been “Completed”. Similarly a recommendatio n to “consider adding … .” with a status of “consideration completed”, pr ovides no explicit documentation of the remedial corrective ac tions taken, if any. In other cases where the investigation team does not believe the recommended action is mandatory, this distinction sh ould also be clearly stated. An example would be the recommen dation of a best practice activity, which could be rejected by management without major consequence. This is one of the reasons why it is useful for each recommendation statement to include comments on the consequences to be averted and the benefits associated with implementing the re commendation. Effectively written recommendations include phrases such as “in order to prevent x, implement y.” It should be noted, however, that some companies have recommendation language protocols in place that may differ from this advice. If a recommendation proposes a change in the process, the change should be managed through the MOC procedure and the associated actions should include a safety assessment which, dependi ng on the change, may include a formal Process Hazard Analysis (PHA) study, such as a HAZOP or other methodology, before implementation. A systematic and formal Hazard Analysis approach identifies and evaluates hazards associated with the

280 Figure 11.2 Valve Gearbox Designs. Ref 11.18 CSB, 2016 11.6 SAFE WORK PRACTICES The reduction of ignition sources, which is partially a design issue and partially a safe work practice i ssue is an example of the use of

DETERM INING ROOT CAUSES 257 Another situation where checklists can be very helpful is when the investigation team has no hypothesis as to what caused an occurrence. The checklist is an example of an inductive approach that can be used to get past a mental block. Checklists used for process safety incident investigation share many similarities with predefined trees. They can comprise a series of questions or statements related to root causes based upon experience of safety management systems. Some checklis ts need care in use because the statements that they contain can infer blame to the casual observer, rather than discourage blame-seeking. Checklists also offer consistency and repeatability by presenting different investigators with the same standard set of potential root causes for each incident. This consistency facilitates statistical trend analysis of multiple incidents involving recurring problems within an organization. While a checklist may not encourage the investigation team to think laterally of other potential causes, it ca n overcome a lack of experience within the team and present causes that the team would not have other wise considered. 10.9.1 Use of Checklists The use of checklists as a primary root ca use analysis tool is virtually identical to the use of predefined trees. This is hardly surprising as most predefined trees are really a succession of checklists organized by subject matter (category) into an arrangement of branches within the tree. A timeline or sequence diagram is first developed, and then causal factors identified. Care should be take n to ensure that the checklist is not used too early. Be sure to determine what happened and how it happened before determining why it happened. Otherwise, the team will think that they have identified the right root cause( s), when in reality only one or two of several multiple causes have been determined. The causal factors are then applied one at a time to each page of th e checklist(s) to iden tify relevant root causes. Those pages th at are not relevant to the particular incident of interest are discarded. Similar quality assurance checks should be applied as those described for predefined trees. The use of checklists to supplem ent another root cause analysis method can be an effective technique; for example, human factors checklist(s) may be used in conjunction with logic trees. The checklist may be used as a guide during development of a logic tree, or as a check after the tree has been

257 Table 12.5 continued Process Phase Example Objectives Hazard Analysis Technique Routine operation Identify employee hazards associated with the operating procedures. Identify ways an overpressure transient might occur. Update previous hazard evaluation to account for operational experience. Identify hazards associated with out-of- service equipment. Checklist What-If What-If / Checklist Hazard and Operability Study Process modification or plant expansion Identify whether changing the feedstock composition will create any new hazards or worsen any existing ones. Identify hazards as sociated with new equipment. 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 Decommissioning Identify how demolition work might affect adjacent units. Identify any fire, explosion, or toxic hazards associated with the residues left in the unit after shutdown. Safety Review Checklist What-If What-If / Checklist Performing a Good Quality Process Hazard Analysis A good process hazard analysis will result in a comprehensive listing of potential incidents with causes and recommendations for managing the hazards. Additionally, the PHA can identify high hazard scenarios that warrant more detaile d consequence and risk analysis (see chapters 13 and 14). Just as with any analysis, an analyst can do a good job or a poor job on a process hazard analysis. In order to conduct a quality PHA, the analyst should plan the preparation, the analysis, and the follow-up. The preparation for an analysis should include gathering the appropriate data and the appropriate expertise. The data can include that listed in Section 12.3.4 and more. The expertise should include people familiar wi th the process, the operation, the local environment, and others as needed. Preparation will also include selecting a location that will encourage the analysis participants to stay engaged in the analysis as opposed to being distracted by their other duties. The type of anal ysis should be identified in advance and will influence the data and expertise needed. The scop e of the analysis should be determined in advance. This could be a single piece of equipm ent, an entire process unit, or a modification and the associated tie-in points in an existing facility. HAZARD IDENTIFICATION

4 EDUCATION FOR MANAGING ABNORMAL SITUATIONS The purpose of this chapter is to provide guidance on training for the management of abnormal situations. It is not possible to provide specific training for abnormal situations that will be appropriate for all operating facilities. The content in this chapter is directed not only at site personnel, but also at those who may be remote from day-to-day operations and may have an influence on the occurren ce, development, and control of abnormal situations. 4.1 EDUCATING THE TRAINER The content includes advice , tools, and techniques fo r the provision of such training, both in the classroom, in local workgroups (such as “tool-box talks”), via E-learning, and using DCS simulators or “digital twins”. As discussed in Chapter 1, simple compu ter-based training material is offered in conjunction with this b ook, consisting of five separate scenarios of abnormal situations that develop on part of an operating process involving distillation, pumps, ta nks, heat exchangers and various instrumentation/controls. These training modules can be used by supervisors, plant engineers, and trainers to train operating teams in the diagnosis of an abnormal situation. The material presents background information, develops the scenarios, and provides prompts for trainers and trainees to evaluate what is ha ppening and why. This should enable discussions on abnormal situations and their diagnosis, actions to be taken, learning, and relevance to their operation. This type of material could be developed and extended by manageme nt/training personnel to include other situations such as startups, shutdowns and other non-steady state operations specific to their operations. Details on how to access this material is in Appendix A. The example incidents (such as Exam ple Incident 4.1) also provide useful examples to the various groups that can influence the successful management of abnormal situations.

Piping and Instrumentation Diagram Development 162 There are some fire‐based heating systems that are very common and are applicable in smaller size containers. For example fire‐tube equipped containers are mainly applicable for vessels and not tanks. Submerged combus - tion is another method, but it is not very common. Steam injection is a method that is sometimes used.Heat tracing, heat blankets, and steam jackets are generally considered as winterization systems and not heating systems. However, they can also be used for heating small vessels. Winterization methods will be discussed in Chapter 17. 9.14 Mixing in Containers Sometimes it is necessary to carry out mixing on streams after merging them together. If mixing can be done in a pipe only a static mixer is enough. However, the other type of mixing can happen in containers. There are two main types of mixing in containers: the hydraulic type and the mechanical type. Different types of mixing in containers are shown in Figure 9.30. The mechanical type of mixing is basically using a mechanical mixer for the purpose of mixing. Mixers are a type of impeller that is connected to an electric motor. The types of mixer impellers are different and depend of the type of mixing, and are beyond the scope of this book. Electric motors could be connected to impellers directly or through gear boxes. However, we don’t see this in the P&ID unless the gear box is huge and needs some sort of lubrication system. Mechanical mixers can be installed in two different positions in containers: on the top and center of the container and at the side of the container. Installing the mixer on the top of the container is a better option; however, it is not always applicable unless the diameter of the container is small (say <5 m in diameter). If the diameter of container is large as in tanks (say >5 m diameter) using a top mounted mechanical mixer is not applicable and a side mounted mixer or mix - ers should be used. If the mechanical mixer is top mounted the best posi- tion is at the center of the container. Installing the mechan-ical mixer at the center of the container minimizes stress and vibration on the container. However, to make sure that mixing happens, rather than rotation of the whole bulk of fluid inside of the container, some baffles should be installed on the container. If, for whatever reason (e.g. risk of plugging because of liquid dirtiness), installing baffles is not applicable we may choose to install a draft tube or off‐center top mounted mechanical mixer. The other option is installing a top mounted, tilted mixer. However, the top mounted center mixers are the best choice. For large tanks side mounted mixers are more common. Even though operators always have concerns about leak - age from side mounted mixer nozzles, sometimes using them is inevitable. Side mounted mixers could be installed on the shell side of tanks from one to three or more units. The second type of mixing in containers is hydraulic mixing. Hydraulic mixing is not a very efficient type of mixing in comparison to mechanical mixing. If, for whatever reason, mechanical mixing cannot be done, for example, when the fluid is very aggressive against the mixer impellor, hydraulic mixing can be done. Hydraulic mixing can be in the form of recirculation or direct injection. Recirculation means sending back a portion of liquid from the discharge side of the pump back to the tank, but in direct injection just the incoming fluid goes through the nozzle and through the sparger in the tank. The recirculation pipe is connected to a nozzle on the tank and the nozzle internally is connected to a piece of pipe in simple form or in the conductor type. 9.15 Container Internals The necessity for and type of container internals are specified during the design of equipment. However it should be mentioned here that the tank internals are never a welcomed item! We put them if we have to. Internals in containers are very difficult for inspection, and if they are malfunctioning it is not easy to spot them. We do not always show the container internals on P&IDs, but if we choose to show them (because of their importance) we show them in the form of dashed lines. 9.16 Tank Roofs There are several types of tank roofs but the one that can be recognized in P&IDs are dome roofs, fixed cone roofs, fixed cone-internal floating roofs, and external floating roofs. Fixed roof tanks are for non‐volatile liquids. Recirculation Sparging Top mounted Side mount edHydraulic mixing Mechanical mixing Figure 9.30 Differ ent types of mixing in containers.

186 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS 7.2.3 Outline Process Description of Milford Haven Refinery The description has been taken from the more detailed HSE report (HSE UK 1997), which should be referred to if greater understanding is required. However, this extracted in formation should provide sufficient detail for the reader to understand the learning associated with the management of abnormal situations. The Milford Haven refinery in South Wales includes a crude distillation unit (CDU) that processes 190,000 barrel s per day and incorporates several other process units including a vacuum distillation unit (VDU), FCCU, butane isomerization unit and HF alkylation unit. The CDU separates crude oil by fractional distillation into intermedia te products including naphtha, gas, kerosene, diesel, and heavier components. The heavy fractions supply the VDU that in turn feeds the FCCU, which cracks the long chain hydrocarbons into lighter fuel products including light naphtha, butanes, propanes, ethane, methane, etc. The heavier material from the FCCU includes gas oil and fuel oil. The key process area relevant to the incident is the separation section, as seen in Figure 7.4 (HSE UK 1997) th at is fed with hydrocarbons from the top of the main fractionation column, downstream of the FCCU. The overheads stream cools and passes through the primary and then the secondary overhead accumulator, F-203. Liquids collect in the accumulator and the remaining vapors are compressed in the wet gas compressor and then cooled before entering th e high-pressure separator, F-310. Liquid from the high-pressure separa tor is pumped to the Deethanizer column, F-302, which removes mainly C2’s (ethane) at the top and the bottom product supplies the Debutani zer column F-304, which removes C3’s and C4’s at the top, which comprise the LPG product. The bottom of the Debutanizer is fed to the Naphtha splitter F-305, which produced Light Cycle Naphtha (LCN) that is used to blend into gasoline product. The flare system collects relief streams from various parts of the process and feed them to the Flare knockout drum F-319. This is designed to remove any liquids from the hydr ocarbon stream, allowing gases only forward to the flare stack.

304 | Appendix E Process Safety Culture Case Histories Over tim e the Coordinator noticed that the PSM S elem ents had become rigid and that the Manager resisted any improvement ideas, regardless of the source. The Manager even rejected im provement suggestions from the corporate process safety team . It became clear that the Manager’s inflexibility was simply protecting his turf. When is being rigid about m aintaining consistent practices good, and when it is bad? Provide Strong Leadership, Empower Individuals to Successfully Fulfill their Safety Responsibilities, Defer to Expertise. E.18 PHA Silos A large facility performed com plete PHAs that com prehensively identified and identified controls for process safety risks. These studies were carefully revalidated over the years to keep them up-to-date. The recomm endations were resolved prom ptly and there were good records of these practices. However, a closer look at PHA practices revealed that although recomm endation management was excellent, the thoroughly perform ed PHAs were not used for any other purpose. The AI/M I team did not receive the report so they could ensure that critical equipment identified in the PHA was included in the MI ITPM and QA programs. The training team was unaware of why recomm ended training was needed. The emergency response planning team was unaware of the potential consequences they needed to plan for. The PHA program, as good as it was, had become a silo activity. How can this happen in a large facility? How can cross- fertilization be encouraged when the PHA team is in a com pletely different organizational structure and where so m any people are involved? Actual Case History

YYWJJJ


INVESTIGATING PROCESS SAFETY INCIDENTS The third edition was authored by Baker Engineering and Risk Consultants, Inc. The authors at BakerRisk were: Quentin A. Baker Michael P. Broadribb Cheryl A. Grounds Thomas V. Rodante Roger C. Stokes Dan Sliva was the CCPS staff liaison and was responsible for overall administration of the project. CCPS also gratefully acknowl edges the comments and suggestions received from the following peer reviewers: Amy Breathat, NOVA Chemicals Corporation Steven D. Emerson, Emerson Analysis Patrick Fortune, Suncor Energy Walter L. Frank, Frank Risk Solutions, Inc. Barry Guillory, Louisiana State University Jerry L. Jones, CFEISBC Global Gerald A. King, Armstrong Teasdale LLP Susan M. Lee, Andeavor William (Bill) D. Mosier, Syngenta Crop Protection, LLC Mike Munsil, PSRG Pamela Nelson, Solvay Group Katherine Pearson, BP Americas S. Gill Sigmon, AdvanSix Their insights, comments, and su ggestions helped ensure a balanced perspective to this Guideline. The efforts of the document edito r at BakerRisk are gratefully acknowledged for contributions in editing, layout, and assembly of the book. The document editor was Phyllis Whiteaker. The members of the CCPS Inciden t Investigation Subcommittee wish to thank their employers for allowing them to participate in this project and lastly, we wish to thank Anil Gokhale of the CCPS staff for his support and guidance.

278 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION which can significantly alter the heat transfer be havior. For liquids having normal boiling points near or above ambient temperature, diffusional or mass transfer evaporation is the limiting mechanism. The vaporization rates for this situation are not as high as for flashing liquids or boiling pools. A simplified approach for smaller releases of liquids with normal boiling points well below ambient temperature is to assume all the liquid enters the vapor cloud, either by immediate flash plus entrainment of aerosol, or by rapid evaporation of any rainout. Table 13.3. Input and output for evaporation models Input for boiling liquid pools Input for nonboiling liquid pools leak rate pool area (for spills onto land) wind velocity ambient temperature pool temperature ground density specific heat thermal conductivity solar radiation input parameters leak rate pool area (for spills onto land) wind velocity ambient temperature pool temperature saturation vapor pressure mass transfer coefficient solar thermal input Output: The time-dependent mass rate of boiling or vaporization. Pool Spread Models. An important parameter in all of the evaporation models is the area of the pool. Pool spreading models are based primarily on the opposing forces of gravity and flow resistance and typically assume a smooth, horizontal surface. If the liquid is contained within a diked or other physically bounded area, then the area of the pool is determined from these physical bounds if the spill has a large enou gh volume to fill the area. If the pool is unbounded, then the pool can be expected to sp read out and grow in area as a function of time. The size of the pool and its spread is highly dependent on the level and roughness of the terrain surface -most models assume a level an d smooth surface. One approach is to assume a constant liquid thickness throughout the pool. Th e pool area is then determined directly from the total volume of material. This approach pr oduces a conservative result, assuming the spill is on a flat surface, the pool growth is not co nstrained, and the pool growth will be radial and uniform from the point of the spill. More comp lex models include consideration of gravity spread and flow resistance terms for both la minar and turbulent flow but does not include evaporation or boiling effects. The approach is si gnificantly different if the pool is on water versus land. A pool spread model solves the simultaneous, time-dependent, heat, mass, and momentum balances. Factors important to this pool spread modeling include the following.

(VJEFMJOFT
GPS
*OWFTUJHBUJOH
1SPDFTT
4BGFUZ
*ODJEFOUT
5IJSE
&EJUJPO By 5IF
"NFSJDBO
*OTUJUVUF
PG
$IFNJDBM
&OHJOFFST Copyright
¥

5IF
"NFSJDBO
*OTUJUVUF
PG
$IFNJDBM
&OHJOFFST 314 14 IM PLEM ENTING RECOM M ENDATIONS The ultimate goal of incident inve stigation is preventing recurrence of a specific incident scenario or related similar incidents. Considerable time and resources are expended in determinin g an incident’s ca usal factors and associated root causes (incident fi ndings) and identify ing recommendations (preventive actions to remedy deficie ncies or mitigate consequences). Despite this effort, the potential for a similar occurrence at other facilities remains unchanged until the incident details are communicated, findings are evaluated across other sites, and reme dial recommendations evaluated and implemented. The value of the investigation is entirely dependent on the effectiveness of follow-up activi ties. This chapter focuses on implementation and co mmunication of the team’s conclusions . The investigation team’s charter is typically complete when the recommendations have been submitted in the final incident report; however, the company’s responsibilities are far from over. Management needs to approve, or in some cases formulate recommendations. Other portions of the organization may be responsible for assessing the applicability of specific recommendations to remedy similar situations at their operations, while others may be assigned responsibility for evaluation, implementation and follow-up of findings identified by the investigation team. Implementation of reco mmendations is a good and necessary business practice for a variety of reasons, most notably the desire to prevent repeat or similar events. In addition, recommend ation implementation often leads to strengthened management systems th at improve operations across the board (safety, productivity, quality, etc.) and positively impact employee morale. There has also been an increased emphasis wi thin organizations to identify, investigate, pub licize, and take action on near-miss occurrences. This chapter addresses: • Major activities related to implementing recommendations, • Examples of repeat incidents wh ere previous incident findings were not validated and/or fo llowed-up adequately, and • Practical suggestions for achiev ing successful implementation.

220 | 6 Where do you Start? group should include a diversity of functions, to help the m oderator understand how various parts of the organization interact with each other. Sm aller groups may not have that diversity, although mini-focus groups of 3 to 4 participants can be useful for evaluating individual functions. B ased on the goals of the focus groups and the structure of the site, the total num ber focus groups and the actual participants of each group should be identified. Schedule and location of each focus group can then be determined. The questions for each group then should be selected. Focus groups should limit the num ber of questions to 5-6, to allow sufficient tim e to discuss each question. More than one group m ay, and in many cases, should be asked the same questions. If m ore questions need to be discussed, additional groups can be formed. Moderators should not feel pressured to get answers to all questions. Rushing at the end to answer remaining questions will not allow the depth of discussion needed. If a group fails to discuss all questions, the group should be reconvened later, or the questions asked of another group. The focus group location should, like individual interview room s, be familiar and comfortable to the interviewees. Ideally, the meeting room should be arranged with chairs in a circle with no table or desks in the way. This allows participants to face each other with no barriers between them . Disruptions such as radios and cell phones should be turned off or, better, not brought into the room. In planning focus groups, realize that people who have already attended focus group sessions will talk about them. Consider encouraging them to do so. This can help jumpstart subsequent focus group discussions. Also anticipate that som e attendees may linger afterwards to say things they were not comfortable sharing with the group. For the same reason, moderators should leave

Appendices 189 Q T R III. Location of the Motor Control Center Is the motor control center (MCC) located so that it is easily accessible to operators? Are circuit breakers easy to identify? Can operators/electricians safely open circuit breakers? Have they been trained? Is the MCC designed such that it could not be an ignition source? Are the doors always closed? Is a no-s moking policy strictly enforced? Is the MCC being improperly used for shelter? IV. Location and Construction of O ccupied Structures and Control Rooms Is the occupied structure built to satisfy current company overpressure and safe-haven standards? Does the construction basis satisfy acceptable criteria (e.g., the Factory Mutual recommendations)? Are workers in occupied structures (or escape routes from them) protected from the following: • Toxic, corrosive, or flammable sprays, fumes, mists, or vapors? • Thermal radiation from fires (including flares)? • Overpressure and projectiles from explosions? • Contamination from spills or runoff? • Noise? • Contamination of utilities (e.g., breathing air)? • Transport of hazardous materials from other sites? • Possibility of long-term exposure to low concentrations of process material (e.g., benzene)? • Impacts (e.g., from a forklift)? • Flooding (e.g., ruptured storage tank)? Are vessels containing HHCs located sufficiently far from occupied structures? Were the following characteristics considered when the occupied structure’s location was determined: • Types of construction of the room? • Types/quantities of materials? • Direction and velocity of prevailing winds? • Types of reactions and processes? • Operating pressures and temperatures? • Fire protection? • Drainage? If windows are installed, has proper consideration been given to glazing hazards?

370 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION on how the human works (both physically and me ntally), and how the system can be designed to support their success, can also support successful process safety performance. An ethical aspect of engineering is to work wi thin your area of expertise and to call for the assistance of others when outside of that area of expertise. Human factors may not be an area of expertise for the typical engineer. Reach out to human factors experts, outside of engineering, that can assist and build competency in the area. Other Incidents This chapter began with a description of the Formosa Plastics explosion. Other incidents relevant to human factors include the following. NASA Challenger Disaster, Florida, U.S., 1986 Celanese Pampa Explosion, Texas, U.S., 1987 Piper Alpha Platform, North Sea, U.K., 1988 Texaco Oil Refinery Explosion and Fire, U.K., 1994 Mars Climate Orbiter, U.S., 1999 NASA Columbia Loss, Texas, U.S., 2003 Buncefield Storage Tank Overfl ow and Explosion, U.K., 2005 BP Isomerization Unit Explosion, Texas City, Texas, U.S., 2005 Air France Flight 447, Brazil, 2009 U.S. Airways 1549, U.S., 2009 Deepwater Horizon Well Blowout, Gulf of Mexico, U.S., 2010 Chevron Richmond Refinery Fire, California, U.S., 2012 MGPI Processing Plant, Kansas, U.S., 2016 Exercises List 3 RBPS elements evident in the Formos a Plastics VCM explosion summarized at the beginning of this chapter. Describe their shortcomings as related to this accident. Considering the Formosa Plastics VCM, what actions could have been taken to reduce the risk of this incident? Training, incentives, and supervision can prevent humans making mistakes. True or false? Explain your answer. A refinery explosion investigation found that in the 11 minutes before the explosion, 275 alarms requiring operator action activa ted. This illustrates which human factors topic? Explain your answer. The BP Refinery explosion in Texas City is described in Chapter 2. What human factors topics were illustrated in this incident? Explain your answer. The Deepwater Horizon well blowout is described in Chapter 21. What human factors topics were illustrated in this incident? Explain your answer. A complicated catalyst regeneration proc edure takes place over three days. This process is done only once every few years. What human factors topics could impact this procedure? Explain your answer.

34 Human Factors Handbook 3.5 Key learning points from this Chapter Supporting successful human performance an d reducing the possibility of error and mistakes can be achieved by understanding the nature of the tasks, the type of human performance required and the ca uses of potential error and mistakes, and then designing the work environmen t taking these points into account. It is vital to create the conditions for successful task performance. In addition, key learning points include: • Each type of human performance is re lated to different types of error and mistakes. • Each type of human performance may benefit from a different form of support.

Piping and Instrumentation Diagram Development 100 accordingly. If for whatever reason there is financial con- straint for designing and also buying different pipe sup-ports, eccentric reducer or enlarger with FOB can be used instead (Figure 6.82). Another reason that may dictate using eccentric reducer or enlarger is specific for liquid transferring pipes. An example is facilitating the draining operation of a pipe circuit. For liquid transferring pipes, the reducer or enlarger is a conventional concentric is used and full draining is impossible (Figure 6.83). A reducer or enlarger that is installed FOB can solve the problem. An example of this application is for the reducers of control valves (Figure 6.84). The full draining is not always important. If the pipe is a small bore (possibly less than 4″ ) or the liquid is inno- cent, it is possible that there is no need to provide full draining capability such as also in pumping a liquid with dissolved gases as it was discussed previously (Figure 6.85). The discharge of the centrifugal pump possibly does not need the eccentric FOT enlarger because it is most likely a vertical pipe.6.14.1.3 Three‐Way Connections In Section 6.9, a multipoint source and multipoint desti-nation pipe arrangements are discussed. When there is a pipe arrangement with multiple sources or multiple destinations, the concept of tying‐in and branching off comes to the picture. Both of these concepts boil down to the question: How is a pipe con-nected to the middle of another pipe? Generally the response would be a requirement for a three‐way con-nection. But there are several types of three‐way connec - tions. To know which type of three‐way connection should be used, a P&ID developer needs to consult the branch table available in the piping material spec table of the project. There are different fittings available to do that including tee’s, reduced‐tee’s, and different types of O‐lets (e.g. Weldolets, Threadolet). However, Table 6.8 can be used in deciding on the type of three‐way connection to be used, and it also shows the  representation of different three‐way connections on P&IDs. As can be seen from the table, the type of intersection from P&ID cannot be figured out. The three‐way con-nection could be a tee, reduced tee, or O‐let. There is no specific symbol to recognize a tee from a reduced tee or from an O‐let. Therefore, a P&ID will not show whether a three‐way connection is a tee, reduced tee, or O‐let. If someone wants to see if the intersection is tee, reduced tee, or O‐let, the piping spec must be consulted. 6.14.1.4 Pipe Connections The pipe manufacturers do not fabricate pipes in infinite lengths. During the construction, the pipes should be connected to each other, end to end, to make a suitable length of a pipe route. There are different ways of con-necting pipes and are shown in Table 6.9. The only type of pipe connection that is visible on the P&ID is the flange. However, flanges are not shown unless there is specific need for them. 6.14.1.5 End‐of‐Pipe S ystems End‐of‐pipe systems are applied to uncoupled pipes. Uncoupled pipes are the pipes that are not connected to other pipes and are not extended. There are several available options for the end‐of‐pipe provision listed in Table 6.10. For pipe sizes less than 2″, a screwed cap or plug can be used. The tendency is toward using a screwed cap in the newer plants. For pipe sizes more than 2″, a blind flange or welded cap can be used. If there is a plan to extend the pipe in the future, a blind flange is generally used, but a welded cap is enough. If a frequent connection to other hose is necessary, a quick connection is the best. Off‐loading systems are using quick connections because the plan is frequent transferring of fluid to and from a transportation system. Figure 6.82 Need for ec centric reducer to use identical pipe supports. Figure 6.83 Using FOB ec centric reducer or enlarger to satisfy full draining of liquid pipes. Figure 6.84 Using FOB ec centric reducer for control valve. Figure 6.85 Using FO T eccentric reducer in suction of centrifugal pump.

210 INVESTIGATING PROCESS SAFETY INCIDENTS management system deficiency has not been reached, the team has stopped too soon at a symptom, immediate failure, or an event that contributed to the incident. Sometimes, the investigation team may not be able to proceed due to a lack of knowledge or informat ion, in which case further evidence gathering and analysis may be required. Most incidents do not have a single root cause. In order to identify multiple root causes, the technique sh ould be repeated asking a different sequence of questions each time. For ex ample, the investigation team should examine other possible reasons fo r the original causal factor before starting with a different negative event or unde sirable condition that influenced the course of activity leading up to the incident. The 5 Whys technique may be used indi vidually or to assist development of a fishbone diagram (also known as Ishikawa or Cause & Effect diagram). A fishbone diagram is used to examine potential causes of an incident or equipment failure, and the 5 Whys may be used to uncover the root causes. The incident is shown as the fish's head with the causes extending to the left as fishbones. Causes are usually groupe d into major categories (e.g., people, process equipment, etc.), and branch off the backbone as ribs with sub- branches for root causes. Figure 10.2 illustrates a typical fishbone diagram. Figure 10.2 Example of Is hikawa Fishbone Diagram While companies in differ ent industries have su ccessfully used 5 Whys, the technique has some inherent limit ations. Table 10.1 illustrates some of its strengths and weaknesses. The fish bone diagram has many of the same

Table C-1 continued HF Competency Performance/ Knowledge Criteria Level 1 - Operator Level 2 - Supervisor* Level 3 - Manager** Non-technical skills Communication Understands the concept of effective communication and related topics (e.g., barriers and enablers to effective communication in emergency situations) Can describe what is meant by effective communication Can identify barriers and enablers to effective communication Can recognize when communication is being impaired Can recognize barriers to communication in work context Is able to lead discussion on effective communication in normal and abnormal situations, including barriers and enablers to communication Is able to communicate effectively in emergency situations Can apply effective communication techniques. Can apply and train people in techniques to improve communication Is able to review and develop techniques to enhance communication in various operating conditions, including emergency situations

APPLICATION OF PROCESS SAFETY TO ONSHORE PRODUCTION 93 Great care is attached to protecting the environment and wildlife in Alaska as is seen in this common scene of a caribou gr azing beside a pipeline in a North Slope production facility (Figure 5-3). Key Process Safety Measure(s) For leaks, all the elements of RBPS, SEMS and PSM are important, but two are especially relevant and are listed below. Hazard Identification and Risk Analysis: Through the use of HIRA tools, a better understanding is obtained regarding the prec ise nature of the hazards presented by leaks from production equipment. This helps facility designers identify the proper safeguards to prevent or mitigate leak events. These include good operating procedures and safe work practices, inspection and maintenance, flammable and toxic gas detection, emergency shutdown systems, consequence calculations to estimate potential hazard distances, and area classification to reduce ignition likelihood. Emergency Management: Control rooms on larger facilities not only protect personnel but serve important roles in emergency response. Modern practice is to either separate the occupied building from the hazards or to make the control room resistant to the types of consequences iden tified in the HIRA. Guidance is provided in API 752 and 753, with additional de tails in CCPS (2018b). Fire protection measures are summarized in CCPS (2003) with additional guidance in NFPA and API Standards. Figure 5-3. Caribou grazing by pipeline in production facility

58 INVESTIGATING PROCESS SAFETY INCIDENTS 4.2 TYPICAL M ANAGEM ENT SYSTEM TOPICS As stated in the introduction, the inci dent investigation management system is a written document that defines the roles, responsibilities, protocols, and specific activities to be carried out by personnel performing an incident investigation. The management system may incl ude a purpose statement, definitions, incident classifications, and investigation responsibilities. It provides the structure for activities such as evidence gathering, witness interviewing, and data control as well as standard practices for notification, reporting, and follow up. The fo llowing sections summarize the recommended elements of a mana gement system for incident investigation. 4.2.1 Classifying Incidents When developing an incident investigation managemen t system, it is important to defin e common terms and classifications (ASSE Dictionary, 1988). Several incident categories can be used to develop a classification system. Classification has three main purposes: 1. Determining the significance of the incident an d the resulting consequences. This often dictates team leadership, size, composition, and investigative techniques. 2. Determining how investigation results will be communicated and to whom (including regulatory-r equired communications). 3. Provide consistent data for tr ending and other analytics. The system should describe specific mechanisms for deciding to activate an investigation team and the team composition for each incident classification. There should also be a mechanism that des cribes required internal and external notification. This is usually captured in a procedure and associated routing forms. The inci dent investigation management system should specify: • Who will make the notification • Who is to be notified • How and when they are to be notified Chapter 5 provides descriptions, incident classifications, and examples of functions and organizations that might need to be notified.

Fundamentals of Instrumentation and Control 257 Temperature sensor arrangements: when the pipe size is small it is not easy to place a thermocouple in it. “Small” is defined differently in different companies. Some companies consider small to be sizes smaller than 4″ but some other companies consider 3″ or even 6″. There are two solutions when faced with the issue of placing a thermocouple in small size pipes. One is using a combination of enlarger–reducer to enlarge the pipe size and then placing the thermocouple (Figure 13.20). The other solution is placing the thermocouple on a bend of the pipe (Figure 13.21). If the purpose of temperature measurement is skin temperature, it could be mentioned beside the balloon on the P&ID, or showing the connection to the surface. Figure 13.22 shows a P&ID with both of them together. 13.11.1.2 Pr essure Measurement Pressure measurement can be done everywhere on the flow of gases or liquids. However, measuring the pressure of liquids is not as common as gases and vapors.Temperature sensors generally don’t have any specific symbol for P&IDs. Table 13.17 is a non‐exhaustive list of common pres - sure sensors. Pressure sensor arrangements: some companies shows PGs and PIs simply as in Figure 13.23. However, this doesn’t necessary mean that there is nothing between the gauge and the process. There are generally different items on the connecting tube. As a minimum there should be one isolating valve – a root valve – and a drain/calibration valve (Figure 13.24). In some critical cases, instead of one root valve, two root valves and a drain can be placed (Figure 13.25). In the cases where the process fluid is harmful to the internal parts of the sensor, a diaphragm can be placed and the space between the diaphragm and the sensor TE 2/uni2033 2/uni2033 Figure 13.20 Ther mocouple installation on narrow pipes (option 1). TT 123 TE 123SKIN Figure 13.22 Skin t emperature sensor.2/uni2033 2/uni2033TENote 1Note 1:On pipe elbow Figure 13.21 Ther mocouple installation on narrow pipes (option 2). Table 13.17 Pr essure sensors. Type Unique advantage Unique disadvantage Application ●Bellows type ●Diaphragm type ●Bourdon tube typeSimple system More prone to mechanical failureDefault choice Piezoelectric No moving parts Limited range General process pressure measurementPGPI 1056 Figure 13.23 Pr essure gauge and pressure indicator connected to process. PG Figure 13.24 Pr essure gauge with root valve and drain/calibration valve.

Chapter No.: 1 Title Name: Toghraei c07.indd Comp. by: PVijaya Date: 25 Feb 2019 Time: 12:21:35 PM Stage: Proof WorkFlow: <WORKFLOW> Page Number: 105 105 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 Valves are a type of pipe appurtenances. Other pipe appurtenances, fittings, and specialty items were already discussed in Chapter 6. 7.1 Valve Naming Valves are named based on their action on the service fluid (e.g. throttling valve, stopping (blocking) valve, or diverting valve) or based on their operating mechanism (e.g. motor‐operated, solenoid, or manual valve). Valves can also be named based on their plug type; they can be a gate, globe, or butterfly valve. Valves can also be named based on the valve’s location in the piping (e.g. foot or root valve) or on its duty in the process (e.g. shutdown, blowdown, or flow control valve). 7.2 Valve Functions Valves are piping appurtenances that actively affect flows. Active means the valve has a movable part. For example, a gate valve has a moving stem, and a check valve has a moving flap. Any valve with moving mechanism that does not take order from the outside of the valve can be categorized as special valve. Check valves, air release valves, and excess flow valves are examples of special valves. Special valves are a large group of valves that have specific function. A few subgroups of special valves will be discussed at the end of this chapter. Special valves are diverse, which makes it difficult to classify them, but other valves can be easily categorized. 7.3 Valve Structure Valves have two main parts: the operator and the plug (see Figure 7.1). The valve operator is the part that takes orders from an external source to act on the fluid flow. The external orders could be an operator’s hand or a nonhuman actuator. The valve plug is the part of the valve whose internal portion is in contact with the fluid. A valve plug can be built such that the valve is designed either as a throttling valve, an isolation valve, or a diverting valve. The valve operator is connected to the valve plug by a piece of rod called the stem. The stem can move up and down or can turn depending on the type of the valve. It is important not to get confused with plug : a plug is a part in all valves and a plug valve is a specific type of valve too. 7.4 Classification of Valves There are at least three ways to classify valves: based on the action of the service fluid and their functions, based on the number of ports, and based on the number of seats. The function of non‐special valves in processing industries could be any of the following: 1) Adju sting flow valves 2) St opping flow valves Valves that adjust flow are valves that change the magnitude of flow in the pipes. These valves are called throttling valves. In these valves the valve plug is respon­sible for changing the magnitude of flow. In theory the valve plugs can take infinite position between fully open and fully closed, and hence valves can be categorized as either partially closed or partially open valves. The movement of the stem can change the plug position manually by an operator or via a remotely operated oper ­ ator. These so‐called valve operators will be discussed later in this chapter. On the other hand valves that stop flow are valves whose plug can only remain in fully closed or fully open position. In both positions the valve allows a stream to flow. Therefore these valves are called blocking valves.7 Manual Valves and Automatic Valves

Direct chlorination using 95% chlo rine vapor direct from chlorine cells rather than purified chlori ne eliminates the need for liquid chlorine storage (F igure 15.14) (OK). Figure 15.14: Use of Chlorine Vapor for Liquid Chlorine Dow’s licensed EO METEORTM tec hnology significantly reduces the portion of the process using concentrated EO (Figure 15.15). 15.8.2 Substitute Replace a material with a less hazardous substance. Manufacture of acrylic esters by oxidation of propylene to produce acrylic acid, followed by esterification to manufacture the various esters, is inherent ly safer than the older Reppe 425

Khan), abbreviated here as KA. Overton and King in their article, “Inherently Safer Technology: An Evolutionary Thing,” describe inherent safety applications by The Dow Chem ical Company (Ref 15.11 Overton), abbreviated here as OK. This sectio n will present worked examples and case studies which represent the principles and concepts discussed in this book. They are illustrative of the many approaches that companies may take to achieve inherently safer processes. 15.8.1 Minimize Use smaller quantities of hazardous substances. Although the reaction of nitric ac id and glycerin is quite rapid, early industrial implementation of this chemistry used large batch reactors. Curre nt technology has enabled efficient contacting of the reactants, maki ng the actual reaction process continuous and rapid. These cont inuous reactors are orders of magnitude smaller than the older technology batch reactors, thus greatly reducing the potential hazard from this extremely hazardous reaction (KA). A continuous process has been developed for manufacturing phosgene on demand, thus eliminating the need for storage of liquid phosgene. Various import ant issues, such as quality control, understanding of transient reactor operation, and process control, were successfully resolved by a fundamental understanding of the chemical re action. This enabled the design of a system that met all of th e user’s requirements in an inherently safer way (KA). It is understood that mixing an d gas-liquid phase mass transfer controls the rate and efficiency of chlorination reactions. Replacing a stirred tank reactor wi th a loop reactor, specifically designed to optimize mixing and gas-liquid phase transfer has been shown to significantly reduce reactor size, processing time, and chlorine usage (KA). A 50-L loop reactor replaced a 5000-L batch reactor in a polymerization process (KA). 423

130 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS also published Guidelines for Integrating Management Systems and Metrics to Improve Process Safety Performance (CCPS 2016c). Both books provide excellent guidance on establishing and maintaining an effective metrics process. Metrics are discussed in Chapter 6 Section 6.2. 5.5 OPERATING PROCEDURES Table 5.5 lists two tools for evaluating procedures to assess if they are current, correct, and complete. Procedures are often considered the first tier of human response safeguards to prevent an unwanted process situation, however when the proc edures are outdated, missing key information, or incorrect, the likelih ood of a process upset occurring is increased. Table 5.5 Techniques for Reviewing Operating Procedures Common Tools and Methods Strengths Weaknesses Transient Operation HAZOP Focuses on hazards during transient operations that history shows are more likely than hazards during normal operation. Not all abnormal situations can be predicted, therefore, all-purpose emergency protocols will always be needed. Procedure HAZOP Like a standard HAZOP, this type of study provides structure to a review of written procedures. If the procedures are not already in good condition (see bullets in Section 3.4.1), this will be time-consuming.

155 12 REAL MODEL SCENARIO: OVERFILLING “Don't let your learning lead to knowledge. Let your learning lead to action.”—Jim Rohn, Entrepreneur, Author, and Motivational speaker Gouda Terminal is located along the Nieuwe Maas River in Rotterdam, Netherlands. The terminal is a main distribution hub that not only serves the Netherlands, but also Belgium and Germany. Its central location allows for petroleum and chemical products to be loaded from trucks, rail tank cars, and marine tankers. The terminal stores crude oil, gasoline, jet fuels, and diesel fuel from the Middle East, North Sea, and Russia. The facility consists of 25 steel aboveground, atmospheric storage tanks with a total capacity of over 400,000 m3. The petroleum products are piped in through an automated flow control system. Float-and-tape gauges are installed in all tanks to measure the actual level in the tanks. These mechanical gauges have been in use for decades and operate via a pulley system consisting of a large float inside the tank attached to a counterweight by a perforated tape. The liquid level readings are displayed in a unit located on the side of the tank. This unit is equipped with electronics that transmit level data to the company’s inventory management system, allowing for remote monitoring of tank levels. The company has also installed two high-level alarms in each tank. The operators and shift supervisors work in three shifts: 7 a.m. to 3 p.m., 3 p.m. to 11 p.m., and 11 p.m. to 7 a.m. At the beginning of each shift, the operators manually record the levels of each tank and report this to the planning and operations team. The members of this team are responsible for monitoring the materials and levels in the tanks. The Planning & Operations team is also trained to do manual calculations if there are any anomalies in the automated systems. The team manager is a stickler for making sure that The individuals and company in this chapter are completely fictional. Driving Continuous Process Safety Improvement From Investigated Incidents By CCPS and EI © 2021 the American Institute of Chemical Engineers

198 | Appendix: Index of Publicly Evaluated Incidents Section 1. RBPS Elements (In Element order) Process Safety Culture See specific Culture Core Principles as applicable Compliance with Standards – Primary Findings A5 C3, C5, C6, C18, C20, C21, C30, C37, C38, C39, C45, C47, C56, C66, C69, C77 D18, D19, D25, D33 HA3, HA6, HB4 J61, J82, J146, J164, J179, J194, J202, J203, J209, J210, J212, J215, J224, J236, J269 S9, S15 Compliance with Standards—Secondary Findings A4 C10, C17, C25, C29, C33, C35, C43, C44, C68, C72, C75 D32 HB5 J18, J27, J106, J201, J213, J229, J230, J237, J244 Competency—Primary Findings A5 C12, C19, C22, C26, C41, C42, C57 J23, J25, J26, J28, J37, J38, J39, J41, J98, J143, J144, J149, J150, J151, J152, J154, J155, J168, J170, J173, J176, J177, J180, J181, J182, J184, J185, J186, J189, J190, J192, J193, J194, J196, J197, J216, J217, J218, J220, J222, J228, J241, J251, J259, J266, J269, J270 S9, S10, S12 Competency—Secondary Findings A6, A7 C8, C9, C10, C23, C24, C43, C44, C45, C52 D32 HB4 J17, J18, J20, J21, J22, J24, J42, J50, J51, J54, J82, J96, J100, J101, J116, J127, J130, J132, J133, J134, J145, J158, J159, J160, J161, J163, J178, J187, J188, J195, J206, J214, J234, J236, J250, J252, J255, J271 S4 Workforce Involvement—Primary Findings C58 J196 S10

178 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS THE FINAL DESCENT The continued nose-up attitude, coup led with a reduction in actual airspeed when the aircraft was close to the edge of its safe operating window, triggered a stall. Despite the repeated stall warnings, the pilots continued to pitch the aircraft upwards. The report states that in the first minute after the autopilot disconnected, the aircraft was outside its flight envelope. Once the Captain entered th e cockpit, the aircraft was in a rapid descent, and he appeared unable to diagnose the problem in time. By 02:12:00, some 18 seconds after the Captain re-entered the cockpit, the altitude was descending throug h 31,500 ft and the angle of attack was about 40 degrees. The report states: Only an extremely purposeful crew with a good comprehension of the situation could have carried out a maneuver that would have made it possible to perhaps recover control of the aeroplane. The FDR stopped recording at 2:14:28, presumably when the aircraft struck the surface of the ocean. 7.1.7 Lessons Learned Relevant to Abnormal Situation Management There were many lessons learned in relation to the aviation industry, most of which are also directly rele vant to the process industries. The following section details are some of the key lessons that can be extracted from this case study, using some of the same categories and terms that were used in earlier parts of this book; it is suggested that the reader reviews the full report to obta in the most benefit from all of the learning and recommendations. 7.1.7.1 Design and HMI: Flight control systems are, by thei r nature, complex systems and require a high level of reliability. They typically have several layers of redundancy, as was the case with th e airspeed measurement. Problems with the blockage of the pitot tubes due to icing under certain meteorological conditions had occu rred in the past. This had led to disconnection of the automatic contro ls but no serious incidents. An improved design of the hardware wa s available and was being fitted to the fleet but had not been fitted to this particular aircraft.

APPLICATION OF PROCESS SAFETY TO ONSHORE PRODUCTION 101 design (see Chapter 7 for further details). Teams review these for adequacy and may recommend additional safeguards. Hazard and Operability Study (HAZOP) HAZOP is a frequently used hazard iden tification method described in CCPS (2008a). A multidisciplinary team is convened to conduct the study. The process drawing is divided into nodes. A node might be a pipe flowing between two vessels, or the vessel itself. There can be multiple equipment items in a node including valves, pumps, filters, etc. A HAZOP study uses a selection of parameters (e.g., flow, level, temperature, pressure) combin ed with guidewords (no, more, less, as well as, part of, reverse, other than) applie d to each node. This results in a rigorous identification of potential deviations. When a deviation is identified, the team assesses the potential consequences and documents any existing safeguards. Risk ranking approaches are commonly employed to assist decision making. If the team thinks the safeguards listed are not sufficient, they may recommend additional ones. Further details on the HAZOP method and risk ranking are provided in CCPS (2008a). Layer of Protection Analysis (LOPA) The LOPA method is an order-of-magnitude semi-quantitative risk analysis technique. It is being increasingly employ ed in onshore production facilities to better understand the expected frequency of a specified consequence of interest. CCPS has published the method and two additional guidelines with suggested IPLs and conditional modifiers to use in LOPA assessments (CCPS, 2001, 2013a, 2015). It should be recognized that the LOPA approach is not an identification method – it relies on HAZOP or PHA to generate the scenarios for further evaluation. Also, it considers one risk at a time and does not evaluate cumulative risks. This technique is frequently used during the design stage and is described in the CCPS references above. Facility Siting A process safety methodology often employed at larger onshore production sites is facility siting. This method is defined in API RP 752 for permanent occupied buildings and in RP 753 for temporary occupied buildings. CCPS (2018b) extends the API documents with additional guidance. The basic aim of facility siting is to a ssess realistic leak scenarios and their predicted potential fire, explosion, and toxic outcomes. These impacts are overlaid on the facility plot plan and this is used to determine if there is sufficient spacing between the hazard source and occupied building locations. Further analysis can assess the layout of process equipment within a process unit. Consideration is given to barriers or barrier elements such as gas detection and ESD, fire and blast walls, drainage arrangements, etc., as thes e affect the predicted consequences. An important aspect of facility siting is to identify what is considered an occupied space. This is not always obvious, and it is usually necessary to engage

9 • Other Transition Time Considerations 170 units, especially during Simultan eous Operations (SIMOPS), when the commissioning and start-up activities are being performed near operating equipment and processes. Guidance of these projects sh ould be coordinated by a commissioning manager who oversees the development of and detailing for a commissioning and in itial start-up plan and estimates of the budget needed to execute the plan. A typical plan for a greenfield project should include, bu t not be limited to, the following (Adapted from [31, p. Table 9.1]): The project’s scope—the equipment, process unit, or facility being commissioned; The commissioning and initial start-up team—the number of personnel, required competencies, roles & responsibilities, etc.; The training requirements for the commissioning and initial start-up team (e.g., operators, mechanics, electricians, engineers, and other technicians, including classroom training covering the technical, operational, and maintenance guidance and vendor instructions on the new equipment and processes, etc.); The contracts for third part y support (e.g., technology vendors, engineering design, etc.); The day-to-day commissioning an d initial start-up resource requirements (such as testing and verification equipment, radios, PPE, water and food, etc.); A detailed, individual task l evel schedule and prioritized sequence of systems that will be inspected, prepared, have chemicals introduced, and operated. ( Note : These tasks either (i) verify that equipment or a system functions as intended or is ready to operate or (ii) in volve operating the equipment, systems, or parts of systems. In addition, scheduling “hold points” should be established, as certain tasks should be completed before other tasks begin, such as commissioning a flare system before the process units.);

3. Options for supporting human performance 33 Figure 3-4: Supporting skill-based performance Procedures, instructions and jo b aids can all help people remember the correct steps and check their progress (which steps they have done) when working on long and complex tasks. Double-checking and peer checking task completion can help people spot and correct slips and lapses. Attention can decline when carrying out skill-based tasks and people can be distracted. Tasks and working environments should be designed to help minimize interruptions, which helps to minimize distractions. Controls and equipment can be designed to help avoid skill-based slips, such as placing the most frequently used controls closer to the operator and ensuring information displays are readable. Workload & fatigue management Task, environment & team design (distractions & interruptions) Procedures, instructions & memory aids Controls & instrumentation design Skill development, instruction & operational experience Task planning & checking See Chapter 16 for more information on task planning. See Chapter 22 for more information on task verification. See Chapter 9 for more information on equipment design.

Piping and Instrumentation Diagram Development 412 ●Technical check for fired heaters and boilers –Che ck to make sure the fired heater/boiler call‐out is complete, accurate and up to date –Che ck to make sure there are adequate devices for the rounding operator to check the flame shape. ●Technical check for instruments and control system –Che ck to make sure instruments are of the right type –Che ck to make sure instrumentation signal lines are of the correct line type 19.3.4 Design C heck The design check means checking to make sure the P&ID follows the other related documents. These “related documents” are generally interpreted as “tier 1” related documents. This means only the documents that are directly related to P&IDs are checked. If the concept is stretched to tier 2 documents checking the sizing calculations and connectivity with heat and material balance tables could also be included. This is generally beyond the scope of P&ID checking. The design check is in two levels, document connec - tivity check and cursory sizing check. ●Document connectivity check – Make sur e the P&ID follows the PFD –Make sur e the control system in the PDF is applied in the P&ID –Make sur e mark ups on the previous P&IDs are implemented correctly (make sure no new set of mistakes is generated by the drafter because of non‐set‐up drawing software) –Che ck the drafting process did not introduce any random new errors or mistakes. (Sometimes the CAD software setting is not perfect and a new set) of issues are generated after fixing some other issues by the CAD drafter. –Make sur e equipment names match other docu- ments like the PFD, equipment list, datasheet, etc. –Making sur e information on the P&ID matches the equipment data sheet. ●Cursory sizing check – Us e rules of thumb to check some sizing –Making sur e the sizes of connected items (e.g. pipe to vessel) “look” correct. 19.4 Methods of P&ID Reviewing and C hecking A P&ID checker/reviewer can decide (or be instructed) to review a P&ID in any of, or a combination of, two methods: the systematic approach and the scanning approach. In either case, it is a good idea to not start and finish the review in one day and/or by one person. To make sure all the problems of the P&IDs are captured we need some sort of “cold eye” every several hours. Otherwise the eyes of the reviewer get used to the elements of P&IDs and may not catch mistakes. Sleeping on a partially checked P&ID helps you to capture more mistakes the next morning. If the deadline doesn’t allow, you may ask someone else to check them, as a cold eye reviewer, at the end. 19.4.1 Sy stematic Approach In this method the checker checks each sheet of P&ID based on the check list he has in hand. This method works for everyone, including less experienced people, but it is time consuming. In this method, for each piece of equipment, pipe, or even instrument a check list is developed. One good source is API14C (Figure 19.6). 19.4.2 Scanning A pproach This could be a quicker method but for trained, good eyes. This is the method of choice for more experienced people who are not directly responsible for checking P&IDs but have to approve them. P&ID vessel check list P&ID tank check list P&ID pump check list P&ID pipe check list P&ID general check list Figure 19.6 Example of check list using API14C.

134 process control program for an upgraded reactor system. The system failed during water-batc hing (prior to commissioning) due to control valves having different fail-safe positions. Safety instrumented systems (SIS ) must be independent of basic process controls. The process must be designed to be tolerant of the failure of control system components or faulty software. The software should be tolerant of hardware failures and correct for them automatically, via self-diagnostics. Alarm functions are typically built into distributed control systems. The minimization strategy can be applied by reducing the quantity of process alarms to the smallest am ount which will provide sufficient alerting to operations personnel, then prioritizing these alarms, and ensuring that a program is in place to manage nuisance/spurious alarms. It is easier to train well on the response to a smaller number of (actual) alarms, ensuring that the pr oper, timely response will be made during process upsets or emergency conditions. 7.8 SUMMARY This chapter describes the application of inherently safer strategies to traditional protection layers, in orde r to improve their effectiveness and robustness. In this approach, the likel ihood of the incident occurring is reduced, perhaps by several orders of magnitude, but remains “non- zero”, in contrast with the direct application of IS strategies to the process, in which the likelihood is virtually reduced to zero. 7.9 REFERENCES 7.1 Center for Chemical Proc ess Safety (CCP S), Guidelines for Engineering Design for Process Safety . New York: American Institute of Chemical Engineers, 1993. 7.2 Center for Chemical Process Safety (CCPS), Human Factors Methods for Improving Performance in the Process Industries. New York: American Institute of Chemical Engineers, 2006.

Part 3: Equipment Human Factors Handbook For Process Plant Operations: Improving Process Safety and System Performance CCPS. © 2022 CCPS. Published 2022 The American Institute of Chemical Engineers.

272 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Source Term Models Source term models are used to quantitatively define the release scenario by estimating discharge rates, total quantity released (or total release duration), extent of flash and evaporation from a liquid pool and aerosol formation. Source term model outputs are the direct input to fireball and jet fire consequenc e models and to dispersion models that model the concentration fields downwind from the source. Scenario Specification Consequence modeling requires a detailed scenario specification. This includes details such as the material discharged, the hole location and size, the release temperature and pressure, the inventory released, and the operator response. A process unit contains many vessels, pipes, an d flanges. Where might the leak occur? It is first important to know where the hazardou s chemical is present in the process. For example, it could be present as a vapor in the to p of a vessel or it could exist as a liquid in the bottom of a vessel. A leak from a pressure vesse l is less likely than from piping. The flanges connecting piping and vessels are likely to leak first in an overpressure. event. Given the number of vessels and lengths of piping in a proce ss unit, it is still hard to know exactly where a leak might occur. It is often sufficient to identi fy sources as lying within an isolatable section and the precise location is not so important. Many release scenarios are possible. For exam ple, the release could be from a pressure relief valve in which case release details are easi er to specify. The release could be from a large storage tank or from a ship. Whatever the case , specify as many of th e scenario details as possible to improve the accuracy of the consequence analysis. Isolatable Section The next question is how much material is leak ed. It is likely that the emergency procedure (see Chapter 19) addressing a leak in the process unit may specify operator actions such as shutting down the process unit and isolating th e section that contains the leak by closing emergency isolation valves. This will limit the le ak duration and volume. The leak duration can be determined based on the specific scenario. This may depend on the detection and reaction time for automatic isolation devices and response time of the operators for manual isolation. The rate of valve closure in longer pipes may require a longer closure time to avoid water hammer and hence can affect the duration of release. An alternative is to use a conservative simplif ication, for example, the U.S. Department of Transportation (DOT 1980) LNG Federal Safety Standards specified a 10 minute leak duration. Other studies (Rijnmond Public Authority,1982) have used 3 minute for systems with leak detection combined with remotely actuated isol ation valves. Other issues to consider when analyzing discharges include the following. The released volume can be no greater than the capacity of the isolated vessels and piping once isolation is achieved. Time dependence of transient releases: De creasing release rates due to decreasing upstream pressure.

80 PROCESS SAFETY IN UPSTREAM OIL & GAS Past optimum well drilling performance becomes a target for the new well and helps to include lessons learned. Bow Tie Analysis The bow tie analysis is a newer method which has a focus on explaining and demonstrating risk management barriers rather than being a tool for decision making directly. It is part of the RBPS element Hazard Identification and Risk Analysis . The bow tie method is described in CCPS (2018c) and briefly in this book in Section 2.7. An example bow tie diagram is shown in Figure 2-11. Bowties are helpful in providing a visual representation of the barriers that prevent or mitigate a risk. They provide an overall view for those charged with managing risks. They also provide a specific view of how the various barriers eliminate or mitigate risk which allows the individual operator or maintenance technician to see how their work supports risk management. The bow tie is not a decision tool directly as it does not quantify the risks of each arm, but it does show wh ere there might be a lack of barriers (e.g., a pathway with only one barrier) or an excess number. A key use and benefit of the bow tie is to identify where and what type of barrier health data should be collected and on what frequency. Bow tie diagrams are created in team se ssions that start with the outcome of some prior hazard evaluation (PHA, What-If or HAZOP) combined with risk ranking. These generate the higher risk sc enarios and the more important scenarios can be selected for bow tie creation. The hazard evaluation minutes contain a column titled safeguards which contains a mixture of barriers (full IPLs) and degradation controls supporting those barriers. The facilitator collects the team results and constructs the bow tie diagram, usually using commercial software (CCPS, 2018c provides several examples). An example bow tie diagram was shown earlier in Figure 2-11. Bow tie diagrams get complex if there are many different threat arms and software can display or hide barrier decay mechanism arms as appropriate to make diagrams easier to understand (Figure 4-5) (CCPS, 2018c). 4.3.3 Asset Integrity and Reliability Developing a Well Integrity Plan is key for well construction. The well integrity plan is similar to the Operating Procedures and Safe Work Practices elements in RBPS. Well integrity is presented in AP I RP 100-1 for onshore fracking wells, Norsk O&G (2012 and 2016) for offshore wells, and ISO 16530-1:2017 (well integrity for the lifecycle) for all types of wells, and by company internal well integrity guidelines. Well integrity both onshore and offshore is addressed in separate API standards which collectively deliver the required well integrity.

EVIDEN CE ANALYSIS & CAUSAL FACTOR DETERM IN ATION 185 9.3.3.1 Evaluate Conditions at the Site - Step 1 Evaluating conditions at the failure site can be a critical important step in the analysis. Understanding the conditions at the site and how the parts or items were used can eliminate some potential failure mechanisms from consideration and support other failure mechanisms. Typical questions to ask include, but are not limited to: • How long had the part been in service? • What were the environmental conditions? • Did the failure occur during startu p, shutdown, abnormal, or normal operation? • Was it a rotating pi ece of equipment? • Did it rub against something? • Was there any fluid or gas flow past the device? • What are the chemicals to wh ich the part is exposed? • What are the materials of construction? • What activities were taking place in the area? • Does any portion of the process utilize reactive chemistry? • Is there a potential for reactive in teractions (caused by inadvertent mixing of incompatible materials) at the site? If so, what are the materials? • Is remaining, relevant equipmen t properly installed (alignment, rotation, etc.)? • Is any equipment used for more than one service? Does it require cleaning before reuse? Investigators can supplement/edit th is list based on the particular circumstances associated with the inci dent they are inve stigating. The answers to these questions should a llow the investigators to focus their subsequent data collection efforts. 9.3.3.2 Perform a Preliminary Component Assessment - Step 2 During Step 2, a preliminary analysis of the parts is performed. Typically, the focus is a visual examination of the items. The investigators should avoid disturbing evidence until necessary, conducting their visual examination without alterations. When it is time to do so, remove the parts in a planned, controlled, careful, and methodical manner. Photog raph each step of the disassembly.