Datasets:

hndc

/

AI4Hazard-small

like 0

Piping and Instrumentation Diagram Development 322 This is a better arrangement when the horizontal cold stream is a process stream and has another responsibility downstream of this heat exchanger. However, the temperature of the horizontal cold stream is still changing and, if it is not tolerable, we can-not do anything for it in this heat exchanger. We can have another variation of the arrangement in Figure 15.56, by putting control valves on both the dis - charge and bypass streams, as shown below. We have already shown split‐range control of valves. In order to maintain a constant flow rate when we have flow from both streams at the same time, we need to go with parallel control. We could substitute the two control valves in Figure 15.56 with a single three‐way valve, as shown in Figure 15.57.The disadvantage of using three‐way valves is that they are not available in large sizes. They are also not  suitable when they receive a stream with large tem perature variations. This will cause different rates of expansion inside the valve, which inevitably leads to leakage. One very interesting variation of bypass control is shown in Figure 15.58. In the arrangement shown in Figure 15.55, we have a pressure differential controller (PDC) on the bypass line. If we have a situation where the pressure drop across the heat exchanger is large, most of the inlet flow will choose to go through the bypass line, where the pressure drop will be smaller. To avoid this, we can put in a PDC to control the pressure drop on the bypass line and thereby control the flow rate through that line. The situation where it is a big advantage to use a PDC on a bypass line is where you have a number of heat exchangers operating in parallel. If you didn’t use a PDC, you would have to have a bypass line with a control valve for each and every heat exchanger. However, if you use a PDC, you only need to install one bypass line across the whole bank of parallel heat exchangers, with one control valve. TC PARALLEL RANGETT TE Figure 15.56 Heat e xchanger bypass control with two control valves.When the fluctuation in temperature may be too great for the equipment downstream, we can solve this by putting another heat exchanger in series with this one to control the temperature for downstream equipment. The second heat exchanger must be a utility heat exchanger. TC TT TEFigure 15.57 Heat e xchanger bypass control with a three‐way valve.

A.4 Report References | 219 NPO Association for the Study of Failure (ASF) of Japan Incident Database (Continued) (For incident reports J1–J163: see www.shippai.org/fkd/en/lisen/cat102.html) Code Investigation J16 Leakage and Fire of Hydrogen from a Mounting Flange of a Safety Valve in a Reactor at a Succinic Acid Manufacturing Plant (1998) J17 Explosion of Coke Oven Gas During Cleaning at a Desulfurization Regeneration Tower of a Coke Oven Gas Refining (1998) J18 Fire of Xylene Remaining in Solid Piperazine Separated in a Centrifuge (1998) J19 Fire of Ethanol Caused Due to Air Intake in the Ejector of a Treatment Drum at a Surfactant Manufacturing Plant (1998) J20 Explosion Caused Due to Generation of a Combustible Gas-Air Mixture at a Naphthalene Oxidation Reaction Plant (1998) J21 Explosion of Acrylic Acid in the Drum Can in the Heating Cabinet for Dissolution (1998) J22 Damage to a Tank Roof Caused Due to Sticking of a Breather Valve During Transfer of Raw Material (1998) J23 Explosion of Silicone Products Dissolved in an Organic Solvent During Subdivision Work (1997) J24 Explosion in the Polycondensation Reaction of Benzyl Chloride (1997) J25 Ignition of Rubber Remained in the Reactor During Cleaning at a Polybutadiene Manufacturing Plant (1997) J26 Explosion of a Machine for Melting and Volume Reduction of Polystyrene Foam (1997) J27 Explosion and Fire Caused Due to Gas Leakage from High-Pressure Ethylene Piping at an Ethanol Manufacturing Plant (1997) J28 Explosion During Charging Operation of Raw Material Powder into a Reactor Containing Dioxane (1997) J29 Explosion of an Air Heater of a Boiler at an Agricultural Chemical Manufacturing Plant (1997) J30 Explosion of an Intermediate Concentration Tank at an Insecticide Manufacturing Plant (1996) J31 Explosion Due to an Incompatible Reaction in a Nitration Workroom for TNT (1996) J32 Explosion and Fire Induced Due to Incompatible Reactions of Residual Contaminant at an Alkylaluminium Manufacturing Plant (1996)

Figure 15.10: Aqueous Ammonia Supply Proposal The design team conducted a Haza rd and Operability analysis that raised several concerns regarding the tank truck delivery system and associated operations: the risk of spills and operator erro rs for the tank truck portion of the delivery system was higher for this option than an anhydrous ammonia system higher capital, operating and maintenance costs the reliability of the addition of pumps in the system. So, the project was recycled back to the option selection phase. 414

Table B.3. Generic Like lihood (L) Descriptors Likelihood Short descriptor Description 1 Low Not expected to occur in life of facility 2 Medium Possible to occur in life of facility 3 High Possible to occur in range of 1 year to 10 years 4 Very High Possible to occur at least once a year B.1 INHERENT SAFETY ANALYSIS – GUIDED CHECKLIST PROCESS HAZARD ANALYSIS (PHA) Table B.4 offers an example of a gu ided checklist approach. The analyst asks the questions from the checklis t (potential opportunities) and the team documents the potential conseq uences of any issue that may be applicable to the process or node under study. Considering the four ISD strategies, the team documents th e potential recomme ndations that may address the concern ranked in the following order: •First order ISD •Second order ISD •Layers of Protection 459

2.6 Understand and Act on Hazards and Risks |49 2.6 UN DERSTAN D AN D ACT UPON HAZARDS/RISKS Flixborough, N orth Lincolnshire, UK, June 1, 1974 A vapor cloud explosion following the failure of temporary bypass piping killed twenty-eight workers. M any other workers suffered injuries and significant onsite and offsite property damage occurred. The tem porary piping had been installed to bypass the fifth oxidation reactor in a chain of six. Reactor five had failed and was being repaired. Supported only by conventional scaffolding, the tem porary piping was installed without first Understanding and Acting Upon the Hazards and Risks . Considering the haste to install the bypass and the close spacing of work areas on the site, the facility appeared to have a weak Sense of Vulnerability . After a two-month exposure to stress, vibration, and fatigue, the piping failed, creating a large release of flammable vapors. The Flixborough incident hastened passage of the UK Health and Safety at Work Act. While it predated the development of form al PSMS elements as we know them, it rem ains a classic example of failures of the Management of Change (MOC) and HIRA/PHA elem ents. B oth elements rely heavily on dedication to understanding hazards and risks, and how they can change as the process changes. Understanding hazards is also a key aspect of the PSM S element “Competency” (see section 5.4). Leaders should understand the difference between hazards, risks, and the safeguards that are used to act on these hazards and risks. The hazard of a m aterial is the harm it can inflict. Process hazards include toxicity, flam mability, reactivity, high and low pressure, and high and low temperature. Physical impact, electrical shock, and suffocation m ay also be process hazards.

190 INVESTIGATING PROCESS SAFETY INCIDENTS 9.5 HYPOTHESIS TESTING The following discussions are intended as an introduction to some special techniques used by experts for technical analysis of evidence and hypothesis validation. Novice investigators and individuals who are not experts in these fields should be cautious when applying these tools. For most minor investigations, review and application of the information in this section is adequate for the investigation team to analyze the data. However, if legal concerns arise during an investigation, ex perts in the forensic analysis of data should be used to ensure a proper an alysis has been performed and correct interpretation of the data has occurred. 9.5.1 Engineering Analysis In addition to physical analytical me thods, engineering analysis tools and methods are also useful during inciden t investigations. Engineering analysis refers to calculations that can be performed to investigate and test various hypotheses. Examples of engineering analyses include: Forces Stresses Fluid motion and pressure Heat transfer/temperature Thermodynamics/en ergy transfer Mass transfer and balance Mass of process fluids and process equipment Concentration of fluid in process equipment Flow rates of fluids through process equipment and through release points Change in levels of tanks over time Rates of chemical reactions Dispersion of a gas Investigators use engineering analys is methods to test the various hypotheses that are put forth duri ng the investigation. Often rough calculations may be all that is needed to determine if a hypothesis is possible. For example, even if the entire conten ts of a tank are released, the volume may not be sufficient to cause an overflow in another part of the process. A simple calculation may be sufficient to eliminate certain hypotheses that have been proposed.

262 document the reasons why items were not considered, for example, if they were not applicable or had been considered previously. Documentation of rationale for rejecting potential IS opportunities (cost, creation of other safety or operability problem, etc.). Recommendations/action plans for further evaluation or implementation of IS alternatives identified during the study. If the IS review was conducted as pa rt of a larger study (i.e., PHA or hazard review), this information shou ld be incorporated into the report of this activity. It is recommended that this information becomes a part of the permanent process safety file and be maintained for the life of the process. Electronic versio ns in an editable format (i.e., MS Word) should be maintained to facilitate futu re updates and revalidations. The rationale for why recommenda tions from IS reviews were rejected should follow the followi ng guidance, which includes for declining recommendations from inci dent investigations and process hazards analyses: The analysis upon which th e recommendations are based contains factual errors. The recommendation is not necessary. For example, the safeguards may be inadequate, but the consequences are operational, or the consequence or severity of the scenario would not result in a significant release. Another IS alternative would prov ide a sufficient level of hazard reduction. (NOTE: Implementing only one option to address identified hazards may not be ad equate to address the greatest hazard reduction or elimination. However, it is not necessary to implement more than one IS altern ative if the implementation of a second IS alternative does not add any significant hazard reduction or has been documented as not feasible.) The recommendation is not feasible due to one or more of the reasons listed below: oThe recommendation is in conf lict with existing federal, state, or local laws.

38 Guidelines for Revalidating a Process Hazard Analysis 3.1 PRIOR PHA ESSENTIAL CRITERIA 3.1.1 Prior PHA Methodology Used The prior PHA methodology refers to its core methodology and any complementary analyses as discussed in Section 1.2. Aside from policy or regulatory demands, two ke y questions should be considered when determining whether the prior PHA methodology was appropriate: 1. Was the PHA methodology appropriate for the complexity of the process? 2. Did the PHA methodology comprehensively identify the hazards of the process, the engineered and administrative risk controls, and the worst credible consequences assuming failure of all those controls? For example, PHA results comprised of only a short one-page checklist with yes/no answers and related comments would be insufficient as a core methodology for a PHA of a complex and hazardous unit involving reactions, separations, and so forth. Likewise, an unstructured What-If Analysis only identifying loss scenarios where the team had a concern or recommendation would generally be considered inadeq uate as a core PHA methodology. However, a thorough What-If Analysis co mbined with equipment- or process- specific checklist analyses might be an appropriate PHA methodology for some processes. Under most circumstances, a PHA including a properly applied and fully documented HAZOP as the core methodology will identify the process hazards, risk controls, and consequences of failure of the controls for each section of a process. A PHA structured with a well-organized and fully documented What- If/Checklist Analysis using appropriat e process- and equipment-specific checklists can also accomplish these objectives. For processes involving compressors, centrifuges, or highly automated systems, an FMEA can be conducted and documented in a manner that meets these objectives. Any of these core methodologies ca n, by themselves, be applied and documented in a manner that identifies process hazards, risk controls, and consequences of failure of the cont rols. However, the most comprehensive PHAs also include complementary analyses, and these are often in the form of checklists.

320 | Appendix E Process Safety Culture Case Histories between the com pany incident com mand and the local emergency response agency confused emergency response organizations and delayed public announcements on actions that should be taken to m inim ize exposure risk. In m anaging the crisis, the com pany reported that “no toxic chem icals were released because they were consumed in the intense fires.” While a reasonable assum ption, investigators found that air monitors placed near the unit to detect toxic chem icals were not operational at the tim e of the incident, so this could not be confirm ed. Managem ent also attempted to prevent public access to inform ation about the accident by asserting that the facility was covered by regulations related to sensitive security information. This assertion was determ ined by the governing authority to be without basis. Managem ent later acknowledged that this was done due to lim it the potential outcry related to existence of the highly toxic chem ical at the plant. The investigators provided num erous exam ples of the com pany using good engineering and operating practices to protect against releases of the highly toxic chem ical, including reducing inventory, locating the m ain storage tank underground, shielding the above-ground day tank, and providing a dump tank if necessary to rapidly em pty the day tank and associated piping. And in fact, these procedures were effective and well-m anaged. While investigators did not exam ine culture, readers can deduct from the investigation report that the process safety culture related to this unit was robust. However, it is not clear that the PSM S and culture was functioning as well in the adjacent unit. If the investigators had exam ined culture, what potential culture gaps m ight the investigators have considered exploring? Did an extra high sense of vulnerability from the highly toxic chem ical reduce com pany em ployees’ sense of vulnerability related to other chem ical and processes?

REACTIVE CHEMICAL HAZARDS 89 The owners did not do any reaction testing such as adiabatic calorimetry (e.g., Accelerating Rate Calorimeter™ (ARC), Vent Sizing Package™ (VSP), Phi-Tec, or Automatic Pressure Tracking Adiabatic Calorimeter® (APTAC)), although this type of testing had been good engineering practice for years. The CSB noted that process safety was not part of the chemical engineering curriculum in almost 90% of universities at the time of the incident. In its report, the CSB recommended to the AIChE and the Accreditation Board for Engineering and Technology, Inc. (ABET) that awareness of reactive chemical hazards be part of the baccalaureate program (CSB 2009). This recommendation was implemented by the ABET, in fact, the CSB notes that the action exceeded the CSBs expectations. Hazard Identification and Risk Analysis . Even though a design consultant recommended that T2 do a Hazard and Oper ability (HAZOP) study on the process, T2 apparently did not do one. If the MCMT proces s had been reviewed by a competent PHA team questions such as, “what happens if the temperatur e is too high?” or “what if the cooling fails?” would have come up. These questions would le ad to recommendations such as: determine what the safe operating temperature is, what ha ppens if it is exceeded, how can we make the cooling system more reliable, or what othe r safeguards can be provided against high temperature and pressure? Asking these questions could also have led to a better understanding of the emergency relief requirements. The emergency relief syst em (ERS) was based on the maximum rate of hydrogen generation in normal operation (CSB 2009). The ERS was inadequate for the reaction that occurred. After subsequent testing in a VSP, the CSB determined that the second exothermic reaction was so fast that the re actor could not have been successfully protected by a relief device. The only way to protect th e reactor from overpressuring was to vent the reactor during the first reaction and allow the energy to be removed by boiling off the diglyme solvent and MCPD. Management of Change (MOC). After one year of production the batch size was increased by one-third, without a safety revi ew. However, without the needed competency to recognize reactive chemical hazard s, an MOC would not have helped. Emergency Management. T2 did not warn emergency responders of the presence of MCMT on site. MCMT is toxic by inhalation and skin contact. Incident Investigation. Prior to the explosion, there had been unexpected exotherms in three of the first ten batches during the first re action step when the pr ocess was scaled up to the main reactor. After the first exotherm (in Batch 1), the response wa s to adjust the batch recipe and to add cooling to the operating pr ocedures. Uncontrolled exotherms also occurred in Batches 5 and 10. Nevertheless, after Batc h 11, the process scale-up was considered successful. The owners did not recognize that the previous exotherms were actually near misses which could have had more severe consequences, and therefore failed to further investigate the causes of these exotherms. A video about the T2 Laboratories explosion can be found on the CSB website at http://www.csb.gov/videos/ .

286 Employing the state-of-the-art in the design of the processing technology is also a use of Substitution and Moderation , both in the process engineering, as well as in the engineering of the equipment. Keeping up-to-date on the state-of-the-a rt is an aspect of process safety competency as it requires proactive activities to obtain and maintain knowledge and expertise in the rese arch, development, and engineering of the type of process technology in use. Also, in a basic sense, process Simplification should result in less required training
BT
MBZFST
PG DPNQ lexity are removed fr om the process, less IVNBO
JOUFSGBDF
JT
SFRVJSFE and
the hu man interface that is required is TJNQMFS
FBTJFS
UP
VOEFS TUBOE
and easier to operate. 11.10 MANAGEMENT OF CHANGE / OPERATIONAL READINESS Management of Change and Operational Readiness (or Pre-startup Safety Review as it is often referred to) are often combined elements by organizations in the same overall MO C procedure in PSM programs, with the OR/PSSR one of the last steps in the MOC process. These elements offer several ways to incorporate the four IS strategies. First and foremost, MOC is required to be used whenever IS changes to existing equipment are being proposed. However, the reverse should also be true, that is, whenever a change is being contemplated to existing equipment, the MOC process should in clude IS considerations. This is just as important as looking for IS opportunities while processing a MOC for other reasons. The use and consider ation of IS strategies should be incorporated into the identification of prospective changes and their technical basis, the MOC safety and hazards reviews or PHAs conducted to examine the potential process safe ty impacts of prospective changes, and the modification of operating an d other procedures pursuant to a change. In particular, the safety review of the proposed change can employ checklists with IS questions similar to what has been described for PHA/HIRA above. The four IS strategies should also be included in MOC/PSSR procedure checklists. The overall MOC process and where IS guidewords and checklists can be used in it is shown in Figure 11.3 (Ref 11.16 Kletz 2010). Table 11.2 shows a list of MOC questions that can be used when reviewing proposed chan ges. As with Process Knowledge Management, the use of IS strategies a n d c o n c e p t s i n a d e s i g n o f a process should be clearly document ed because they are inherent and

380 14.2.4 Safer Technology & Alternatives Analysis – Revised US EPA Risk Management Program (RMP) Rule This final section on specific Un ited States regulations with IS requirements focuses on the 2018 the U.S. Court of Appeals for the DC Circuit decision to vacate the de lay of the final revised RMP Rule published in the Federal Register in 2017. The revised RMP Rule contains a provision to perform a Safer Technology & Alternatives Analysis (STAA) as part of performing PHAs on RMP-covered processes. US EPA modified the PHA provisions in the RMP Rule by adding a requirement for certain industry se ctors to conduct a STAA and to evaluate the practicability of any inherently safer technology (IST) identified. The practicability study will determine the costs and assess the reasonableness of implementing technology alternatives. US EPA limited the applicability of this requ irement to owners or operators of facilities with RMP Program 3 regulated processes in North American Industrial Classification System (NAICS) codes 322 (paper manufacturing), 324 (petroleum and coal products manufacturing), and 325 (chemical manufacturing). In the proposed rulemaking, US EPA specified that the STAA would cons ider, in the following order of preference: IST or inherently safer design (ISD), Passive measures, Active measures, and Procedural measures. US EPA further indicated that the owner or operator would be able to evaluate a combination of th ese risk management measures to reduce risk at the process. US EPA did not mandate the adoption of any IST found to be practicable in part because we recognize that a passive measure or other approach on the ST AA hierarchy may also be effective at risk reduction and left the adoption of particular accident prevention approaches to owners’ and operators’ reasonable judgment. US EPA also added several definition s that relate to an STAA. US EPA defined active measures to mean risk management measures or engineering controls that rely on me chanical, or other energy input to detect and respond to process devi ations. Some examples of active

15. Worked Examples and Case Studies 15.1 INTRODUCTION This chapter illustrates the application of IS principles and concepts in both idealized and actual situations. It also includes a post hoc consideration of IS opportunities as applied to the Bhopal tragedy, dramatically illustrating the potential benefits to both the facility and the surrounding community from identifying and implementing IS opportunities. 15.2 APPLICATION OF AN INHERE NT SAFETY STRATEGIC APPROACH TO A PROCESS As discussed in Chapter 8, inherent safety (IS) concepts can be considered throughout the life cycle of a process. The following example illustrates the concepts described in Chapter 2 (see Figure 2.3), as applied over the life cycle of a process. Reactive Chemicals, Inc., a fictiona l coatings industry supplier, is planning to install a new polymeriza tion unit to produce Intermediate C and Final Product Z. The final product goes into various coatings industry applications. Industry expectations ar e for lower solvent formulations of this type polymer. The following illustrates the processes involved: Intermediate production: A + B = C In the intermediate reaction, raw material A is reacted with raw material B to produce intermediate C. Current production is in a batch reactor with all materials, including the catalyst, in the initial charge. Raw material A is flammable (flash point <100ºF), toxic, and supplied and stored in bulk. Raw material B is a reactive monomer that is corrosive (to human tissue) and combustible (flash point >100ºF) and is typically inhibited with hydroquinone (HQ) or methoxyhydroquinone (MEHQ). Like Raw Material A, it is supplied and stored in bulk. The catalyst used for the intermediate pr oduction is boron trifluoride (BF3), 388 (VJEFMJOFT
GPS
*OIFSFOUMZ
4BGFS
$IFNJDBM
1SPDFTTFT
"
-JGF
$ZDMF
"QQSPBDI #Z
$$14 ¥

5IF
"NFSJDBO
*OTUJUVUF
PG
$IFNJDBM
&OHJOFFST

56 INVESTIGATING PROCESS SAFETY INCIDENTS responsibility a corporation assumes once it has increased knowledge of a hazard or remedy. Failure to act on this knowledge may result in much more significant legal and regulatory consequences. The management system can include actions for companies to take when preparing for an agen cy inspection. Whether or not to consent to an immediate entry by govern ment inspectors in the aftermath of an incident is a difficult question to answer in any situation. It is impossible to answer generically. Consider involving legal co unsel in these situations. Remember that the incident site and evidence may come under regulator control. Facility managers should be aware of the company’s righ ts with regard to unreasonable searches and seizures. A governme nt entry into and search of a facility in the wake of an incident may be unreasonable. It may be appropriate to refuse to consent to entry in some cases. In others, it may be appropriate to consent to the government entry under specific conditions. The conditions might include limits on the scope and duration of the inspection or specific agreemen ts about the taking and sharing of photographs and interviews of employees . Of course if the visitors have one, the terms of a government agency warrant must be followed. In general, cooperating with an agency seeking to perform an investigation is the best approach. In the long term, this appr oach can help forge a good working relationship wi th the agency. Whether or not the agency is admitte d to the facility by the consent of the facility or under a warrant, the agency’s purpose should be kept in mind. That purpose may not be the same as that of the company incident investigation team. The company seeks to identify the factors contributing to the incident and the underlying causes . The agency also seeks to identify regulatory violations and evidence that may lead to an enforcement action. A regulator’s approach to incident investigation has to be different from the company’s as “proof beyond a reas onable doubt” is required if a criminal case is justified. Of course, both parties want to ensure that lessons are learned to prevent future incidents. Especially when an accident causing death or personal injury has occurred, government investigators are likely to assume that a preventable condition caused the incident, that the condition violated a statute or regulation, an d that regulatory penalties should be imposed. Agency involvement presents challen ges from the facility’s perspective. Facility personnel need to manage the incident and its aftermath, but may a l s o b e a s k e d t o d i v e r t resources to accommodate agency personnel. Personnel should cooperate with authorities but should avoid volunteering unnecessary or unconfirmed information. Plant staff may be asked

7. Developing content of a job aid 73 A common approach to engaging oper ational and maintenance staff in developing procedures for existing tasks is to conduct a walk-through or a talk- through, possibly aided by a video recording of the task. It can help to use a questionnaire to assi st with the production of the job aid or procedure. This should include questions, w i t h p r o m p t s o r s u g g e s t i o n s t o h e l p encourage information sharing, and to ca pture details about the task. It should also be used to record responses duri ng the walk-through, to capture detailed information about the task steps. An example process is provided in Figure 7-3. The Human Performance Oil and Gas (H POG) group also provide a Walk Through Talk Through template and guide [ 35]. This is a free resource that also covers capturing task steps, potential errors and ideas on error prevention. In the case of new processes, tasks ca n be viewed or imagined by use of process flow, functional, instrumentation diagrams and/or 3 dimensional models. If available, drawings or mo ck-ups may b e used to help identify task s and sub- tasks. Some important information that can be realized or obtained from a walk- through are 1) assumptions or preconditi ons assumed when starting the task, 2) opportunities for errors, and 3) possible different ways of doing things. Task walk-through The walk/talk-through approach is a simp le process that consists of a person, with knowledge of a task, demonstrating how it is done, while being observed by someone else. It should be a fair and accurate refl ection of how the task is actually performed. A task walk-through should be completed prior to first use of a procedure.

62 INVESTIGATING PROCESS SAFETY INCIDENTS note, email, report or communication as if it would become a public document available to the press, go vernment or the public in general, including competitors. Regulatory requ irements may dictat e that reports on process safety are to be shared with workers, depending on jurisdiction and type of incident. Other protections that may apply include The W ork Product Doctrine and The Self- Critical Analysis Privilege (Adams, 1999). The work product doctrine was created to protect materials prepared in anticipation of litigation from discovery. Although technically speaking a lawyer might no t have to be involved for material to acquire work product protection, atto rneys may need to be involved for several reasons. First, some rulings have favored the involvement of a lawyer. Second, involving a lawyer suggests the matter should not be considered ordinary course of business. Third, the lawyer’s involvement emphasizes that the work is being done in anticipation of litigation. 4.2.3.2 Recording the Facts There may be a perceived conflict between the need of the investigation team to gather information quickly and record observations versus the legal risk the company could face fr om hastily prepared notes or erroneous preliminary conclusions. Haste in making notes without clearly distinguishing between factual observations and speculation can cause unnecessary legal risk to the company. The company could spend a great deal of time and money trying to explain the hasty notes in litigation or enforcement actions. The investigation team should take accurate notes and record only facts. Any opinions or speculation should be clearly noted as such. Facts cannot be altered, but conclusions can change as the investigation continues. In some cases, the legal counsel should review documents that are prepared by the investigation team for outside distribution as well as the final offici al reports as they are drafted. The guidance by legal counsel can help to limit unnecessary liability. Typical guidance to investigators regarding note and report writing may include: • Using header and footer designation s to identify official incident team internal documents. Lega l counsel may recommend adding statements such as, “Privileged and Confiden tial—Attorney–Client Privileged Information” or other designators on each page of certain documents • Refraining from use of superlatives and inflammatory language; rather, use factually accurate statements • Refraining from use of judgmen tal words with special legal

Event: First European Conf. of Young Res. Chem. Eng , July 14-18, 1996 (pp.62-64). Rugby, UK: Institution of Chemical Engineers. Edwards, D.W., Lawrence, D. , and Rushton, A.G. (1996). Quantifying the inherent safety of chemical process routes. In 5th World Congress of Chemical Engineering , July 14-18, 1996, San Diego, CA (Paper 52d). New York: American Instit ute of Chemical Engineers. Eierman, R. G. (1995). Improvin g Inherent Safety With Sealless Pumps. In E.D. Wixom and R.P. Benedetti (Eds.). Proceedings of the 29th Annual Loss Prevention Symposium , July 31-August 2, 1995, Boston, MA (Paper 1e). New York: American Institute of Chemical Engineers. Emsley, J. (12 March, 1994). A cleaner way to make nylon. New Scientist , 15. Englehardt, J. D. (1993). Pollution prevention technologies: A review and classification. Journal of Hazardous Materials, 35, 119-50. Englund, S.M. (1990). Opportunities in the design of inherently safer chemical plants. Advances in Chemical Engineering , 15, 69-135. Englund, S. M. (1990). “The design and operation of inherently safer chemical plants.” Presented at the American Institute of Chemical Engineers 1990 Summer National Meeting, August 20, 1990, San Diego, CA, Session 43. Englund, S. M. (1991a). Design and operate plants for inherent safety - Part 1. Chemical Engineering Progress , 87 (3), 85-91. Englund, S.M. (1991b). Design and operate plants for inherent safety - Part 2. Chemical Engineering Progress , 87 (5), 79-86. Englund, S.M. (1993). Process and design options for inherently safer plants. In V. M. Fthenakis (ed.). Prevention and Control of Accidental Releases of Hazardous Gases (9-62). New York: Van Nostrand Reinhold. Englund, S. M. (1994). “Inherently safer plants—Practical applications.” Presented at the American Institute of Chemical Engineers 1994 Summer National Meet ing, August 14-17, 1994, Denver, CO, Paper No. 47b. 476

EVIDEN CE ANALYSIS & CAUSAL FACTOR DETERM IN ATION 179 determining the cause of loss of containment that led to a flammable material release and fire. Figure 9.1 Scientific M ethod Process Data collection is the second step of the Scientific Method process. This includes examination of the scene, measuring and documenting damage, interviewing witnesses, and data collection activities, as described in Chapters 7 and 8. The collected data are anal yzed in the third step. Analysis refers to all manners of evaluating data, including examination and testing of physical data, engineering calculations, sy stems testing, simulations, and reconstructions as described in this chapter. Observations, measurements, data analysis and other information are used to formulate hypotheses in the fourth st ep. Hypothesis formulation is inductive reasoning. It is important to recognize that inductive reasoning involves postulating a reasonable conclu sion from the available data, but the conclusion may not necessarily be true. For example, it may be hypothesized that a pipe burst because the internal pressure exceeded the pipe’s pressure capacity. However, it remains to be proven that the pipe failed due to excessive pressure rather than corros ion, a material defect, some other cause, or a combination of factors. It may appear to be unproductive to postulate hypothes es that may not be true. However, during the course of an investigation, data may not be available to prove or disprove a hypothes is at the time that a hypothesis is postulated. By postulating the hypoth esis, investigation activities can be developed to evaluate the hypothesis, such as metallurgical examination of

6 • Recovery 105 processes, as was illustra ted in Figure 3.2 and Figure 3.3, respectively. An abnormal situation is defined as “a disturbance in an industrial process with which the Basic Proces s Control System (BPCS) of the process cannot cope” [34]. The recove ry efforts to control successfully the process safety risks depend on the integrity and reliability of the engineering controls and on operational discipline from those responding through the admi nistrative controls [21]. Since relatively “small” process deviations are expected, they can be successfully responded to, as is illustrated for a co ntinuous process in Figure 6.3’s timeline. The process deviations can be anticipated and determined beforehand using the hazards and risks analysis approaches, as discussed briefly for a higher pressure deviation in Section 6.3. Then, with the under standing of the deviations with process safety risks that should be managed, the engineering and administrative controls required for normal operations can be identified, designed, implemented, and maintained. These controls help the operations team safely re turn the process to its standard operating conditions once the deviations are detected. Engineering controls include understanding the dynamic characteristics of the Basic Process Control Sy stem (BPCS) [59]. The administrative controls, the procedures for “normal operations” are written to guide the operations team on the standard (e xpected) operating conditions and often provide minimal troublesh ooting protocols to help with the recovery efforts. As was noted earlier in this chapter, when the recovery efforts for an abnormal situation are unsuccessful, then the operations team will shut the process down.

PROCESS SAFETY AND MANAGEMENT OF ABNORMAL SITUATIONS 11 process hazards to aid managing abnormal situations (CCPS 2007a, 2015, 2011a), Asset Integrity and Reliability to ensure that process equipment and control systems remain fit for purpose and reliable throughout their life to minimize challenges to protection layers (CCPS 2007a, 2017a, 2017c, 2007b), Conduct of Operations and Operational Discipline to ensure that all tasks including those e ssential for safe operation are performed reliably to minimize errors leading to abnormal situations (CCPS 2007a, 2011b, 2018f), Process Safety Culture to maintain the values and behaviors of a sound culture to deliver safe operations and improve human factors to help provide the conditions that support maximum performance of workers during abnormal situations (CCPS 2007a, 2018c, 2006, 2004), and Incident Investigation to learn from experience of prior abnormal situations and take action to strengthen management systems and process control to avoid and/or mitigate future abnormal situations (CCPS 2019). Application of these and other process safety elements for managing abnormal situations is discus sed in detail in Chapter 3. 2.2 THE CASE FOR POSITIVE MANAGEMENT OF ABNORMAL SITUATIONS An abnormal situation typically starts with one or more operating parameters drifting outside normal limits that may impact product yield and quality. However, if this cond ition is not managed positively and quickly, the situation can rapidly escal ate to a more dangerous and costly event that may include downtime, lo st production, equipment damage, or injuries, as well as external prop erty, environmental, and reputational damage. Figure 2.2 (BakerRisk 2021) illustrat es the concept of operating limits, showing the deviation of an oper ating parameter from “normal”, through a “troubleshooting” zone, in to an “emergency” zone. Once the

TOOLS AND METHODS FOR MANAGING ABNORMAL SITUATIONS 145 Example Incident 5.4 – Flight 173 DC-8 Crash in Portland, 1978 (cont.) The NTSB considered th at the accident was an example of a recurring problem: “… A breakdown in cockpit management and teamwork during a situation involving malfunctions of aircraft systems in flight .” The report continued: Admittedly, the stature of a captain and his management style may exert subtle pressure on his crew to conform to his way of thinking. It may hinder interaction and adequate monitoring and force another crewmember to yield his right to express an opinion. The first officer’s main responsibility is to monitor the captain. In particular, he provides feedback for the captain. If the captain infers from the first officer’s actions or inactions that his judgment is correct, the captain could receive reinforcement for an error or poor judgment. The final recommendation in the NTSB report was as follows: “Issue an operations bulletin to all air carrier operations inspectors directing them to ur ge their assigned operators to ensure that their flight crews are indoctrinated in principles of flightdeck resource management, with particular emphasis on the merits of participative management for captains and assertiveness training for other cockpit crew members.” The investigation led to the development of assessment and training on Crew Resource Management (CRM). Today, CRM has evolved to cover many issues that are highly relevant to the management of abnormal situations. Outside the aviation indu stry, it is sometimes called Team Resource Management (T RM) or Non-Technical Sk ills (NTS, or NOTECHS). It can be defined as “ the cognitive, social and personal resource skills that complement technical skills, and contribute to safe and efficient task performance ” (Flin et al 2003). It primarily focuses on leadership techniques and effective management of resource s, but also concerns the cognitive skills that are required to gain and maintain situation (or situational) awareness, particular ly in stressful situations. The International Association of Oil an d Gas Producers (IOGP) produced a

334 reduced - that the increased probab ility of a release of less hazardous materials presents a lower risk than a lower probability of release of a higher hazard material. Another example of potential shifts, rather than reductions in risk is the question of whether converting from chlorine gas to bleach shifts risk from the population around th e water treatment facility to the facility producing the bleach. If the bl each supplier also supplies chlorine gas by taking large quantities of chlorine and repack aging them into smaller containers, then the facility may be able to readjust the amount of chlorine from repackaging to bl each production. However, a bleach supplier that does not repackage chlo rine may be required to increase the amount of elemental chlorine used at that facility in order to meet the increased demand for bleach. If th e bleach supplier is also in a more densely populated area, the increase d chlorine needed at the facility could increase the risk to that new population. Again, the question of whether overall risk is reduced or shifted will depend on the specifics of an individual water treatment plant, as well as the specifics of the treatment plant’s supplier (Ref 13.26 Overton). 13.5 INHERENT SAFETY AND ECONOMIC CONFLICTS 13.5.1 Existing plants – operationa l vs. re-investment economics in a capital-intensive industry The following example illustrates the selection of an inherently safer design solution for an existing process. The design problem was to avoid a significant leak in several water- cooled heat exchangers. These exchan gers used material on the process side that reacted violently with wate r, producing corrosive and toxic by- products. Alternative solutions cons idered included combinations of passive (double tube sheet or falling film exchangers), active (multiple sensor leak detection with automa ted isolation), and procedural (a variety of nondestructive testing/ in spection techniques, periodic leak testing with inert gas, improved clea ning procedures) strategies. All of these design alternatives resulted in a lower risk level than the original design. However, none was totally acceptable (see Table 13.3). When management studied the effort and commitment of resources necessary to maintain a less than sati sfactory risk level, they chose a

LESSONS LEARNED 351 Examples of old newsletters that still convey highly relevant learning T are shown in the ICI newsletter (Kletz, IChemE website), in Figure 16.3 and Figure 16.4. Figure 16.3 ICI Safety Newsletter No. 96/ 1 & 2

TOOLS AND METHODS FOR MANAGING ABNORMAL SITUATIONS 125 Table 5.4 Policies and Administrative Procedures Common Tools and Methods Strengths Weaknesses Organizational chain of command, hierarchy Defines responsibilities and authorities Limitations of authority can create problems in abnormal situations when critical decisions must be made quickly. At large facilities, area management can devolve into ‘kingdoms’ that result in inconsistent standards across the site. Communications between shifts – verbal, written logbooks and electronic Provides seamless link between shifts so that transient conditions are managed. In practice, the quality of shift change communications is highly variable, and requires supervisor monitoring. Auditing of conformance to policies and administrative procedures. Can detect gradual degradation of systems and behaviors that may not be apparent to people working day-to-day. Can also detect specific faults that would aggravate an abnormal situation. Findings are a snapshot in time so that failings that occur between audits may be present for a year or more before being detected. Process Metrics Key Performance Indicators (KPIs) can be established to monitor the alarm system data as well as process parameters. Excellent for checking current process conditions against target parameters. Similar metrics may apply across multiple processes. May provide a superficial view of “symptoms” rather than underlying faults.

31 Independent protection layers (IPL ) that meet the prerequisites of independence, effectiveness, an d auditability may be credited qualitatively or semi-quantitatively in formal Layers of Protection Analysis (LOPA) to determine if the co llective protective features (i.e., the layers) are adequate to reduce the ri sk of an undesired hazard scenario to a tolerable level. These are generally the most reliable and robust layers and may include safety instru mented systems (SIS), basic process control systems (BPCS), relief devices, operator response to alarms, certain mitigation systems, and certain key design features related to preventing process safety incident s. Creating multiple layers of protection, and those that qualify as IPLs, between a hazard and potentially impacted people, prop erty and the environment can be highly effective. Their application ha s significantly improved the safety and process safety perf ormance of the chemical /processing industry. However, such an approach may have significant disadvantages: The basic hazard(s) remains, and some combination of failures of the layers of protection may re sult in an incident, thereby allowing the consequences to be re alized. Every active or passive layer has a certain likelihood of fa ilure, due either to mechanical means or management systems failures, such as not maintaining or keeping administ rative controls active. The outcome of the event may be limited by whatever passive or inherent layers have been applied. Potential impacts could be realized by some unanticipated route or mechanism . Hazardous event can occur by means beyond what were anticipated by process sa fety engineers. Accidents can occur by mechanisms that were unanticipated or that were poorly understood. Complicated and overlapping layers of protection increase the possib ility of unanticipated failure routes, particularly when common cause failures are shared by some of the layers, i.e., they ar e not independent. Therefore, the actual risk may be the same or even increase after the application of additional layers of protection. For example, a complicated shutdown system may cause an inadvertent sudden shutdown, which presents an overpressure, leading to a release.

136 | 4 Applying the Core Pr inciples of Process Safety Culture proactive approach to process safety as well as other environment, health, safety, security, and quality, and encourages collaboration with regulators. This can help strengthen PSMSs and contribute towards establishing the imperative for process safety . Various country and regional programs led by regulators such as OSHA VPP (USA), Safer Together (Australia), and Step Change in Safety (UK) seek sim ilar goals. The threat of a routine regulatory inspection is generally not an incentive to im prove culture or PSM S perform ance. In general, regulatory agency staffing levels are rarely sufficient to put teeth in such a threat. Som e agencies have been trying to change this by focusing on only 1-2 PSMS elem ents and certain sub-sectors, such as the National Em phasis Program s (NEPs) used in recent years by OSHA in the USA. Facilities leaders should take care to prevent regulatory focus on just a few elem ents from leading to normalization of deviance or loss of the imperative for process safety in the other elements. While some process safety regulators around the world are them selves process safety experts, the majority are not. Their backgrounds m ay be in occupational safety, environm ental sciences, or sim ilar disciplines that enable them to interpret regulations and understand management system s. In other words, regulators will generally not conduct in-depth technical analysis, but they will understand and evaluate management system performance. They will also be sensitive to cultures that do not take management systems perform ance seriously. Regulators will certainly come to the plant following a m ajor release incident. In such cases, regulators will generally have one or more regulatory findings. Having a collaborative relationship with the regulator while demonstrating a strong culture will help limit findings by keeping the regulators’ focus on the relevant

102 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION that jurisdiction. These codes are a good source of knowledge addressing fire protection and suppression. Of note are the following. (NFPA) NFPA 491M Manual of Hazard ous Chemical Reactions NFPA43 B Storage of Organic Peroxide Formulations NFPA 49 Hazardous Chemicals Data NFPA 325 Fire Hazard Properties of Flamma ble Liquids, Gases, and Volatile Solids NFPA 430 Storage of Liquid and Solid Oxidizers Summary Reactive chemical hazards can occur in a variety of industries and types of equipment. Incidents occur when the reactive hazard is unk nown or when it is underestimated. Wherever chemicals are handled and processed, even where it is not understood that reactive chemistry can occur, the chemical properties should be understood, hazards screened, and safeguards implemented. Do not assume that the chemicals can be co ntrolled in all operating conditions at the facility. If the data are not available to verify this, then conduct testing to gain the data. Many sources are available to gather the chemical property data and useful tools such as the Chemical Reactivity Worksheet can be used to analyze the hazards. Other incidents Other incidents involving reactive chemicals include the following. Rohm & Haas Road Tanker Ex plosion, Teeside, U.K., 1976 Arco Channelview Explosio n, Texas, U.S., 1990 Hickson Welsh Jet Fire, Yorkshire, U.K., 1992 Hoechst Griesheim, Explosion, Frankfurt, Germany, 1993 Port Neal AN Explosion, Sioux City, Iowa, U.S., 1994 Napp Technologies Explosion, Lodi, New Jersey, U.S., 1995 Bartlo Packaging Pesticide Explosion, West Helena, Arkansas, U.S., 1997 Morton International Explosion, Paterson, New Jersey, U.S., 1998 Concept Sciences Hydroxylamine Explosio n, Allentown, Pennsylvania, U.S., 1999 AZF AN Explosion, Toulouse, France, 2001 Synthron Chemcial Explosion, Morg anton, North Carolina, U.S., 2006 Bayer CropScience Runaway Reaction and Pressure Vessel Explosion, Institute, West Virginia, U.S., 2008 West Fertilizer AN Explosion, West, Texas, U.S., 2013 Tianjin AN Explosion, China, 2015 Arkema Fire, Crosby, Texas, U.S., 2017 Seveso Disaster, Seveso, Italy, 1976 Port of Beirut Ammonium Nitr ate Explosion, Lebanon, 2020

APPENDIX A – CONCLUDING EXERCISES 469 8. What inherently safer design options might be considered for this project? 9. Name three failures that might occur with this equipment. 10. A consequence analysis is to be performed. List 3 potential scenarios including source, transport, consequence effects, and potential outcomes. 11. Draw a swiss cheese diagram for one of the scenarios identified in the HAZOP. 12. Suggest how human factors could be considered in conducting the HAZOP. 13. List 5 things you expect to be on the operational readiness plan for this project. 14. As the project is 50% through the detailed engin eering, a proposal is made to increase the reactor size. How should this be handled? 15. List three operating practices and three safe work practices that would be appropriate for this facility when it is operational. 16. List 3 emergencies should be addressed in th e Emergency Response Plan for this facility. 17. List 2 means to engage the workforce in the pr oject. List 2 stakeholder groups that should be involved in the project. 18. List 3 leading and 3 lagging process safety metr ics that might be appropriate for this facility when it is operational. 19. What action might you take to foster a g ood process safety culture on the project? Exercise 3: Ethylene Buffer Tank An operator is preparing an outdoor ethylene buffer tank for maintenance by evacuating the vessel of ethylene to an acceptable level. While lining up the vessel vent line to a flare header, an ethylene release to atmosphere occurred du e to a ¾” bleed valve being inadvertently left open. When you arrive at work, your boss has several questions regarding the incident, and has given you 80 minutes to give him the answers. The first thing you do is gather the process safety information related to the incident from the Cameo database, MS DS, and CRW. This is given in Table A.2.

KEY RELEVANCE TO PROCESS PLANT OPERATIONS 79 3.4.3.2 Issues Associated with Various Chemical Phases Several major industry events have occurred because of personnel not understanding or addressing that an abnormal situation is developing at the interface between different phases of a chemical mixture. Most often this is simply a case of a faulty instru ment (e.g., level transmitter failure). However, in other cases, the instrume nt may be performing correctly but the resulting output is not reflecti ve of the actual situation due to transient conditions. Examples of ea ch are provided in Example Incident 3.18 and Example Incident 3.19. Example Incident 3.18 - Unre liable Interface Detector In a hydrofluoric acid (HF) alkyla tion unit, the interface detector between the HF and hydrocarbon phas es was unreliable, on occasion reading zero level of the hydrocar bon phase when the phase was, in fact, present. One evening this lo ss of level on the instrument was observed as usual but dismissed by the control panel operator because of the prior history of faulty readings. A f t e r a n h o u r o f t h i s o p e r a t i o n , the supervisor noted several other atypical issues in the unit – reduct ion in feed consumption, increase in waste product, strange separator readings, etc. Since these were also symptoms of a zero-level condition, he concluded that the loss of level might be real, and that the unit was actually in an ‘acid runaway’ condition. The fix for such a situation was to stop the feed to the unit and regroup. However, since neit her the supervisor nor the area supervisor had authority to shut th e flow, they deferred to the overall plant night shift supervisor, whose de cision was to wait for four hours and, if the level was still reading ze ro, then initiate a shutdown. About six hours after the initial zero level indication, the unit was finally shut down, but because the shutdown ha d been delayed th e restart took three days to accomplish vs. about on e hour if it had been initiated immediately. Because storage for the intermediate feedstock had not yet been commissioned, about 25,000 ba rrels of LPG had to be flared.

INTRODUCTION AND REGULATORY OVERVIEW 13 OSHA, U.S. Occupational Safety and Health Administration, https://www.osha.gov/laws- regs/regulations/standardnumber/1910/1910.119. Zerbonia 2001, Robert A. Zerbonia , Cybele M. Brockmann , Paul R. Peterson & Denise Housley, “Carbon Bed Fires and the Use of Ca rbon Canisters for Air Emissions Control on Fixed-Roof Tanks”, Journal of the Air & Waste Management Association , 51:12, 1617-1627, DOI: 10.1080/10473289.2001.10464393.

138 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Nature can pose meteorological and geological hazards . (CCPS 2019) Meteorological hazards are those that naturally occur due to the weather cycle or climactic cycles, and include flooding, temperature extremes, snow/ice storms, wildfire, tornado, tropical cyclones, hurricanes, storm surge, wind, lightning, hailstorms, drought, etc. Geological hazards are those occurring due to the movements of the earth and the internal earth forces, and include seismic ev ents, earthquakes, landslides, sinkholes, tsunami, volcanic eruptions, and dam rupture. Natural Hazards Triggering Technological Disa sters (Natech) refers to the interaction between natural disasters and industrial accidents. More information on this area of research can be found at the UN Economics Commission for Europe. (UNECE) Identifying these natural hazards and including their potential impact in design and emergency response preparedness plans is important to prevent their resulting in a process incident. For example, natural hazards can result in loss of access and power to facilities which can, in turn, result in loss of cooling to reactive chemical storage. This was evident in the Arkema organic peroxide decomposition and fire following Hurricane Harvey flooding in 2017. (CSB 2018) Earthquakes can result in failure of equipment and piping resulting in fires. Mining operations often create retention basins or dams to hold tailings. In Hungary, the Kolontar Tailings dam failure released a wave of bauxite into the surrounding area. (AGU) Tailing dam failures can result in both rapid water current and flooding hazards as well as longer term toxic exposures. Guidance is availa ble such as the Mining Association of Canada’s Guide to the Management of Tailings Facilities. (MAC) Seismic hazards for process fac ilities include both potential toppling of tall structures and also the resulting escalation of the initial event. In seismic regions a seismic evaluation should be conducted of tall structures, such as distillati on columns, to determine if actions should be taken to prevent damage due to a seismic even t. Additional design considerations may need to be incorporated into facilities located in seismic zones. Natural hazards can also result in second ary process safety impacts. Consider the eruption of the Iceland volcano that resulted in a shutdown of air travel in Europe as shown in Figure 8.5. (NASA) This meant that work to evaluate hazards, analyze risks, and implement systems to control process safety risks were pu t on hold until travel, and transportation of supplies, could resume.

89 5 SELECTING AN APPROPRIATE PHA REVALIDATION APPROACH As stated in Chapter 1, the primary goal of a revalidation is to verify that the PHA document accurately describes the current risk profile of the subject process unit. While a PHA revalidation can sometimes be accomplished with less effort and in less time than the initial PHA, this is highly dependent upon the scope and quality of the prior PHA and thoughtful planning. The preparation described in Chapters 2 through 4 of this book, and illustrated in Figure 5-1, were all intended to help the reader answer the following pivotal question: Which PHA revalidation approach will be the most resource- efficient and effective way to produce a complete, compliant, up-to-date, and thoroughly documented PHA? Figure 5-1 Revalidation Flowchart – Selecting the Approach

APPLICATION OF PROCESS SAFETY TO WELLS 67 4.2.3 Shallow Gas Risks Shallow gas deposits near to the surface ca n be encountered during drilling before the BOP and surface casing are in place and can lead to a shallow gas well control incident. This can be associated with eith er onshore or offshore drilling. Oil is not normally present with shallow gas. While m ud weight can be increased, if this fails then it may be necessary to drill a relief well to kill the shallow gas flow. During a shallow gas incident, it is not normally advised to try to shut in the well as the surface formation is not strong en ough to provide for containment. A safer option is achieved using a diverter valv e to direct flow away from the rig floor. Key Process Safety Measure(s) Hazard Identification and Risk Analysis : Identification of shallow gas is key to understanding and managing the risks. Shallow gas is hard to detect, but newer digitally enhanced seismic analysis can re veal this hazard. Consequences onshore are primarily related to flammable and potentially toxic gases. Offshore consequences are similar as shallow gas can bubble to the surface under a floating drilling rig and create a flammable atmosphere . It can also damage the sea floor and destabilize a jack-up rig. A shallow gas incident can damage or rupture the drill string and thus reduce the ability to deliver a heavier mud to the problem zone. Personnel require evacuation which can be difficult due to the flammable atmosphere but, as seen in the Snorre A blowout described in the following incident description, can be done successfully. Shallow water hazards are similar but without the flammable or sour gas hazards. They are thus more of an oper ational than process safety problem. However, depending on the source of the water and if there are nearby receptors, there can be a pollution risk. For example, if the water source is due to accumulation of nearby water injection wells, then the water may be contaminated. 4.2.4 High Pressure High Temperature (HPHT) Wells Risks Loss of well control risks are heightened when well construction involves high pressure, high temperature (HPHT) reserv oirs. These have temperatures exceeding 300 ˚F, a pore pressure of at least 0.8 psi/ft, or requiring pressure control equipment exceeding 10,000 psi. Drilling, comple tions, workovers, interventions and abandoning wells in HPHT environments ar e at greater risk due to the complexity associated with the high pressure and high temperature and having a higher probability of well control incidents and equipment failures.

74 INVESTIGATING PROCESS SAFETY INCIDENTS To ensure continuous improvement, an evaluation after each investigation should include: • Team thoroughness in the investigation. • Team effectiveness in applying the techniques. • Team preparedness in advance of the investigation. • Equipment performance during the investigation. • Supply logistics and quality. To ensure that the management system cont inues to provide the intended results, periodic reviews and updates are necessary. This action recognizes that organizations are dy namic, ever-changing, and evolving. Consider the following critique questions. Were the investigation techniques applied correctly and fully? Did the team accurately determine what happened? Did the team find the management system failures that led to the in cident (that is, did they get to root causes)? Was the team documentation adequate? Were the right skills available within the team? What other resources could be used next time? What should be changed next time? Is there evidence to suggest that near-misses are being reported? Have there been any repeat events? Chapter 14 provides guidance on recommendation implementation effectiveness, and Chapter 15 details proven methods for enhancing an incident investigation system. 4.3 M ANAGEM ENT SYSTEM Implementing a new or upgraded management system normally begins with training employees, supervision, and management in their respective roles in the investigation program. Implem entation also includes development and refinement of the incident da ta management systems. The data management system should allow users to easily dev elop consistent reports

204 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS stability tests that were reviewed in 1988 and 1989 when the process of batch distillation in 60 Still Base was introduced. After the modification in 1988, problems started with residue accumulation in parts of the cont inuous distillation section of the process. These were stripped and cl eaned out although operators also noted the accumulation of sludge in 60 Still Base, up to a depth of some 34cm (14 inches). In June 1992, one of the senior process technologists expressed frustration about these problems an d the impaired conditions on the continuous MNT distillation section of the process. A memo was written stating: “It is my view that we are wi thin five years of a major accident on the MNT distillation system.” At some point (date unknown), the steam regulator on the supply to the heater battery became non-oper ational. This was overcome by someone opening a bypass valve aro und the regulator until the relief valve started to lift. On 10 September 1992, some 11 days before the incident, two of the 40% whizzer oil tanks were fully emptie d to the 60 Still Base, in order to clean them out for a change of prod uct. Being a vacuum vessel, this allowed the residual sludge from thes e tanks to be sucked into 60 Still Base, further increasing the level of solids in the vessel. On Thursday, 17 September, the removal of the slud ge from 60 Still Base was discussed by a shift manager and area manager. Since this vessel had never been cleaned out before, they discussed some of the practical measures required including removal of some st eps and provision of a skip (steel dumpster) to collect the sludge. A batch distillation of the material in the 60 Still Base then took place on Saturday 19 September in prepar ation for the work to start the cleanout of 60 Still Base the fo llowing Monday, 21 September 1992. 7.3.4.2 Organization In parallel with the process changes, several organizational changes took place at the Castleford site. The original management structure was a traditional, hierarchical one, wher e plant managers managed individual plants and each shift comprised a shift supervisor and a small number of shift operators.

163 Some batch reactions have the po tential for generating very high energy levels. If all of the reactant s (and catalysts, if applicable) are charged into a reactor before the reac tion is initiated, and two or more of the materials in the reactor reac t exothermically, a runaway reaction may result. The use of continuous or “semi-batch” reactors to limit the energy present and to reduce the risk of a runaway reaction should be considered. The term “semi-batch” refers to a system where one reactant and, if necessary, a cataly st is initially charged to a batch reactor. A second reactant is subs equently fed to the reactor under conditions such that an upset in the reacting conditions can be detected and the flow of the second reactant stopped, thus limiting the total amount of potential energy generated in the reactor. Additional discussion regarding reac tor design strategies is covered in Section 3.2 on Minimization (as an inherently safer design strategy), and in Section 8.4 on the process design. Distillation . There are options to minimize hazards when distilling materials that may be thermally unstable or have a tendency to react with other chemicals presen t. These options include: Trays without outlet weirs Sieve trays Wiped film evaporators An internal baffle in the base section to minimize hold-up Reduced base diameter (Ref 8.52 Kletz 1991) Vacuum distillation to lower temperatures Smaller reflux accumulators an d reboilers (Ref 8.30 Dale) Internal reflux condensers and reboilers (Ref 8.30 Dale) Column internals that minimize holdup without sacrificing operational efficiency (Ref 8.30 Dale) Another option is to remove toxic, corrosive or otherwise hazardous materials early in a distillation sequ ence, reducing the spread of such materials throughout a process (Ref 8.82 Wells). Low-inventory distillation equipment, such as the thin film evaporator, is also available and shou ld be considered for the distillation of hazardous materials. This equipme nt offers the additional advantage

ACRONYMS AND ABBREVIATIONS xxvii MCC Motor Control Center MIE Minimum Ignition Energy MOC Management of Change MOC Minimum Oxygen Concentration MOOC Management of Organizational Change NASA National Aeronautics and Space Administration NDT Nondestructive Testing NFPA National Fire Protection Association OD Operational Discipline OIMS Operational Integrity Management System (ExxonMobil) OSHA U.S. Occupational Safety and Health Administration PAC Protective Action Criteria PFD Process Flow Diagram PFD Probability of Failure on Demand PHA Process Hazard Analysis P&ID Piping and Instrumentation Diagram PLC Programmable Logic Controller PRA Probabilistic Risk Assessment PRD Pressure Relief Device PRV Pressure Relief Valve PSE Process Safety Event PSI Process Safety Information PSI Process Safety Incident PSM Process Safety Management PSO Process Safety Officer PSSR Pre-Startup Safety Review QRA Quantitative Risk Analysis RAGAGEP Recognized and Generally Accepted Good Engineering Practice RBPS Risk Based Process Safety RMP Risk Management Plan

39 potentially affected population. A technology may be inherently safer than another with respect to some ha zards but inherently less safe with respect to others and may not be safe enough to meet societal expectations. IST options can be location and release scenario dependent, and different potentia lly exposed populations may not agree on the relative inherent safety characteristics of the same set of options. ISTs are based on an informed deci sion process: Because an option may be inherently safer with respec t to some hazards and inherently less safe compared to ot hers, decisions about the optimum strategy for managing risks from all hazards are re quired. The decision process must consider the entire life cycle, the fu ll spectrum of hazards and risks, and the potential for transfer of risk from one impacted population to another. Technical and economic feas ibility of options must also be considered. Risk reduction criteria w ill be determined by the nature of the hazards or threats and will requir e consideration of conflicts among multiple hazards and threats. Inherently safer options should also be considered for the entire supply chain, including manufacturing, use, storage, transportation, and disposal. Tradeoffs are involved in terms of moving risk from one location in the supply chain to another (Ref 2.6 Berger, Ref 2.2 ACS). For example, reducing onsite inventor y of a hazardous raw material may require more frequent shipments and subsequent risk along the transportation route, or increased inventory at the supplier location. Marshall (Ref 2.23 Marshall 1990, Ref 2.24 Marshall 1992) discusses accident prevention, control of occu pational disease, and environmental protection in terms of strategic and tactical approaches. Strategic approaches have broad significance and represent “once and for all” decisions. The inherent and passive categories of risk management would usually be classified as strate gic approaches. In general, strategic approaches are best implemented at an early stage in the process or plant design. Tactical approaches, wh ich include active and procedural risk management categories, tend to be implemented much later in the plant design process, or even after the plant is operating, and often involve much repetition, increasing co sts, as well as the potential for failure. However, it is never too late to consider inherently safer alternatives. Major enhancements to inherent safety have also been

8.1 Flixborough, North Lincolnshire, UK, 1974 | 107 Flixborough taught us the importance of formally, thoroughly, and consistently managing process changes, including changes that are considered temporary or emergency. The engineers at Flixborough did evaluate the changes to reactor train throughput that would be required to operate with one less reactor. However, they did not evaluate whether the planned temporary bypass piping had been properly designed to support its weight, handle the vibrations resulting from two angle bends, and endure the expected thermal stresses (Figure 8.1). After two months of exposure to stress, vibration, and fatigue, the piping failed, releasing a large cloud of flammable vapors that ignited and exploded. Figure 8.1 Temporary Bypass on Flixborough Reactor 5 (Source: UKDOE 1975) Most of us stop our consideration of Flixborough there. But we can learn a lot more from this incident. The jacket of reactor 5 also failed due to a stress corrosion crack and thermal stress. The corrosion resulted from a temporary change to spray process water on the reactor head. Nitrates were added to boost heat transfer capacity. The nitrates eventually led to stress corrosion cracking of the jacket, however (Figure 8.2). Today, as then, incidents happen when companies apply temporary fixes rather than having the operational discipline to shut Figure 8.2 Stress Corrosion Crack on Flixborough Reactor 5 (Source: UKDOE 1975)

2 INVESTIGATING PROCESS SAFETY INCIDENTS The first edition provided a timely treatment of incident investigation including: • a detailed examination of the role of incident investigation in a process safety management system, • guidance on implementing an incident investigation system, and • in-depth information on conducting incident investigations, including the tools and techniques mo st useful in understanding the underlying causes. The second edition, released in 2003, built on the first text’s solid foundation. The goal was to retain the know ledge base provided in the original book while simultaneously updating and expanding upon it to reflect the latest thinking. That edition pres ented techniques used by the world’s leading practitioners in the science of process safety incident investigation. This third edition is a further enhancement of the second edition. Specific emphasis has been placed on updating investigation techniques and analytical methodologies, and applying them to example case studies where possible. Expanded topics include scientific va lidation of hypotheses, rigorous physical evidence documentation and examination, scientific analysis, hypothesis rejection and substantiation, learnings from repeat incidents, and means to institutionalize learnings within an organization. 1.2 INVESTIGATION BASICS Successful investigations are dependent on prep lanning, documented procedures, appropriate investigator training and experience, appropriate support from leadership, and necessary resources (personnel, time, and materials), to conduct a thorough investigation. It is imperative that operating organizations conduct careful and comprehensive investigations that are factual and defensible. Develo ping and following written procedures allows organizations to consistently respond promptly and effectively, establishes the basis for continuous improvement, and helps preserve a company’s “license to operate”. 1.2.1 The First Step in conducting a successful incident investigation is to recognize when an incident has occurred so that an Incident Management System (Chapter 4) can be activated. Linked with incident recognition are Initial Notification, Classification, and Investigation (Chapter 5).

160 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS Example Incident 6.1 – Fire Protection System Found Disabled During a field review of safety sy stems in one unit, the auditor noted that the pumps were provided wi th a sprinkler system that was intended to provide a rapid response to a local fire. Curious as to the source of this water, the auditor followed the supply piping back to a manifold. This manifold had several va lved lines tying into it, so that it was not immediately clear which va lve would lead to which user. Furthermore, a painter had left a thic k layer of paint on all the valves, so that a significant effort would be needed to operate the valves in an emergency. Beyond that, the au ditor traced the source of the manifold’s water to yet another manifold, where those valves were blocked in. The condition and position of the two sets of valves were in themselves not an emergency si tuation; however, they were important to the function of a mi tigation system to address an emergency. In the event of an ac tual fire, the response would have been delayed and ineffective. Lessons learned in relation to abnormal situation management: Understanding abnormal situations: A lack of knowledge about the positioning of the block valves, their readiness, and functionality. The valving was abnormal, yet the consequences were not considered. Process Monitoring: Continually mo nitor the readiness of safety systems or include confirmation that a required periodic checklist has been performed. Car Seal Management: The utilization of a car seal open and car seal closed system to ensure valves are maintained in their preferred position is a positive way to minimize valving errors.

EQUIPMENT FAILURE 205 Figure 11.20. Horizontal peeler centrifuge with clean-in-place system and discharge chute (Patnaik) Figure 11.21. Cross sectional view of a continuous pusher centrifuge (Patnaik)

DEVELOPING EFFECTIVE RECOM M ENDATIONS 285 The incident investigation team should consider including recommendations that examine inhere ntly safer design. Changes can be either beneficial or detrimental, so investigators should be alert for features in recommendations that are inherently less safe. Two common examples of design changes that can increase overall risk are the use of flexible joints and the use of glass (rotameters, bulls eyes, sight gl asses, or additional control room windows) (Englund, 1991). Seal-less pumps are generally considered to be inherently safer than pumps with mechanical se als. The failure mode(s) of any recommended new equipment should be carefully considered before a decision is made to implement the change. 12.3.2 Layers of Protection The concept of multiple layers of protection (barriers) has widespread support throughout the refining and chemical processing industry. By providing sufficient layers of protection against an accident scenario, the potential risk associated with that accident can be avoided or at least reduced. For a given scenario, only one barrier must work successfully for the consequence to be prevented. However, since no single barrier is perfectly reliable, multiple layers of protection are often provided to render the risk of the incident tole rable. It should be under stood that these multiple layers of protection are fully indepen dent; otherwise, there could be fewer barriers than expected. This is illust rated in Chapter 2, where the “Swiss Cheese Model” is discussed. The failure of one or more barriers mi ght be identified as part of an incident investigation. Recommendations arising from an individual barrier failure can be made at various levels . Trevor Kletz said that accident investigation was like peeling an onion: “The outer layers deal with the immediate technical causes while the inner layers are concerned with ways of avoiding the hazards and with the und erlying causes, such as weakness in the management system.” He identified three layers of recommendations, as follows: (Kletz, 1988) • First layer remedies use immediate technical recommendations targeted to prevent a particular incident. Consider the case where an employee is injured by inhalation exposure while taking a liquid chlorine process sample. Firs t-layer recommendations would address such items as changes to the sampling procedure, refresher training, and selection and use of personal protective respiratory equipment. • Second layer recommendations focus on avoiding the hazard. A deeper and broader perspective is used for this second layer, and

238 INVESTIGATING PROCESS SAFETY INCIDENTS Figure 10.21 Exit Piping Crack Branch What if the team was not able to obtain any physical evidence? They could use the absence of any corrosion inspection records plus knowledge of the expected corrosion (i nternal and external) of the system as an indicator of whether corrosion was a credible possibility. With no evidence at all, the team might develop each hypothesis as a separate branch of the tree and try to address potential causes of corrosion, improper choice of materials, flange failure, or other items. After collecting and analyzing the available evidence, the incident investigation team constructed the logic tree diagrams shown in Appendix D. These diagrams present, in a logical and systematic format, the sequence of events and conditions that ultimately resulted in the major incident. The simplified qualitative fault-tree indicates various events and conditions that

Appendix B - Major accident case studies 385 B.2 Bayer Crop Science plant explosion in West Virginia, U.S. In 2008, a large explosion led to fatality of two workers at the Bayer Crop Science plant in West Virginia, USA [26]. The fire burned for more than four hours. Two contractors and six firefighters were trea ted for possible toxic exposure [83]. The damaged plant is shown in Figure B-2. ‘What happened’ is summarized after Figure B-2. A thermal runaway reaction (a chemical reaction) occurred inside a 4,500 gallon (17,000 liter) pressurized residue treater, ca using it to fracture. Highly flammable solvent sprayed from the vessel and ignited, causing fire. Figure B-2 Bayer Crop Science plant damage (reproduced from www.csb.gov) The incident happened during the first methomyl restart after an extended outage to install a new process control system and a stainless-steel pressure vessel. The steps leading to this accident are outlined next: • A methomyl unit was due to be restarted after replacing the control system and the residue treater vessel. • Prior to start-up, the vessel should have been loaded with solvent and the solvent preheated. Neither of these actions were done. • At 04:00 the outside operator manually opened a feed valve to start filling the residue treater vessel with methomyl. The methomyl should have been instantly mixed with the missing solvent. • With a low flow rate, more than 24 hours is required to fill the residue treater to the normal operating level (50%). During this time the mix was

PHA Revalidation Requirements 27 would be aware of every possible RAGAGEP. Hence, as mentioned in Chapter 1, organizations (e.g., companies, trade associations, and professional societies) often develop checklists to remind people of key points to verify during original designs, PHAs, and revalidations. An excellent example of this is Bulletin 109, published by the International Institute of Ammonia Refrigeration for the mutual benefit of its membership [28]. It comp iles safety criteria for a variety of equipment (e.g., compressors, condense rs, evaporators, and piping) normally used in refrigeration systems. The foll owing paragraphs briefly describe some common RAGAGEPs that PHA reva lidation teams might consider: International Electrotechnical Commission (IEC) 61508, “Standard for Functional Safety of Electrical/Elect ronic/Programmable Electronic Safety- Related Systems” [29]. This standard co vers safety-related systems, which may include anything from simple actuated valves, relays, and switches up to complex programmable logic controllers. The standard specifically addresses safety instrumented systems and how their failure to execute their safety functions either on demand or conti nuously can result in safety-related consequences. The overall program to en sure system reliability is defined as “functional safety.” IEC 61511 is the proc ess industry derivative of IEC 61508. Risk judgments in a PHA of an automated process may involve supplemental risk assessments (e.g., LOPA) that are cont ingent upon the reliability of safety systems covered by this standard; in these cases, it is essential to have a revalidation team member(s) familiar with its guidance. American National Standards Institute/National Fire Protection Associ- ation (ANSI/NFPA) 70, “National Electric Code” [30]. As its title indicates, this code applies to a wide range of electr ical power systems and their components. Various chapters address everything fr om basic definitions and rules for installations (e.g., voltages, connections , markings, and circuit protection) to general-purpose equipment (e.g., conducto rs, cables, receptacles, and switches) to special equipment and conditions (e.g ., signs, alarms, emergency systems, and communications). Articles 500 to 506 are of particular interest to revalidation teams trying to ensure th at process changes have not created or altered areas with potentially explosiv e concentrations of dusts or vapors. European Directive 94/9/EC, commonly referred to as the Atmospheres Explosible (ATEX) Directive, similarl y addresses equipment in potentially explosive atmospheres. NFPA 652, “Standard on the Fundamentals of Combustible Dust” [15]. Dust fires and explosions have resulted in fatalit ies and extensive property damage. Yet the materials involved may be as seemin gly innocuous as sugar or polyethylene, so they are not classified as highly hazardous materials. Dust hazard analyses

Appendices 193 Q T R XII. Electrical Classification Is there an electrical classification document? Does the electrical classification appear correct and complete? Has the electrical classification do cument been recently revised? Have significant changes made si nce the system was originally constructed (addition of new mate rials, new sources of flammable gases or vapors, new low points [e.g., sumps or trenches] at grade) been included in the electrical classification document? Are the design and maintenance of ventilation systems adequate? Are there safeguards to alert operators when a ventilation system fails? Are ventilation systems being properly maintained, and are alarms and interlocks on these systems periodically function checked? Is adequate maintenance being do ne to function check natural ventilation systems? Are there technical bases for design changes to the ventilation systems? Are ventilation systems verified to be adequate for new gas or vapor loads? Are there adequate controls to ensu re that electrically classified equipment is replaced with equipment of equal or higher classification? Are boundaries between electrically classified areas physical boundaries? Are Division 1 areas necessary (if there are any)? Are there adequate controls (e.g., a hot work permit system) on repair and construction activities , including work by contractor personnel? Are there specific requirements for pe rsonnel in classified areas (e.g., static dissipative attire, prohibition of ignition sources such as cell phones)? Does the electrical classification adequately reflect the effects of different modes of operation (e.g., normal operation, maintenance, startup, infrequent operating modes such as reactor regeneration or operation with a portion of the system bypassed)? XIII. Contingency Planning What expansion or modification pl ans are there for the facility? Can the unit be built and maintained without lifting heavy items over operating equipment and piping?

Pipes 97 pipe around the trimmer unit. This bypass pipe can be used during the start‐up of the plant when there is no real need to bring all the units in operation at once, and the trimmer unit can be bypassed (Figure 6.73). 6.12.4 Piping f or Units in Parallel Units can be placed in parallel, and their piping is a bit tricky. Here we want to focus on the pipe sizes in such arrangement. It is obvious that if there are two similar equipment in parallel and one of them is the operating piece of equip-ment and the other, a spare one (it means 2 × 100% s paring philosophy), the pipe size does not change after splitting. This concept is shown in Figure 6.74. But if both pumps are operating (it means 2 × 50% spar ing philosophy), the flow splitting on the inlet side of the pipes in a way that each pump will receive half of the flow. Here the rule of thumb says that the size of each branch is 2 of the size of main header, as shown in Figure 6.75. Based on this, when a pipe flow is branched to three even flows, the size of each branch would be 3 of the size of main header. 6.12.5 Piping f or Pressure Equalization The pipe for pressure transfer or pressure equalization can be much smaller than the main pipe size (i.e. two or three size smaller). The pipe for pressure equalization can be connected to two sides of blocking valves for ease of opening or between two tanks to allow initiating the liquid flow between them.6.13 Pipe Size Rule of Thumbs It is not easy to tell from a P&ID whether the pipe size is correct. However, when two or more pipes are con-nected to each other, it is easier to check the accuracy of pipe sizing. Below are some cases through which the pipe size can be checked: 1) When two (or mor e) pipes are merging together, the resultant pipe may have a larger size. Here the word may is used because there are some cases that this rule is not valid. For example, when a 2″ pipe is merged to a 20″ pipe, the size of pipe after connection of the 2″ pipe is less likely to be changed to a larger size, like 22″. 2) When a pipe i s split into two or more branches, the size of branches may be smaller than the main pipe. 6.14 Pipe Appurtenances Pipe appurtenances are mainly classified into three main groups of valves: fittings, process items, and non-process items. Valves are piping components that actively affect flows. Active means they should have a movable part. A gate valve has a moving stem. A check valve has a moving flap. Fittings are piping appurtenances that passively affect flows. Examples of fittings are elbows and reducers. Process items are piping appurtenances that are installed on piping and change some process feature of the flow. Examples are strainers and silencers. Process items generally tagged in P&IDs as SP item.Recirculation pump(b) (a)Unit Unit Figure 6.72 (a, b) Unit r ecirculation pipe. Unit 1 Unit 2 Figure 6.73 Ser ies units pipes.Unit 1X/uni2033 X/uni2033 X/uni2033Unit 22×100% Figure 6.74 Pipe siz es of parallel units, a spare one operating. Unit 1X/uni2033.(2)0.5 X/uni2033.(2)0.5X/uni2033 Unit 22×50% Figure 6.75 Pipe siz es of parallel units, both operating.

131 10 REAL MODEL SCENARIO: LEAKING HOSES AND UNEXPECTED IMPACTS OF CHANGE “A [person] who carries a cat by the tail learns something [they] can learn in no other way.”—Mark Twain, Author and Humorist Feijoada Pharma produces a range of generic drugs in a modern facility outside São Paulo, Brazil. Each of its five primary reactor trains run one- to three- month campaigns to produce active pharmaceutical intermediates. The reactors are then reconfigured for the next product. Each reactor can be fed via multiple routes, including: • dip pipe • free-fall from a nozzle in the reactor head • free fall via one or more spray balls in the reactor head • through a recirculation line. These feed routes are accessed through four hose connections near the reactor that are grouped close together for convenience but are clearly labeled. The materials to be fed are similarly piped close to the reactor, terminating in hose connections. These, too, are grouped close together for convenience and clearly labeled. Operators connect the desired feed material to the desired feed port using the appropriate flexible hose designated in the operating procedure. Raw materials can also be fed to the reactor through any of the ports from drums. Some hoses are kept connected for the duration of the campaign, while others are purged and disconnected to allow a different raw material to be fed via a given feed route. 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

72 INVESTIGATING PROCESS SAFETY INCIDENTS apply lessons learned from an incident, no t only at a facility level, but also across the organization. Employees are affected by the recommendations. Their responsibilities include: Using new or modified equipment properly. Abiding by procedural improvements. Giving feedback to management wh en something is not working as expected. Sharing their knowledge when they find a better or safer way to address the problems identi fied in the investigation In summary, developing the recommendations is a responsibility of the incident investigation team. Accepting and implementing the recommendations is a management responsibility. The inclusion of the elements of the recommendation in dail y work practice is the responsibility of each individual affected by the recommended action. 4.2.11 Implementing the Recommendations and Follow-up Activities Resolving recommendations and followi ng up on their effectiveness is a cornerstone of all manage ment systems. Once a recommendation has been accepted for implementation, a clear, auditable document trail should be established and maintained. The recommendations should not only be implemented but also, they need to be su stained. For lastin g results, it is wise to audit implemented recommendat ions periodically to ensure that they are continuing to achieve the intended objectives. It is the prevailing opinion of many regulatory agencies that any changes in the originally accepted recommendation should be thoroughly documented. If a recommendation is modified in scope or time commitment, or is otherwise not implem ented as originally planned, then the basis for this decision should also be documented. The concept of an auditable trail is mentioned in regu latory and legal activities. If a recommendation is rejected or modified, the basis for the rejection or change should be thoroughly documented after review with the investigation team. These requirements should be reflected in the incident investigation management system and sh ould be emphasized wh en personnel at all levels are trained. The management system sh ould indicate the importance (priority) of the recommendation, assignment of responsibility, and method for verifying and

EQUIPMENT FAILURE 199 A B Figure 11.16.A. Example distillation column schematic (Bouck) Figure 11.16.B. Typical industrial distillation column (©Sulzer Chemtech Ltd.) Figure 11.17. Schematic of carbon bed adsorber system (OSHA a)

206 | 6 Where do you Start? broadly, so that all sites and units feel equally included in the process and change happens more quickly. Culture surveys should be perform ed anonym ously, and ideally by an independent party. At the outset of culture improvement efforts, m utual trust m ay not yet have been developed, so respondents m ay hesitate to give full open and frank input to assessors who represent com pany management. It m ay not be necessary to use an independent party for re-surveys of organizations where m utual trust is already high, but is still a good idea just in case there has been some slippage in the culture. Whoever conducts the surveys, anonymity should be preserved both in collecting and in reporting the data. When sub-segmenting the data, the number of people in a sub-segm ent should be large enough to prevent identifying individual respondents. While culture surveys typically produce narrative data, it is important for statistical analysis purposes to develop clear definitions m apping narrative input to num erical scores. This will likely require identifying a range of potential responses to every question, and decide how each would fit on the scale of potential responses, perhaps from 0% to 100%. In culture surveys, it is not unusual for some employees to respond very negatively to questions they do not really feel negative about. They may do this thinking they are punishing m anagement, or m ay wish to emphasize a negative response to other somewhat related questions. In some cases, the fear of m anagement reprisals m ay lead one member of the group to answer negatively on behalf of the rest, to deliver the message while saving their peers. Therefore, a m eans of interpreting extreme input is also needed. This may involve observing the work group with the apparently outlying feedback or asking members of the group why som e of their peers reacted as they did. Follow-up interviews m ay also be warranted. Of course, the negative input m ay also be

B.2 INHERENT SAFETY ANALYSIS - INDEPENDENT PROCESS HAZARD ANALYSIS (PHA) Table B.5 is an example of an IS Anal ysis approach, which is similar to a typical PHA, but focuses exclusively on inherent safety. The analyst considers a hazard, such as runaway reaction caused by water reactivity in a reactor and sets a safety objective, such as “minimize potential for runaway reaction in the feed to the reacto r.” The team then documents each potential cause of th e hazard being evaluated, reviews the consequences, existing safe guards, and potential means of eliminating it or reducing its risk through ISD strategies. Considering the four ISD strategies, the team documents potential recommendations that may address the concern using the order of First Order ISD, Second Order ISD, follo wed by Layers of Protection. Each strategy is considered. Ideas that are feasible, practical, and best address the hazard are generated. The approach acknowledges that other risk management strategies besides ISD may be more effective. 464

348 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION The mechanical design and control of proce ss equipment should maintain the process at the operating temperature and pressure. Alarms are typically provided to alert operators when a temperature, pressure, or level operating limit is being reached. The operators can then take action to bring the process back to normal op erating conditions. If the level continues to exceed limits, then an instrumented system ma y be provided to take action (e.g. shut/open valves) to prevent an unsafe condition being reached. These concepts were discussed in Section 10.5. A pressure relief valve is an active mitigation device that is often referred to as the last line of defense as it is intended to protect equipment should all other systems fail. If an overpressure occurs in a vessel and associated piping, pressure relief valves are designed to open at a predetermined pressure to allow the pressure to be relieved before the equipment fails. The three common types of relief valves are conventional, balanced bellows, and pilot operated. They have different operating charac teristics that make them appropriate for different operating conditions. Common to all is that they are designed to operate at a specified set pressure that is related to the MA WP and operating limits (refer to section 10.6). In some processes, pressure relief valves disc harge to a flare system where the diverted gas and fluid are safely burnt. API STD 520 Sizing, Se lection and Installation of Pressure-relieving Devices and API STD 521 Pressure-relieving and Depressuring Systems provide guidance on these topics. (API STD 520 and API STD 521) If a loss of containment occurs such as a tank overfilling, a containment dike (passive) can be provided to contain the fluid and avoid its flowing in areas containing ignition sources or environmentally sensitive areas. If a loss of cont ainment and fire occurs in a process unit, an automatic suppression system (active) may be acti vated to control the fire and sloped drainage provided to limit the extent of the fire. Occupied buildings may be protected by blas t or fire walls (passive) to mitigate the potential impacts on building occupants.

Conducting PHA Revalidation Meetings 151 7.4 PRINCIPLES FOR SUCCESSFUL REVALIDATION MEETINGS Actual experience in conducting revalid ations, by PHA practitioners across a number of companies, has highlighted so me keys for success, as well as some things that can impede success. While none of the items listed are absolute rules, they do provide valuable guidance. Successful Practices: • Having a formal code of conduct or team charter that fosters participation by all team members • Ensuring the team includes members with all the required skills and diverse experience • Including process experts who will be available as needed for the revalidation sessions and include them in the opening meeting • Informing the team members of the revalidation strategy for each aspect of the PHA (core, complementary, and supplemental analyses) and explaining the rationale for each • Refreshing the team members’ understanding of the organization’s risk tolerance criteria • Ensuring that some team members were not involved in performing the prior PHA • Preparing appropriate worksheets to capture revalidation team discussions • Providing team members with an overview of process operations, an overview of the physical layout , and a refresher in the analysis techniques, as necessary • Providing team members with an overview of changes since the prior PHA • Suspending a team session when an adequate team is not present and rescheduling a later time when required participants (or substitutes with similar qualifications) will be available • Providing team members with an overview of recent, relevant, process safety incidents • Performing a detailed review when Updating the prior PHA. A detailed review affords the PHA revalidation team the time and resources to help identify new concerns/hazards that may have been overlooked in the prior PHA

18 Inherently safer concepts will enhance overall risk management programs, whether directed towa rd reducing the frequency or consequences of potential incidents. 2.4 INHERENTLY SAFER STRATEGIES Approaches to achieving inherently safer processes and plants have been grouped into four major inhere nt safety strategies. The names or titles of these strategies were estab lished in the 1st and 2nd editions of this book. These groupings follow the original treatment of the topic of inherent safety as published by IChemE and IPSG (Ref 2.17 IChemE) and Kletz (Ref 2.20 Kletz 1984, Ref 2.19 Kl etz 1991). However, in this earlier, seminal work, different names/titles we re used. In general, these may be considered subsets of the four core strategies. These different names/titles and how some strategies have been incorporated into the four used herein are presented in this section for clarity and as a reference in interpreting older work on inherent safety: Minimize Using or having smaller quantities of hazardous substances (also calle d intensification). Substitute Replace a chemical/material with a benign or less hazardous substance; or repl ace a process or processing technology with one that is benign or is less hazardous. Moderate Use less hazardous or energetic processing or storage conditions, a less hazardous form of a material, or facilities that minimize the impa ct of a release of hazardous material or energy (also called Attenuation and Limitation of Effects ). Simplify Design facilities which eliminate unnecessary complexity and make operating erro rs less likely, and which are forgiving of errors that are made (also called Error Tolerance). These inherently safer design strategies are discussed in more detail in Chapters 3-6. Examples can be found in Chapter 8, which discusses inherently safer solution opportunities throughout the life cycle of a process. Table 2.1 presents a mapping of th e original concepts to the CCPS ISD strategies as restated by Kletz in 1998, as well as others who have

Heat Transfer Units 209 valves on the bottom heat exchanger. Generally there is not enough room to place drain or vent valves between the stacked heat exchangers. Even chemical cleaning can be reduced in this arrangement. As there is generally not enough room between the two stacked heat exchangers, the sensors can be deleted and only the “location” for the portable sensors are left. The concept of location for portable sensors will be cov-ered in Chapter 13. 11.8.2 Heat Ex changers in Parallel Heat exchangers can be placed in parallel when all of them are operating, or in some less common cases to provide spare capacity for the heat exchanger. If the flow rate to a heat exchanger is huge, and/or the required temperature change is high, a heat exchanger with an overly high heat transfer area may be needed. Each type of heat exchanger has a limitation on the heat transfer area. Therefore, sometimes splitting the flow rate and putting two or more heat exchangers in parallel may be needed to avoid placing a heat exchanger that needs an overly high heat transfer area. The limitation on the heat transfer area for each heat exchanger is dict ated by technical and economical factors but some- times can be deviated from. For example very large heat exchangers with very large heat transfer areas will require bringing a heavy duty crane to the plant when the heat exchangers need to be sent to the workshop for maintenance. Large cranes are not available in all plants and usually only large plants can afford to keep them. Therefore, we can have multiple heat exchangers in par - allel and all of them are working, for example, 2 × 50% a pplications. The other case where we may have parallel heat exchang- ers is when we need a spare heat exchanger. Generally speaking, heat exchangers are too expensive to allow us to keep them as a spare in plants. Therefore it is very rare to see a parallel heat exchanger 2 × 100%, which me ans one operating heat exchanger and another spare heat exchanger. However, in some cases, for example if the service fluid is very fouling, we may have a spare heat exchanger.11.9 Aerial Coolers If it is decided to use ambient air as the cooling medium on the shell side of a S&T HX, the type of heat exchanger can be used is an “aerial cooler. ” In aerial coolers, a bundle of tubes (generally finned tubes) is secured in a frame and there is no shell at all. Air blows through the tube bundle with the help of fan(s). Aerial coolers have the advantage that they have no need for a cooling medium because they use ambient air for cooling purposes. However, they need electricity for the operation of the motor that is connected to the fan. Aerial coolers are generally used in multiple units in plants. The smallest component of an aerial cooler is the “tube bundle. ” Each tube bundle has one set of dedicated head-ers, inlet header and outlet header. Each tube bundle may have one or two (even up to four) inlet process flows and the same number of outlet process flows (Figure 11.9). A “unit” is several side‐by‐side tube bundles that work as a single piece of equipment and carry one tag on the P&ID. Aerial coolers can be seen in process plants as “banks” , which are large pieces of equipment. Each bank of aerial coolers can be more than one tagged aerial cooler in a P&ID. One (or more) specific area of a plant may be ded-icated to aerial coolers and all the aerial coolers of the plant will be located there as a “bank” (Figure 11.10). No gauges in between DrainVent TW PPTW PP Figure 11.8 Stacked hea t exchangers. Figure 11.9 The tube bundle of an aer ial cooler. An aerial cooler bank A unit A unit Figure 11.10 The unit and bank in aer ial coolers.

Piping and Instrumentation Diagram Development 400 However if non‐coincident “design pressure @ design temperature” is selected one check should be done to make sure nothing is missed. This check is: “the pressure that an item can tolerate at the highest absolute temperature should be checked too. ” This check basically means making sure that the pressure corresponding to the non‐coincident design temperature is not higher than the selected design pressure. The design temperature of an item could be decided based on a specific margin on the HHT (high‐high temperature) of the item. The margin could be any number from 5 to 30 °C and is instructed by the company guidelines. 18.8.2 Sour ces of Rebel Pressures Design pressure is decided based on rebel pressure. However there are some cases that rebel pressure can-not be used for the purpose of design pressure specifica-tion. These are the cases where there is no specific maximum sustainable rebel pressure. For example if the discharge side of a positive displacement valve is closed off, the pressure will rebel and increase. However as this pressure doesn’t eventually stay at a specific value, it cannot be used as the design pressure. Such cases can only be handled by placing a pressure safety device, as stated in Chapter 12. In Table  18.11 some reasons for rebelling pressures, either on the high pressure side or low pressure side, are listed. Here we explain one important rebel high pressure and one rebel low pressure scenario. The rebel high pressure can be decided to be the “dead head pressure or shut‐off pressure. ” It is very common to see the “design pressure” of items on the discharge side of a dynamic fluid mover if it is decided based on the “dead head differential pressure” of the fluid mover.The rebel low pressure can be decided to be “full vac - uum” for the equipment that may need “steaming out” during its life cycle in the plant. If a piece of equipment deals with oily material it may need steaming out for cleaning purposes. It has been observed before that a ves - sel was cleaned by steaming‐out, and a quick but harsh rain caused quick condensation in the vessel, which led to a vacuum inside of the vessel and the vessel crumpled like a piece of paper. For such cases the design vacuum of the vessel should be specified as “full vacuum. ” 18.8.3 Sour ces of Rebel Temperatures Listing the scenarios for rebel temperatures are easier where there is something that change the temperature of the process fluid. For example in an exothermic reactor, the HHT could be decided based on the maximum temperature attained when the cooling water to the reactor jacket is  –  for whatever reason – stopped. The other example is the HHT for a piece of equipment downstream of two heat exchanger in series may be decided when one heat exchanger (possible the one with larger duty) fails to do its functionality. However there are some other cases that rebel tem- peratures exists because of other reasons. In Table 18.12, some reasons for rebelling temperatures, either on the high temperature side or low temperature side, are listed. 18.8.4 Design P ressure and Design Temperature of Single Process Elements Below the design pressures of several items are discussed. ●Tank: the tank design pressures are requested by pro- cess engineers and are provided by the mechanical engineers of the fabricator. The requested design pressures, however, should be less than the maximum allowable design pressure dictated by the associated Table 18.11 Specific rebel pr essure scenarios. Rebel high pressure Rebel low pressure ●Centrifugal fluid mover discharge closing off: dead head pressure ●Fail open of control valve or regulator where it is connected to high pressure reservoir ●Runaway reaction where the products of side‐reactions are gaseous ●Vaporization (because of abnormal heat) ●Fail open of control valve or regulator where it is connected to vacuum reservoir ●Runaway reaction where the product of side‐reactions are liquid while the raw material are gaseous ●Vapor condensation (because of abnormal cooling)Table 18.12 Specific rebel t emperature scenarios. Rebel high temperature Rebel low temperature ●Runaway exothermic reaction ●Fail open of control valve or regulator where it is connected to high temperature fluid ●Runaway endothermic reaction ●Fail open of control valve or regulator where it is connected to high temperature fluid ●Quick vaporization (because of pressure drop) ●Quick pressure drop of a liquefied gas because of Joule–Thomson effect (in some cases temperature increases)

251 HAZOP WORKSHEET Area: Unit: Node: Drawings: Design Intent: No. Guideword Deviation Causes Consequences Safeguards Recommendations Action by Figure 12.5. Example HAZOP analysis worksheet (enggcyclopedia) Table 12.4. HAZOP overview Typically Used During Resource Requirements Type of Results Advantages and Disadvantages Pilot plant operation Detailed engineering Routine operation Expansion or modification Material, physical, and chemical data Basic process chemistry Process flow diagram Piping and Instrumentation Diagrams Scenario-based documentation of deviations, causes, consequences, safeguards, risk ranking, and recommendations, if any. Provides a structured methodology to systematically and consistently analyze hazard scenarios. Provides input to Layer of Protection Analysis by identifying high consequence scenarios. Potential for redundancy in covering hazards. HAZOP, like the other hazard identification me thods, is a qualitative analysis. The higher risk scenarios from a HAZOP analysis are frequently used as the foundation for a Layer of Protection Analysis (LOPA). LOPA uses simplifying rules to evaluate initiating event frequency, independent layers of protection, and conseq uences to provide an order-of-magnitude estimate of risk. The primary purpose of LOPA is to determine if the scenario has sufficient layers of protection to prevent or mitigate th e consequences. LOPA is discussed in Chapter 14 along with other risk assessment techniques. Failure Modes and Effects Analysis (FMEA) The FMEA method originated in the U.S. M ilitary where it was used to assess potential equipment failures and reliability issues. The purpose of an FMEA is to identify single HAZARD IDENTIFICATION

6.2 Assess the Organization’s Pr ocess Safety Culture |213 fam iliar with. Follow-up with a more general question about how that role works. Conduct the Interview. Work through the interview protocol, using follow-up questions to clarify answers and to assure com pleteness. Typically, interviewers will use three types of questions: Open-ended questions seek inform ation in the interviewee’s own words. Questions like “What does a good process safety culture m an to you?” and “What needs to be done to reach your view of a good process safety culture?” allow the interviewee to provide their opinion more fully. While the answers to open-ended questions can be harder to evaluate, their inform ation is more valuable. Open-ended questions can sometimes lead to extraneous information and tangential stories, which the interviewer can manage with other forms of questions. Leading questions help steer the direction of the conversation. A leading question like “Can you tell me m ore about (the desired focus of the original question) can be useful to bring a tangential question back on track. However, avoid leading question like “You follow procedures, don’t you?” These can sometimes direct the desired answer or be perceived by the interviewee as a trap. Closed questions seek concrete answers, typically “Yes” or “No.” These provide the most precise information but limit the respondent’s ability to provide valuable detail. For exam ple, a closed question such as “Has the Alkylation Unit PHA been revalidated yet?” m ay result in the answer “No.” Once “No” has been stated, the interviewee may become defensive and information m ay be lost. However, closed questions like “Do I understand correctly what you said that … ?” can be very useful to check understanding. Take care to avoid close-ended questions that feel like a legal cross-examination. • • •

Piping and Instrumentation Diagram Development 406 3) Think about potential weakness points of the equip- men t. Consider these words: thin, tight, non‐metallic, multi‐component part, expensive part, moving part, vibrating part, etc. 4) Is the e quipment sensitive toward suspended solids and are there suspended solids in the incoming flow? If the answer is yes, a strainer may be needed upstream of the equipment. 5) Is the e quipment sensitive toward surged flow and is the incoming flow likely to surge? If the answer is yes, a surge dampener may be needed upstream of the equipment. 6) What i s the plan if an off‐spec product is produced? Can it be recycled or should it be discarded? 7) What do you need to put in for easier inspection of the equipment? You need to put more facilitating things if more frequent inspection is needed or if missing an inspection has large consequences. 8) You mo st likely need to put an isolation system (includ- ing isolation valves) around your equipment unless you can not afford pulling off the operation of the equipment during the normal operation of the plant. 9) What do you w ant to do with incoming flow when the equipment is non‐operational? 19.2.2 P&ID Dev elopment: Control and Instruments After developing the piping and equipment portion, the next step is the instrumentation and control portion. However, after finishing this task you may need to check the piping and equipment portion again. P&ID development, like other design processes, is not a straightforward process. Here again we need to cover the four stages of opera- tion. However instrument and control are mainly needed for the first two stages: normal operation and non‐routine operation of equipment. For easy decision making a matrix similar to that shown in Figure 19.2 can be developed for each piece of equipment. The different process and non‐process parameters are given in the first column and the different wrapping layers are placed in the first row. A check mark shows if it is decided to use each wrapping layer around each parameter. Decision making for each steering loop was discussed in detail in Chapter 16.However, here we provide a simple method as a preliminary step. To decide about parameters for each steering loop, all the applicable parameters should be classified in five groups of “barely important parameters, ” “mildly impor - tant parameters, ” “very important parameters, ” and “criti-cal parameters. ” Each importance level is connected to an I&C requirement: “nothing” for the first group, “monitor field” for the second group, “monitor control room” for the third group, “regulatory control loop” for the fourth group, and “safety interlock function” for the last group. Such correspondence is shown in Table 19.1. There are some cases that a parameter is not definable for a piece of equipment. For example, for a vessel flow rate is not definable. In some other cases a parameter is not important. For example composition generally is not important for pumps. Barely important parameters are chosen based on screening of the non‐important parameters and mildly important parameters. Mildly important parameters are the parameters that affect the operation of the unit but are not the main parameters of the unit. Very important parameters are the parameters that are related to the main duty of the unit. Critical parameters are the parameters for which their violation creates risk to personnel, assets and the envi-ronment. The increased risk could be through increasing the probability or consequences or both. Table  19.2 shows a parameter matrix for a typical pump in a hot water service. In Table 19.2 different process parameters are exam- ined against their criticality to come up with the required monitoring and control system. We generally don’t care about the temperature in pumps but as this pump works with hot water then there is chance of having cavitation when the temperature is high. It is a good idea to put a ICSS action Parameter Pressure Temperaure Flowrate Level Composition MonitoringField Control Regulatory control Interlock Figure 19.2 St eering component selection matrix. Raw material Conversion unit Seperation unitProduct By-product Figure 19.1 Pr ocess plant, a bird’s eye view.

EQUIPMENT FAILURE 223 Piping Most facilities have miles of piping. This pipi ng supports the flow of feed, product, and everything in between to and from the equipment previously discussed in this chapter. Piping comes in all sizes and materials. The piping materi al, the chemicals it is transporting, how it is protected, and its routing are fact ors in how it may be damaged. Figure 11.36. Piping rupture (CSB 2015) Example 1. Piping may corrode over time depending on the piping material and the composition of the chemicals flowing through th e piping. An example of this is the Chevron Refinery in Richmond, California suffered a fire in 2012 (Figure 11.36). The source of the fire was a rupture of unit piping due to sulfidation co rrosion applicable at high temperature. Other metallurgy and other chemical combinations can cause different types of corrosion. Example 2. Through science and research, an im proved understanding may be gained regarding the appropriate type of metallurgy for use with specific chemicals. Facilities that were constructed years ago used the understa nding and materials of that time period. A hydrofluoric (HF) acid alkylation unit with a piping elbow installed in about 1973 contained more nickel and copper than would be recommended today in API RP 751 “Safe Operation of Hydrofluoric Acid Alkylation Units” or in a Na tional Association of Corrosion Engineers (NACE a) paper. The elbow failed resulting in a fire, expl osion, and release of toxic HF acid. Images of the HF unit before and after the explosion and fire are shown in Figure 11.37.

274 INVESTIGATING PROCESS SAFETY INCIDENTS Table 11.1 Human Factors Issues (cont.) PERSONNEL FACTORS Mental States Physiological States Physical/Mental Focused attention Physiological state Reaction time Complacency Physical health Vision/hearing Distraction Influence by medication Knowledge Mental fatigue Physical capability Haste Fatigue Situational awareness Motivation Task saturation Language/cultural differences Shift cooperation/teamwork WORKPLACE FACTORS Design Maintenance Environmental Instrumentation clarity Poorly maintained equipment Illumination / visibility Layout work space, access Poorly maintained workspace Storm Communications equipment Poorly maintained communications equipment Temperature (hot or cold) Equipment provided for the job Labeling Wind Noise level Incident investigations must include human performance considerations and human fa ctor issues. The use of ch ecklists and flowcharts is a helpful technique to aid investigators i n addressing human performance issues. For example, checklists can be built using the information in the tables shown above in this section. Checklists may be strengthened with input from a human psychologist, an expert on human reliability analysis, and experienced incident investigators. Numerous interface devices have been developed that translat e theoretical models of human error causation into easy-to- understand engineer ing terms. Some of these devices are in the form of logic trees or checklists. Chapter 10 describes the use of checklists in root cause analysis.

INVESTIGATION M ANAGEM ENT SYSTEM 49 4.1 SYSTEM CONSIDERATIONS 4.1.1 An Organization’s Responsibilities Incident investigation is only one of the many elements of a process safety management program (CCPS, 2007), and is notably one that plays an essential role in identifying overall management syst em weaknesses on a continuous basis. Establishing a high quality incident investigation program begins with management’s support, comm itment, and action. To demonstrate support, it is common pr actice to establish a written policy regarding incident notification, investigation, and dissemination of findings; to communicate this policy to the workforce; and to sustain the policy over time by committing resources for continuous improvement (see Chapter 15). This is often expressed in a formal statement written to achieve the following goals. • Communicate management’s commitment to prevent recurrences by determining causal factors and root causes, evaluating preventive measures, and taking follow-up action. • Recognize the importance of implementing investigation findings as a strategic risk control mechanism. • Strongly support reporting and investigating near-misses. • Clearly focus on finding causal fa ctors, root causes, and management system weaknesses, while avoiding assignment of blame. • Endorse sustained commitment of resources for the investigation program, including training team members. This supports employee participation in the investigation program and the appropriate and timely implementation of recommendations. • Emphasize the value and necessity of communicating and sharing the lessons learned from the investigation to all that could reasonably benefit. • Support a system to ensure that all recommendations and findings are resolved and that decisions and actions are documented. • Establish a mechanism to fo ster continuous improvement. Management demonstrates support fo r this policy by nurturing an atmosphere of trust and respect that encourages openness in reporting incidents throughout the organization. Failure to achieve this positive atmosphere may result in hidden incide nts and low or no reporting of near- misses, which results in lost learning opportunities that could have potentially led to avoida nce of future accidents.

8 inherent safety in determining the r oot causes of an incident. By means of an illustrative example, Gupta et al . (Ref 1.7 Gupta) provided evidence of the linkages between inherent safety and the cost of process safety. Their work helps to establish a clear bu siness case for the use of inherent safety principles in management ef forts directed at enhancing process safety. Further motivation for the curre nt research is found in the comments of employees who have reviewed the field of inherent safety and inherently safer design, including (Ref 1.3 Bollinger), (Ref 1.7 Gupta), (Ref 1.12 Kletz), (Ref 1.10 Khan) and (Ref 1.26 Vaughen 2012b). For example, Khan and Amyotte (Ref 1.10 Khan) have remarked that the various elements of process safety management can be seen to have at least a partial basis in inherent safety . This fact has been recognized by companies that have incorporated inhe rent safety as a “named feature” in their safety management document ation and have developed internal standards for the use of inherent safety principles. Yet the term inherent safety is typically not named as such in the general description of process safety management systems. Per Bo llinger, et al. (Ref 1.3 Bollinger), explicit use of inherent safety te rminology within such management systems is a possible means of furtheri ng the adoption of inherent safety principles in industry. Over the years, the interest in inherent safety from government, industry, and the public has increase d. Inherent safety’s promise has produced heightened expectations and it is seen almost as a panacea to reducing risks in the chemical proces s industries as the public becomes aware of the concept. Inherent safety is incorporated into safety and security regulations in pa rts of the United States at the local, state, and federal levels. Inherent safety has been proposed as a leading requirement for chemical security re gulations in the U.S. Congress. To better clarify and more precisely define the terminology, the US C h e m i c a l S e c u r i t y A n a l y s i s C e n t e r ( C S A C ) o f t h e D e p a r t m e n t o f Homeland Security (DHS) contracted with CCPS to provide a technology-based definition of Inherently Safer Technology (IST) (Ref 1.4 CCPS/DHS). Some of the best compilations of in formation on IS can be found in the works of (Ref 1.14 Kletz 1978), (Ref 1.15 Kletz 1996), (Ref 1.16 Lees) and the final report of the INSIDE (INherent SHE in DEsign) project in

REACTIVE CHEMICAL HAZARDS 99 Chemical Reactivity Worksheet (CRW) The Chemical Reactivity Work sheet (CRW) is a free software program providing extensive process safety information (Refer to Sectio n 2.5) The CRW includes data required to understand the hazards associated with the inadvertent and intentional mixing of reactive chemicals. This includes the chemical reactivity of thousands of common hazardous chemicals, compatibility of absorbents, and suitability of mate rials of construction in chemical processes. It is designed to be used by emergency resp onders and planners, as well as the chemical industry, to help prevent hazardous chemical incidents. It is available at https://www.aiche.org/ccps/resour ces/chemical-reactivity-worksheet . (CCPS) Versions of the CRW were developed by th e collaboration of several organizations including the Center for Chemical Process Sa fety, Environmental Protection Agency, NOAA's Office of Response and Restoration, The Ma terials Technology Institute, Dow Chemical Company, Dupont, and Phillips. The CRW contains a database of chemical datasheets for thousands of chemicals. The chemical datasheets in the CRW database cont ain information about the intrinsic reactive hazards of each chemical, such as flammability, the ability to form peroxides, the ability to self- polymerize, explosivity, strong oxidizer or reducer capability, water or air reactivity, pyrophoricity, known catalytic activity, instabilit y, and radioactivity. Datasheets also contain general descriptions of the chemicals, physical properties, and toxicity information. They also include case histories on specific chemical inci dents, with references. The CRW also allows the creation of custom chemical datasheets, for ex ample, to use in documenting properties of a proprietary chemical that is not in the CRW database. The CRW uses chemical pairs. In order to fully understand inadvertent and intentional mixing, the reactivity of the entire mixture must be understood, not just the pairs. The CRW includes a reactivity prediction worksh eet to virtually "mix" chemicals to simulate accidental chemical mixtures, such as in the case of a train derailment, to learn what dangers could arise from the accidental mixing. For exampl e, if the reaction is predicted to generate gases, the CRW will list the potential gaseous products, along with literature citations. The CRW has two additional modules of particul ar interest to the chemical industry. One of them discusses known incompatibilities between certain chemicals and common absorbents which are used in the cleanup of small spills. The other module, new in CRW 4, contains information about known incompatibilit ies between certain chemicals and materials that are used in the construction of containers, pipes, and valving systems on industrial chemical sites. The Mixture Manager screen provides a search for chemicals in the CRW's database, a preview of the information on the chemical datash eets, and the creation of virtual mixtures of chemicals. It also provides access to all the ot her features of the program from this screen, including the compatibility chart and hazards report for any mixture created, reference information about the reactive groups used in the CRW, and information about absorbent incompatibilities with certain chemicals. The Compatibility Chart shows the predicted hazards of mixing the chemicals in a mixture in an easy-to-use graphical interface. The reactivity predictions are color-coded, and

465

FIRE AND EXPLOSION HAZARDS 69 Figure 4.14. Relationships between the different types of explosions (Crowl 2003) Table 4.3 provides examples of the types of ex plosions. You can observe that some incidents can involve multiple types of explosions, for example a vessel rupture leading to a BLEVE. Physical Explosion - The catastrophic rupture of a pressurized gas/vapor-filled vessel by means ot her than reaction, or the sudden phase-change from liquid to vapor of a superheated liquid (CCPS Glossary) Boiling Liquid Expanding Vapor Explosion (BLEVE) A type of rapid phase transition in which a liquid contained above its atmospheric boiling point is rapidly depressurized, causing a nearly instantaneous transition from 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. (CCPS Glossary)

308 | Appendix E Process Safety Culture Case Histories explosion that resulted in injuries to personnel and significant property dam age to the facility. In the investigation that followed, the operator stated that he did not feel com fortable taking SWA action and that a supervisor should have been there to m ake that call. When asked why he was not comfortable, the operator responded that in over the years, when SWA was used, there had been a lot of second-guessing by investigators after the fact. Further review showed that the incident investigation reports described alternative actions that the operators could have taken in response to the indications they were receiving at the control board that would have abated the transient but kept the process running. Som e reports also suggested disciplinary action, although none was taken. When operators exercise SWA, it is certainly possible that options existed for them to bring the process under control. B ut under duress, it is hard to know if such an option exists or not, which is why SWA is so important. How can incident investigators address potential alternative actions without underm ining SWA? Foster Mutual Trust, Combat the Normalization of Deviance. E.23 SWPs by the N um bers Safe work permits are involved activities used to help ensure that the hazardous work is fully prepared before any work begins. In a large facility, these permits (e.g. Safe Work, General Work, Hot Work, Confined Space Entry, Line Breaking, and others) had been issued by the on-duty operators. The very large num ber of perm its being sought at the beginning of day shift would overwhelm the board operator and com pletely distract him from running the equipment. To address this problem, the company appointed a set of perm it approvers especially for this “rush hour.”Actual Case History

52 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS 3.3.5 Additional RBPS Elements Related to Management of Abnormal Situations The RBPS elements in this section ar e briefly discussed in this chapter for awareness. Some of them, such as MOC, will be discussed in Chapter 5. 3.3.5.1 Compliance with Standards Includes applicable regulations, standards, codes, and other requirements issued by nation al, state/provincial, and local governments, consensus standards or ganizations, and the corporation. Interpretation and implementation of these requirements include development activities for corporate, consensus, and governmental standards. 3.3.5.2 Process Safety Competency Addresses skills and resources that the company needs to have in the right places to manage its process haza rds. Provides verification that the company collectively has these skills and resources and applies this information in succession planning and management of organizational change. 3.3.5.3 Asset Integrity and Reliability Activities to ensure that importan t equipment remains suitable for its intended purpose throughout its serv ice. Includes proper selection of materials of construction; inspec tion, testing, and preventive maintenance; and design for maintainability. 3.3.5.4 Management of Change Process of reviewing and authoriz ing proposed changes to facility design, operations, organization, or activities prior to implementing them, and updating the process sa fety information accordingly. 3.3.5.5 Conduct of Operations Means by which management and operational tasks required for process safety are carried out in a deliberate, faithful, and structured manner. Managers ensure workers carry out the required tasks and prevent deviations from expected performance:

182 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Figure 11.2. Buncefield storage depot before the explosion (HSE 2017) Figure 11.3. Buncefield Terminal site and wider area after explosion (HSE 2017)

Documenting and Following Up on a PHA Revalidation 165 8.5 PRINCIPLES FOR SUCCESSFUL DOCUMENTATION AND FOLLOW-UP Actual experience in conducting revalid ations, by PHA practitioners across a number of companies, has highlighted so me keys for success, as well as some things that can impede success. While none of the items listed are absolute rules, they do provide valuable guidance. Successful Practices: • Documenting the revalidation to maximize its value to future users • Documenting the PHA in a way that simplifies future revalidations (e.g., facilitating the Update approach) • Describing complete loss scenarios, from initiating cause, through intermediate events, to potential consequences • Documenting the changes made in an Update clearly (e.g., by using a distinct font color, using software features to track changes, or making detailed annotations in a comment column) • Consolidating the core, complementary, and or supplemental analyses, along with any relevant portions of prior PHAs in a single revalidated PHA document that satisfies both regulatory and organizational requirements • Writing recommendations that are clear to people who did not participate in the revalidation • Minimizing the time between the PHA team developing recommendations and communica ting them to management • Resolving recommendations as soon as possible/practical • Retaining a complete set of records for the next revalidation Obstacles to Success: • Documenting team discussions with abbreviated or cryptic notes, or using terminology unfamiliar to operating personnel • Leaving blanks in worksheets, tables, or checklists • Having no clear rationale for Update vs. Redo documentation style • Documenting the PHA in Update style when a Redo is performed • Using Update documentation on many sequential revalidations • Failing to revise all the core, complementary, and supplemental study documentation affected by the revalidation

140 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION New engineer activities on this topic will depend on their working location. Coastal locations may have exposure to hurricanes and associated flooding, while other areas may be concerned with wind and wildfires. Northern loca tions such as Canada and Alaska will be most concerned about very low temperatures. Desert locations may need to address sandstorms. Locations such as California and Turkey sh ould address earthquake preparedness as demonstrated by the Fukushima incident. Tools Kinetic and potential energy hazards ca n be identified and managed through use of engineering tools learned in undergraduate engineering curriculum. CCPS Monograph: Assessment of and planning for Natural Hazards. lists many data sources and approaches for identifying meteorological and geological hazards and addressing them in design and emergence response preparedness plans. (CCPS 2019) The monograph includes reference to the following data sources and design criteria. Data sources. • Federal Emergency Management Agency (FEMA) flood maps - https://msc.fema.gov/portal/home • United States Geological Survey (USGS) seismic maps - https://earthquake.usgs.gov/hazards/designmaps/usdesign.php • American Society of Civil Engineers (ASCE) seismic guide - https://hazards.atcouncil.org/#/ • National Oceanic and Atmospheric Adminis tration (NOAA) tornado prediction - https://www.spc.noaa.gov/new/SV Rclimo/climo.php?parm=allTorn • Tornado Wind Prediction - https://hazards.atcouncil.org/#/ (ATC) • National Hurricane Center (NHC) Storm surge maps - https://www.nhc.noaa.gov/nationalsurge/ (NHC and NOAA 2019 a) • Snow load - https://hazards.atcouncil.org/#/ (ATC) • NOAA Hurricane center - https://www.nhc.noaa.gov/climo/ Design guidance. • ASCE Flood Resistant Design and Construction, ASCE 24 (ASCE 2014) • ASCE Minimum Design Loads and Associated Criteria for Buildings and Other Structures, ASCE /SEI 7-16 (ASCE 2016) • CCPS Guidelines for Safe Warehousing of Chemicals (CCPS 1998) • CCPS Guidelines for Safe Storage and Handling of Reactive Materials (CCPS 1995) • Guidelines for Siting and Layout of Facilities, 2nd Edition (CCPS 2018) • FM Global Property Loss Prevention Data Sheets 1-2 Earthquakes (FM Global 2021)

Appendix 215 Other References in Table A.2-2 o (Behie 2008) Dolphin Energy [44] o (Bloch 2016) Bhopal [95] o (EPA 2018) Tosco Avon Refinery [118] o (EPSC 2019) Flare System [119] o (Fogler 2011) Monsanto [62] o (Meshkati 2014) [77] o (NFPA 2011) Hydrocracker Excursions [120] o (UK HSE 1997) Texaco Pembroke [68]

4.5 Process Safety Culture Metr ics |145 m eetings including substantial discussion of process safety. An imbalance in emphasis m ay be indicative of a m anagement attitude that process safety is less important. Frequency with which relevant process safety statistics are shared with the organization. A low value m ay indicate that m anagement does not adequately appreciate the value of informing the workforce of the organization’s process safety perform ance. Manager attendance at management review meetings. Poor attendance m ay indicate a low interest in process safety perform ance, or in communicating m anagement expectations. Foster Mutual Trust This core principle can be subjective, and leaders and workers may have a different opinion regarding trust. Those opinions may be difficult to elicit from fixed surveys. Therefore, this core principle should be assessed primarily by interviewing leaders and workers. Generally, the interviews should consider whether interviewees feel that: o A just system exists where honest errors can be reported without fear of reprisals, o Submitted information will be acted upon, o B ad ideas can be challenged, discussed, and resolved satisfactorily; and o Errors will not be punished unless the act was reckless, deliberate, or unjustifiable. Since trust between peers is also im portant, the sam e approach can be applied to peer interactions. Ensure Open and Frank Communications Do employees exercise stop-work authority? When they do, does leadership thank them and take care to avoid second-guessing their decisions? • • • • •

Overview of the PHA Revalidation Process 21 1.8 RELATIONSHIP OF RBPS PILLARS TO A PHA REVALIDATION The RBPS book [3] describes the RBPS approach and recognizes that not all hazards and risks are equal. It advocates that resources should be focused on more significant hazards and higher risks. The RBPS approach is built on the four accident prevention pillars. Included is a summary of how PHA revalidation is related to each pillar and its associated elements. For detailed information on the RBPS accident prevention pillars an d their elements, see the RBPS book: Commit to Process Safety . Process safety culture is generally defined as “How we do things around here” or “How we behave when no one is watching.” In a facility or company with a strong process safety culture, a PHA revalidation is an important and highly valued exercise. In order to meet the process safety goals of an organization, a revalidation team consists of a diverse team of well-trained process experts who understand the importance and value of the revalidation. Conversely, in a facility where process safety culture could be improved, the PHA revalidation exercise is seen as a rote task that must be completed (e.g., to satisfy regulatory or company require- ments). The goal for such a PHA is to be able to claim it was done on time but required little or no effort to perform and created minimal additional work for follow-up. Understand Hazards and Risk. Process knowledge is a critical key to the success of any PHA revalidation and includes both process documentation and the competency of process experts (wit h many areas of experience including operations, engineering, and applicatio n of PHA methods). A revalidated PHA that incorporates current knowledge of the process and its hazards is foundational for all the other RBPS elements that rely on the PHA, such as the asset integrity of safety-critical equi pment, MOC, or emergency response planning. Process Safety Culture In a facility with a strong process safety culture, there is trust that the PHA revalidation will be of value and the results of the study will be acted upon. This trust and outreach/ communication flows from opera- tions, through the PHA revalidation team, to management, and then back again.

302 Human Factors Handbook 23.2.2 Why did this happen? This incident had multiple causes. The Ch emical Safety Board report noted that: • Repairing a cracked seal loop was po stponed (allowing vinyl fluoride to flow to the slurry tank). • The equalizer line, which provided the direct path for the flammable vapor to enter tank 1, was not blinded and was not included on the lock out card for the slurry tank. • The hot work permit procedure did not require testing the atmosphere inside the tank. There were issues with the role of contractor management, which are discussed next. DuPont recognized that contractors ma y be unfamiliar with process safety or activities on their sites. DuPont’s inte ntion was to ensure that everyone would understand the work and potential hazards. In this instance, DuPont intended that the construction field engineer and the area manager would help the contractor understand the hazards. The contractor submitted a “hot work permit”. However, the section of the permit which asked if flammable material would be within 35 feet (10 meters) of the work was not completed. The hot work was within 35 feet (10 meters) of the slurry flash tank that vented vinyl fluoride to the atmosphere. It was concluded that: “The contractors were unfamiliar with the Tedlar® process and the process equipment involved. The contractors did not know what the slurry flash tank was or which chemicals were presen t inside it.” p10, CSB [98] The permit was signed off by the DuPont construction engineer and by the area manager. It was reported that: • The DuPont construction engineer for the slurry tank work had no working knowledge of the Tedlar® process. • The construction engineer expected the area manager would advise the contractors of plant-specific proce ss safety information for hot work. • An area manager signature was ob tained by someone in a service department. The area manager lacked knowledge of the area and of the Tedlar® process. In addition, they did not perform the required “walk down” of the area before signing the permit. • The area manager assumed the constr uction engineer was briefing the contractors on-the-job and the hazards.

5.5 References | 67 6.Prepare. The evaluator works with peers to develop an action plan to implement the changes. The plan includes resources, capital, training, communication, and other key factors. 7.Implement. After company leadership has endorsed the plan, the plan is implemented. The implementation team may include both corporate experts and site leaders. Implementation includes leadership, workforce involvement, training, conduct of operations, metrics, and ongoing management review. 8.Embed and Refresh. Company and site leadership now manage according to the changes as implemented. Ongoing communications and training remind, maintain the sense of vulnerability, and reinforce the need to maintain the new way of doing things. These components of the REAL Model will be described in greater detail in chapters 6 and 7. Chapters 9–14 will provide some hypothetical examples of how the REAL Model may be used. 5.5 References 5.1 CCPS (2019). Process Safety Leadership from the Boardroom to the Frontline. Hoboken, NJ: AIChE/Wiley. 5.2 Gardner, H. (1995). Reflections on multiple intelligences: myths and messages. Phi Delta Kappan 77: 200–209. 5.3 Gardner, H. (2011). Frames of the Mind: The Theory of Multiple Intelligences. New York: Perseus Books Group. 5.4 Hiatt, J.M. (2006). ADKAR: A Model for Change in Business, Government, and Our Community. Fort Collins, CO: Prosci Learning Center Publications. 5.5 International Association of Oil & Gas Producers (2016). Components of Organizational Learning from Events. IOGP Report No: 552. 5.6 Joshi, S. (2009). How we learn and grow. blog.practicalsanskrit.com/ 2009/12/how-we-learn-and-grow.html (accessed January 2020). 5.7 Kotter, J.P. (2012). Leading Change. Brighton, MA: Harvard Business Review Press. 5.8 Levasseur, R.E. (2001). People skills: Change management tools— Lewin’s change model. FOX Consulting Group Newsletter, July-August 2001. 5.9 Lombardo, M. M. and Eichinger, R. W. (2010). Career Architect Development Planner, 5th Edition. Minneapolis, MN: Lominger.

28 PROCESS SAFETY IN UPSTREAM OIL & GAS Other hazardous chemicals present onshore or offshore include methanol if used for flow assurance (flammable and toxic), glycols (e.g., TEG) for dehydration, corrosion and other inhibitors, acids for various treating, and amines (MEA, DEA, MDEA) for H 2S and carbon dioxide (CO 2) removal. H 2S separated from amine solutions can be almost pure and is ve ry hazardous. Storage may be required for diesel fuel for emergency generators, fire pumps and jet fuel for helicopters. An environmental issue relates to naturally occurring radioactive materials (NORM) which may leach from the formation and be transported to the surface in produced water, oil and gas. These can precipitate ou t and form solid waste. Dumping of these locally is not permitted. There are multiple chemicals used for well stimulation and water flood, but these normally do not pose a process safety ri sk. An exception is the use of hydrogen fluoride (HF) which is very toxic as well as flammable. Getting it mixed and delivered to a well, safely injected, and the returns properly handled, is a challenge. 2.5 WELL WORKOVERS AND INTERVENTIONS During the plateau and decline phase of the well (Figure 2-1), the composition of produced fluids changes. Ma ny reservoirs flow naturally initially because of the reservoir pressure and the gas content. In itially lighter materials may dominate, but over time the reservoir pressure generally decreases, heavier components may increase, and produced water may also increase. Also, in some cases sand can be produced. While initial flows may be free flowing, the change in reservoir characteristics to heavier materials may requ ire some form of flow assistance (i.e., artificial lift) to prod uce the well – either gas lift or water injection into the reservoir to maintain reservoir pressure. Other forms of enhanced oil recovery are possible using, for example, hydrocarbons, heat, or carbon dioxide. It may also be feasible to increase production and do well maintenance through workovers or interventions. Process Safety Issues There are process safety implications when the well is opened for workover or intervention. Many aspects of risk are similar to well construction (see Section 2.2), including loss of well control or blowout. The likelihood may be lower if the pressure is reduced due to reservoir depletion. 2.6 DECOMMISSIONING PHASE The end-of-life stage of a we ll is reached when flow rates produced are no longer economical and it is decided to abandon the well. The proper plugging and abandonment of an onshore or offshore well is defined by regulators globally. While differences exist, they all seek to accomplish the following. 1.Isolate and protect all subsurface freshwater zones

13 2 The Upstream Industry 2.1 UPSTREAM INDUSTRY 2.1.1 Life Cycle Stages This chapter provides an overview of the upstream industry and follows the upstream life cycle. Greater details on we ll construction, well work, and onshore and offshore production are provided in Chapters 4, 5 and 6, respectively. Engineering design, construction, and installation are covered in Chapter 7. The main life cycle stages are shown in Fi gure 2-1 and are described as follows. ●Exploration and Well Construction, including discovery and appraisa l wells ●Engineering design, construction and installation ●Production phase (covering first oil, build-up, plateau and decline) ●Well workovers and interventions to maintain well integrity and boos t production during decline ●Decommissioning / abandonmen t The complete life cycle is more comple x than the figure shows as additional wells may be constructed and well stimulation activities or other enhanced oil recovery methods may be implemented to maintain production levels. Figure 2-1. Typical upstream life cycle Exploration well Appraisal Well First Oil Build-upPlateau Decline Abandonment Economic limitProduction TimeProcess Safety in Upstream Oil and Gas © 2021 the American Institute of Chemical Engineers

386 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION The design was based on chalk drawings on th e workshop floor. No engineering review by a qualified mechanical engineer was undertaken to review the mechanical adequacy of the connection. The investigation found that “No co nsideration was given to the bending moments or hydraulic thrusts that would be imposed on the assembly due to its dogleg design. There was no reference made to vendor manuals for the expansion bellows, no r to relevant British Standards. No drawing was made for the design.” (HMSO 1975). Lessons Management of Change. Changes to a process or equi pment must be reviewed and implemented by people with knowledge appropriate to the situation. This incident is important in the history of process safety as the prime example of the importance of an MOC program. The site made no engineering review of this ch ange. As seen in the cause section, important mechanical design features were not considered during the change. Flixborough highlights the importance of Manage ment of Organizational Change (MOOC) as well as physical change. At Flixborough, “the works engineer had left early in the year and had not yet been replaced. At the time the by pass line was being planned and installed, no engineer was on site with the qualifications to perform a proper mechanical design, or to provide critical technical review on related issues. There were chemical and electrical engineers on staff, but no other mechanical engineers.” A statement often used in relation to the modifications at Flixborough is that “they didn’t know what they didn’t know”. Although the presence of a mechanical engineer may not ha ve changed the outcome if no MOC review was held, it is more likely that the significance of the change could have been recognized by someone at the plant. MOOC covers modification of work schedules, personnel turnover, task allocation changes, organizational hierarchy changes, and organizational policy changes. Guidelines for Managing Process Safety Risks During Organizational Change (CCPS 2013) covers this topic in more detail. Compliance with Standards. As stated in the summary, the site office building was destroyed. At the time, 1974, no facility siting and layout standards existed. This event is an example of why such a standard, API RP 752, “Management of Hazards Associated with Location of Process Pl ant Buildings” was developed. Introduction to Management of Change Much thought goes into the design and engineering of a facility to support it operating safely. T h i s i n c l u d e s a d e s i g n r e v i e w w i t h P H A , a pre-startup safety review, sound operating procedures, and ongoing mechanical integrity processes. Management of change addresses any changes in design, equipment, and operation. Examples of types of changes are given in Table 18.1. One of the key things to understand is what constitutes a change. Many things may be changed over the life cycle of a facility for exam ple, to improve operations, expand production, add new products, and replace worn equipment. The one example that is not a change is replacement-in-kind. This deserves a clos e review as changes that may seem like a replacement-in-kind may have minor variations that warrant consideration as a change.

120 | 4 Applying the Core Pr inciples of Process Safety Culture his previous position, the only time Congress ever directly intervened in the employment of a government contractor. Managing Ethics Ethical behavior is not an innate activity. Ethics can, and should, be m anaged. Many com panies define standards guiding ethical behavior that is encouraged as well as unethical behavior that is not tolerated. Likewise, professions and professional societies, as well as trade organizations, have had similar codes for decades and even centuries. Codes generally represent high- level aspirations of conduct, and establish consistent expectations for the conduct of m embers of the group. M oreover, they declare to the public the behavioral expectations of the group. Organizations associated with process safety that have ethical codes include the Board of Certified Safety Professionals, the B oard of Environmental Auditor Certification, and the National Society of Professional Engineers, among others. The American Institute of Chemical Engineers (AIChE) stands out am ong technical societies for its Code of Ethics. The pream ble to AIChE’s code states (Ref 4.9): “Members of the American Institute of Chemical Engineers shall uphold and advance the integrity, honor, and dignity of the engineering profession by: being honest and impartial and serving with fidelity their employers, their clients, and the public; striving to increase the competence and prestige of the engineering profession; and using their knowledge and skill for the enhancement of human welfare.” The AIChE code goes on to address three topics closely related to process safety culture: (1) Hold paramount the safety, health and welfare of the public and protect the environment in performance of •

Conducting PHA Revalidation Meetings 133 The main difference between a Redo and an Update is the starting point. In a “pure” Redo , the meeting is conducted in a ma nner similar to the initial PHA. More often, the team makes some use of the previous PHA documentation if it is of high quality. For example, the previ ous PHA may be used as a checklist. As the current team concludes its discussion of a particular loss scenario, the team might look at (or display) the previou s PHA and ask, “Is there anything we missed?” The differences might be factual (e.g., the previous team included a cause that the current team overlooked) or judgmental (e.g., the previous team took credit for a safeguard that the current team judged ineffective). Regardless, the study leader should continue the team discussion and resolve any discrepancies before moving to the next topic. Extensive use of the previous PHA is possible as long as it does not compromise the reason the Redo approach was selected. For example, perhaps the PHA is being Redone simply because the study leader believed it was the more efficient approach, given the large number of changes that affected most of the nodes. With no indication that the previous PHA was deficient, the study leader might pre-populate the revalidat ion worksheets with some information from the previous PHA and use it to expedite the current discussions as if it were an Update . On the other hand, if the Redo approach was selected because the organization truly wanted a fresh look at the risks of the entire process, such extensive use of the previous PHA migh t seriously compromise achievement of that goal. In any case, the revalidation team should be fully involved in developing and analyzing loss scenarios. The facilitator should ensure that any pre-populated entry is discussed, and th at additional brainstorming is allowed (and encouraged) beyond those borrowed entries. Updating the Core Analysis. The Update approach is usually selected when there are relatively few, specific changes in the process equipment or procedures, and the core methodology, existing node definitions, and risk tolerance criteria are unchanged. In that case, the revalidation approach is relatively simple and quick to apply, but it still requires thoughtful consideration. The revalidation leader guides the team through the existing documentation, soliciting the team to either confirm or correct the current information. The more detail (e.g., specific valve or instrument numbers) documented in worksheets of the prior PHA report, the easier it is to Update . (If the prior PHA lacks such detail, the team should cons ider adding it to facilitate future revalidations.) When known changes are encountered, the team appropriately edits the affected documentation. Even if there are no known changes, the team should critically evaluate each node for technical accuracy and thoroughness in identifying hazards. The previous team may have made a mistake, a change may have eluded the management of change system, a safeguard relied upon by the previous team may no longer be effecti ve, or incidents may have shed new light

Heat Transfer Units 203 Each enclosure is for each stream, one for a cooling stream and the other one for a heating stream. To refer to each of these enclosures we have to use better terminology than “closure one of the heat exchanger” and “closure two of the heat exchanger. ” In S&T HXs, a stream flows inside of tube and the other stream flows outside of the tubes or in shell. Therefore we can name the former enclosure the tube‐side and the latter the shell‐side. This applies specifically to S&T HXs. The other types of heat exchangers don’t have shells and tubes but they still have two enclosures. Table 11.2 lists the terminology for the two enclosures in different types of heat exchangers. 11.4 Different Types of Heat Tr ansfer Fluids and Their Selection As it was mentioned before we have two types of heat exchangers of process heat exchangers and utility heat exchangers, and process heat exchangers have higher priority for use in plants. We can, then, say the best heat transfer fluid is the existing process fluid and then utility fluids. Amongst cooling utility streams the best fluids are air and water, which are abundant resources. Therefore they are on the top of the list of preferred utilities. If they cannot be used the other options can be considered. If sea water is available it could be very attractive choice for cooling. A list of cooling streams is given in Table 11.3.For heating purposes the fluids on the top of list are hot water and steam. A list of heating streams is given in Table 11.4.If the temperature of the target stream needs to be increased to more than 400 °C, t he only choice is probably a fired heater. Using steam as a source of heat in heat exchangers is very common. However there are some points regarding the usage of steam as a heat transfer medium.The first point is that only saturated steam, and not superheated steam, can work as the heat providing fluid. If superheated steam is available, and it is intended to be used for heating purpose in a heat exchanger, it should be converted to saturated steam before usage. A superheated steam is nothing except a “gas, ” but it can be converted to a heating medium through a desuperheater. The second point is that it should be ensured that the  steam is completely “used” in the heat exchanger before  leaving it. The complete usage of steam means complete conversion of the steam to condensate. We have to make sure there is no amount of steam remaining in the stream exiting a heat exchanger. This can be done by a placing steam trap on outlet side of the hot fluid of the heat exchanger. Using steam as a heat transfer medium is economically justifiable when the required temperature of the heat transfer medium is less than 150 °C. When the required Table 11.2 Ter minology of twin enclosures in heat exchangers. Heat exchanger type Enclosure Name Shell and tube (S&T) heat exchanger Shell side, tube side Double pipe heat exchanger Pipe side, annular side Plate and frame (P&F) heat exchanger Hot side, cold sideSpiral heat exchanger Hot side, cold side Aerial cooler Tube sideTable 11.3 Utilit y stream choices for cooling. Cooling streams Application Cooling by air in aerial coolerWhen cooling down to approximately 65 °C is ade quate Cooling by “cooling water” or “cold glycol”When cooling down to 65 °C is not enoug h but down to approximately 40 °C is ade quate Cooling by “chilled water”When cooling down to 40 °C is not enoug h but down to approximately 20 °C is ade quate Cooling by “refrigerated water”When cooling down to 20 °C is not enoug h but down to approximately 10 °C is ade quate Table 11.4 Utilit y stream choices for heating. Heating streams Application Heating by hot water or hot glycolWhen heating up to approximately 100 °C is ade quate Heating by steam or “hot glycol”When a heating up to 100 °C is not enoug h but up to approximately 150 °C is ade quate Heating by non‐water based hot liquidsWhen heating up to 150 °C is not enoug h but up to approximately 400 °C is ade quate

2. Human performance and error 15 A Human Factors principle is that it is vital to ask how and why errors occur. This includes asking: • How an individual’s performance is influenced by the conditions they work in; • Whether the information and equi pment they have been given are suitable and sufficient; • Whether the training they have been given is sufficient; and • How an individual’s performance is influenced by the prevailing culture. An understanding of human errors and mistakes makes it possible to identify how to reduce the possibility they o ccur. Consequently, it enables the improvement of human performance. 2.3.3 Performance influencing factors and human error Many factors contribute to an individual or a team making a mistake. These include the operator’s level of experience, the complexity of a task, the clarity of operating instructions, the duration of working hours, organizational culture, as well as many others. These factors are sometimes called Performance Influencing Factors (PIFs). Some common PIFs are illustrated in Figure 2-2. Directors, managers and supervisors should identify which of these factors influence the performance of a particular task. It is then possible to create conditions to successfully carry out tasks. Later Chapters in this handbook provide advice on creating these conditions. “Performance Influencing Factors (PIFs) are the characteristics of the job, the individual and the organization that influence human performance.” UK Health and Safety Executive [4]

AN INTRODUCTION TO PROCESS SAFETY FOR UPSTREAM 7 4.This book can help upstream personnel improve their understanding and communication of the concepts of process safety management. 1.5 SCOPE OF THIS BOOK The upstream oil and gas industry is diverse. This concept book provides an overview of process safety as it applies in the upstream industry. Hopefully, this book will spur the interest in developing subsequent, more detailed, guideline texts. After this Introduction chapter, this book provides an overview of upstream operations in Chapter 2, including an introduction to safety barriers, and a short review of international regulations. This is followed by a summary of Risk Based Process Safety (RBPS) in Chapter 3, along with short descriptions for each element. The book then covers the application of the various RBPS concepts throughout upstream operations: well construction (both onshore and offshore) in Chapter 4, onshore production in Chapter 5, offshore production in Chapter 6, engineering design, construction, and installation in Ch apter 7, and future topics and research needs in Chapter 8. As noted earlier, the focus is on process safety (i.e., loss of containment events), so other major incident hazards (adverse we ather, marine events, structural failure, transportation incidents) which would require extensive discussion, are not covered in this concept level book. However, these events can be initiating events for loss of containment – e.g., the Mumbai High event in Table 1-1. The methods described in this book can be applied to these other major incident hazards as well. Similarly, occupational safety is not addressed other than toxicity or fire and explosion that can affect many people at once. Liquefied Natural Gas (LNG) can be thought of as upstream or midstream. In this book, it is covered briefly where the liquefaction occurs offshore in floating facilities (FLNG units), but not onshore in fu ll scale liquefaction plants as these are very similar to downstream facilities and thus are already covered in the existing CCPS library. A figure showing the topics which are in scope and those that are not in scope is shown in Figure 1-2. The column topics are addressed in Chapters 4, 5, 6, and 7 as indicated. 1.6 UPSTREAM SAFETY PERFORMANCE Upstream incidents are tabulated by several organizations. Offshore, individual regulators collect their own safety data (e .g., BSEE, UK HSE, PSA, etc.). They do their own reporting, but also share this in formation to the International Regulators Forum ( https://irfoffshoresafety.com ) allowing for easier comparison between regions using standardized categories. In the US, upstream onshore activity is less

8 • Emergency Shutdowns 156 8.7 How the RBPS elements apply All of the Risk Based Process Safety elements (RBPS) apply when setting up a process safety and risk management program to manage the process safety risks effectively. Effective anticipating for and activating shut-downs during emerge ncies is a result of an effective process safety program. For safe shut-downs at this time—the subject of this chapter —it is essential that the hazards be understood, the risks evaluated, and the engineering and administrative controls be identified, designed, implemented, and sustained for the life of the process. Effective proc ess safety and risk management programs are the subject of considerable guidance today, noting that the knowledge of how to identify, design, implemen t, and sustain the technologies for these emergency response programs continues to evolve. Additional guidance for applying and auditing these RBPS elements for an effective overall process safety and risk management program is provided other resources [40].

474 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Table A.4. HAZOP Worksheet Node 1 – T-1 WWT Equalization Tank Node: 1. T-1 WWT Equalization Tank Drawing Number: P&ID Example Figure A.5. Risk Matrix

10 Human performance and operational competency 10.1 Learning objectives of this Chapter This Chapter provides an overview of the Human Factors of operational (process operations, production and maintenance) Competency Management, and explains how competency leads to safer performance. The term ‘operational’ refers to the skills and knowledge such as understan ding of process hazards and how to operate and maintain equipment. By the end of this chapter, the reader should be able to: • Understand what is meant by the terms competency and Competency Manage ment. • Recognize the importance of operational competency in safety critical tasks, and error prevention. 10.2 What is competency? Competency is commonly regarded as a set of skills, knowledge, and practical experience or abilities that enable people to reliably perform tasks efficiently and safely. This includes routine tasks, and unexpected situations and changes to usual activities. Competency is also defined as the abilit y to perform work activities reliably and consistently, to the required standards. Competency can be measured against these standards. A term such as “Suitably Qualified and Experienced Person (SQEP)” can be used to indicate that a pers on is competent in their role/in the tasks they are conducting. This includes rout ine and non-routine tasks; abnormal and upset; first line emergency response; safe ty-critical maintenance, inspection and testing activities. The CCPS “Guidelines for Risk Based Pr ocess Safety” [24] cites “Process Safety Competency” and “Training and Performance Assurance” as elements. The CCPS “Guidelines for Defining Process Safety Competency Requirements” [50] provides an over view of process safety competency. This Chapter and Chapters 11, 12 and 13 build on the CCPS guidelines by providing additional insights and advice on the Human Factors of competency, learning, and Competen cy Management. These Chapters focus on operational competency. Human Factors Handbook For Process Plant Operations: Improving Process Safety and System Performance CCPS. © 2022 CCPS. Published 2022 The American Institute of Chemical Engineers.

66 Guidelines for Revalidating a Process Hazard Analysis • Relying on a prior PHA with inadequate information (e.g., incomplete or inaccurate P& IDs or operating procedures) • Relying on a prior PHA with no node or process section descriptions • Relying on a prior PHA where the team failed to identify or document all credible hazards asso ciated with the process, capture important initiating events, or address all operating modes (e.g., startup, shutdown, clean outs, catalyst changes) • Using a prior PHA where there were fundamental issues with application of the core methodology. For example, not carrying consequences to their final conclusion (e.g., stopping at "high pressure in vessel" rather than considering the possibility of "vessel rupture" and its subsequent cons equences) or not following loss scenarios to their conclusion in equipment or processes beyond the physical boundaries of the PHA being evaluated • Using a prior PHA where the team risk judgments were suspect. For example, considering ineffective or unclear safeguards when evaluating likelihood (e.g., taking credit for operator intervention, when in reality the event would develop too rapidly for the operator to respond), or failing to specif ically identify engineering or administrative controls (e.g., listing “high level switch” rather than an instrument tag number or “operating procedures” rather than a procedure number) • Using a prior PHA that had been documented by exception (i.e., only those deviations or scenar ios for which severe consequences were deemed likely or resulted in a consequence of interest were documented, such that the reviewer cannot tell whether other scenarios were overlooked, or if they were considered and discounted) • Updating a prior PHA where the previous intention was to not have any recommendations or a prior PHA with an improbably small number of recommendations indicating, perhaps, a too-cursory analysis • Using a prior PHA with no or limit ed documentation (e.g., there is no explanation of the methodology used by the prior PHA team) • Failing to consider future PHA revalidation needs when documenting and implementing MOCs between revalidation cycles, resulting in increased effort for the revalidation team and potential for unexpected results

Piping and Instrumentation Diagram Development 386 In the process plant world, winterization basically means implementing specific features in a plant design and P&ID development to prevent the impact of cold weather on plants in shutdown conditions. Winterization also can refer to activities to make a piece of equipment functional even in the harsh condi-tions of winter. Therefore in a broader sense, winterization is activities to prevent freezing, frosting, or setting of matter in a process plant in its all operation phases. The other name of winterization is “Frost prevention” or “freeze protection. ” When a plant is shut down, either partially or com- pletely, after de‐energizing the plant elements, the next step is to drain all the pipes, equipment and containers to make sure there is no trapped liquid in enclosed spaces, and that all of the enclosures are empty. This is to protect the plant during post‐shutdown time to keep the plant safe against anomalies caused by trapped liquids. Trapped water in a plant may freeze and expand, and this expan-sion of frozen water can rupture pipes, equipment or con-tainers. Trapping very heavy oil in plant enclosures will cause it to set and become hard to move. After a long shutdown, this makes the plant’s start‐up very difficult. For all of these reasons, all the trapped liquid should be emptied by the plant operator. However, the problem is that each plant may have a few hundred or more drain valves and only a small number of operators. Therefore, draining all the trapped liquid through a few hundred drain valves may take weeks to complete and during this time if the ambient temperature reaches a low level (and the word “winterization” comes from here), it may endanger the plant’s equipment. There are passive winterization protection and active winterization protection methods. The passive winterization protection methods are generally more inexpensive than active methods. The passive methods could be implemented instead of or in addition to active methods. The winterization methods are listed in Table 18.7.One passive method of dealing with winterization is minimizing the exposed area of process items. Such min-imized exposure can be obtained by putting process items indoors, inside buildings or inside sheds/cabinets or by burying underground pipes or containers below the “frost line. ” The frost line is an imaginary surface below the ground surface under which the soil is less affected by the atmospheric temperature. It is assumed that the (wet) soil below the frost line does not freeze in the winter. For fluid‐in instruments like Bourdon tube pressure gauges, winterization can be attained by essentially not sending service fluid inside of the instruments. When freezing is an issue in a specific environment and a spe-cific service fluid, specific pressure gauges with filled liquid can be used. In such Bourdon tube pressure gauges the Burdon tube is filled with a non‐freezing liquid (like g lycol) and capped with an elastic mem- brane. The membrane (diaphragm) allows pressure to transmit to the Bourdon tube without allowing the problematic service fluid getting in to the Bourdon tube. The other good practice regarding dealing with win- terization is providing sloped piping toward “tolerable” equipment. This technique is known as providing “inter - nal natural free draining. ” An example of more tolerable items against freezing are tanks. Elimination of dead legs is important if implementing the concept of “natural internal drainage. ” Dead legs are Table 18.5 P&ID presen tation of heat conservation insulation for different items. Pipe Equipment Instrument H 50 And/or in pipe tag 114–ASL–COS –100–3 8H EHS–07171–2–B 9W–50H And/or in equipment call‐outNot common Table 18.6 Pipes tha t most likely do not need HC insulation. ●Pipes go to ponds ●Pipes go to coolers (cooling stream of heat exchangers) ●Pipes of “used” cold streams (e.g. CWR) ●Short and large diameter pipes

178 Human Factors Handbook 15.4 Key learning points from this Chapter Key learning points include: • Fatigue can greatly increase the potential for error. • Causes of fatigue are many, and th ey vary by task and by person. • Established good practices help managing fatigue. • Fatigue can be monitored and managed in an operational setting.

180 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS altitude. A combination of lack of feedback from the sidestick and this loss of protection made it more likely that excessive changes to the flight control surfaces would occur. The design of the HMI did not make it easy for the crew to assimilate the information and form an accurate mental model of what was happening to the aircraft. Improvements to the HMI design including a display of angle of attack were recommended in the report. 7.1.7.2 Abnormal Situation Recognition: Clearly, as soon as the automatic systems disengaged, the aircrew were aware of an abnormal situation. Howe ver, they were unable to diagnose what was happening to the aircraft within the available time, which was a matter of some two minutes. The instruments presented them with conflicting information: the altitude suddenly appeared to decrease, but so did the speed, although the me asurements were inconsistent. Although the display includ ed an artificial horizon, it did not include the angle of attack. The action to gain height is understandable although in hindsight, this was the wrong action to take and led to the stall. The stall indicators sounded repeatedly, but th e PF did not put the nose of the aircraft down, except very briefly. Similar situations can occur in the process industries when there are sudden changes in the weather condit ions. For example, if instrument air systems are not kept very dry (typically to a dew point of -40 °C (-40 °F), moisture can freeze in lines or in mechanisms leading to erroneous readings and stuck valves. 7.1.7.3 Human Factors and Crew Resource Management: It is typical with incidents of this nature for Human Factors to contribute to most of the causes. The startle effe ct was a key factor that led to the large input to the sidestick and the ai rcraft rapidly coming out of a safe flight envelope. The sudden increase in workload led to a degradation of the communication between the pilots. The report refers to the surprise generated by the autopilot disco nnection and the loss of cognitive control of the situation. The report highlighted that the initial and refresher training provided did not adequately address this type of sudden scenario and recommended improvements in this area including reinforcement of Crew Re source Management (CRM) training and improved training simulators.

Ancillary Systems and Additional Considerations 399 is meaningless. The “design pressure” should be men- tioned “at” a “design temperature” as a pair; e.g. the design pressure of this vessel is 900 KP ag at design tem- perature of 80 °C or “900 KP ag @ 80 °C. ” For all process items a wise pair of “design pressure @ design temperature” should be selected and requested from the item vendor. Moreover, this pair should be coincident. This means that during the operation of a process item a pressure as high as the selected design pressure could happen during the time the temperature is as high as the design temperature. 18.8.1 Decision on “Desig n Pressure @ Design Temperature” Pair See Figure 18.26, which shows pressure and temperature changes in a piece of equipment during its life time. Pressure and temperature go up and down. “Bubble 1” shows the absolute highest temperature. “Bubble 2” shows absolute highest pressure. From a pure theoretical view point the pair of “design pressure @ design temperature” should be selected as the pair inside of “bubble 3. ” This is because it represents the highest pressure at the highest temperature. However, we generally and negligently report “bubble 4” as the pair of “design pressure @ design temperature” . “Bubble 4” is obviously not a coincident pair.In the next two sections we discuss decision process for selecting design pressure and design temperature. 18.8.1.1 Deciding on “Desig n Pressure” If you remember, in the level system of pressure mentioned in Chapter 5 the design pressure is a level of pressure higher than the HHP (high‐high pressure) and the HP (high pressure) and NP (normal pressure). For the selection of design pressure we can go high, very much higher than the HHP as much as we want, but this increases the cost of process elements. Therefore we need to bring down the design pressure to a level that is inexpensive while safe. The way we define the “process design” is firstly define it through a minimalistic approach and the increase is to go higher than (or equal to) the HHP . What is the minimum safe level of pressure that could be selected as design pressure? As it is attempted to keep the pressure on and con- trolled at the normal pressure (NP) one may say the design pressure could be placed at the normal pressure! However the controllers that try to bring the pressure to the normal pressure are not perfect. Control loops have an “overshoot. ” In nutshell, “overshoot” is a magnitude of deviation from the pre‐determined set point of a controller when it tries to keep the parameter at the set point. As the control technologists generally adjust an over - shoot of 10% in process plants, the design pressure could be selected as 10% higher than normal pressure as minimum. After preliminary selection of design pressure as “NP + 10%” we should c heck to make sure the design pressure is higher than the HHP and, if it is needed, increase the design pressure to make sure the rest of the pressure levels, HHP and HP , are laid down somewhere in a band between NP and design pressure. 18.8.1.2 Deciding on  “Design Temperature” As was mentioned, the design temperature in the pair of “design pressure @ design temperature” should be a coincident value. However this temperature is generally decided independently of the design pressure. This is the meaning of selecting “bubble 4” in Figure  18.26. What we did basically is selecting the highest absolute pressure and selecting the highest absolute tempera-ture and tied them “nominally” into a “pair. ” We chose this selection and not the more accurate pair of “bubble 3” just because it is easier to do that. It is very difficult to estimate the maximum upset temperature at the upset pressure. The selected pair may call for stronger, more expensive equipment but if the additional cost is acceptable, this easier methodology can be used. Bubble 1Bubble 4Bubble 3Bubble 2 Pressure Temperature Life time Figure 18.26 Pr essure–temperature pair fluctuations in a piece of equipment.

PROCESS SAFETY REGULATIONS, CODES, AND STANDARDS 49 CI, The Chlorine Institute, Arlington, VA, www.chlorineinstitute.org, http://www.chlorineinstitute.org/pu b/ed19b46c-c6a7-acca -ce7f-43f496d0dcae . Contra Costa, “Contra Costa Count y Industrial Safety Ordinance”, www.co.contra-costa.ca.us/ . CSA, “Process Safety Management Standard”, CSA Z767-17, CSA Group, Toronto, Ontario, Canada. DNREC, “Extremely Hazardou s Substances Risk Management Act”, Regulation 1201, Accidental Release Prevention Regulation, Dela ware Department of Natural Resources and Environmental Control, March 11, 2006. DOSH, Malaysia – Department of Occupational Safety and Health (DOSH) Ministry of Human Resources Malaysia, Section 16 of Act 512. DOT 2015, “Rule Summary: Enhanced Tank Car St andards and Operational Controls for High- Hazard Flammable Trains”, U.S. Department of Transportation, https://www.transportation.gov/mi ssion/safety/rail-rule-summary . EC, “Environmental Emergency Regulation s” (SOR/2003-307), Environment Canada. EC, “The Seveso Directive – Technological Disaster Risk Reduction”, https://ec.europa.eu/environment/seveso. EM 2009, “Operations Integrity Management System”, ExxonMobil, Irving, Texas, July, https://corporate.exxonmobil.com/-/media/glob al/files/risk-management-and-safety/oims- framework-brochure.pdf. EPA, “Accidental Release Prevention Requirements: Risk Management Programs ” Under Clean Air Act Section 112(r)(7), 20 CFR 68, U. S. Environmental Protection Agency, June 20, 1996, https://www.epa.gov/rmp. EPA b, Clean Air Act Section 112(r)(1) – “Preven tion of Accidental Releases – Purpose and general duty”, Public Law No. 101-529, U.S. Environmental Protection Agency, November 1990. FMG, FM Global Datasheets, https://www.fmglobal.com/fmglobalregistration . HSE, “Control of Major Accident Hazards Re gulations (COMAH)”, U.K. Health and Safety Executive, https://www.hse.gov.uk/comah/ . HSE 2015, “The Offshore Installations (Offsho re Safety Directive) (Safety Case etc.) Regulations 2015. Guidance on Regulation s”, U.K. Health and Safety Executive https://www.hse.gov.uk/offshore/safetycases.htm. IEC, International Electrotechnical Commission, Geneva, Switzerland, IEC Central Office, https://www.iec.ch/. ISA, The Instrumentation, Systems, and Automa tion Society, Research Triangle Park, N.C., www.isa.org. ISO, International Organization for Standa rdization, https://www.iso.org/home.html. Merriam-Webster, https://www.merriam-webster.com/dictionary .