Datasets:

hndc

/

AI4Hazard-small

like 0

KEY RELEVANCE TO PROCESS PLANT OPERATIONS 57 Example Incident 3.6 – Relief Valve Opening While returning from lunch one day, a unit engineer noticed a trail of condensate running in a line along the ground under the pipe rack. The engineer knew from experience that such a trail is frequently a sign that a light material such as LPG is vaporizing in a pipe (in this case a flare header), which cooled the pipe and caused atmospheric moisture to condense. As he followed the trail back to the source (an LPG treater), he noticed an operator at the top of a vessel actively blocking in a relief valve. When asked, the operator replied that he was blocking in the relief “because my supervisor said it is making the flare too large.” The unit engineer stopped him before a disaster occurred, but this illustrates those issues such as culture and hypersensitivity to events that are visible to outsiders can motivate workers to take inappropriate and potentially dangerous actions. Lessons learned in relation to abnormal situation management: Understanding abnormal situations/process safety competency: In this case, the operator (and his supervisor) failed to consider the potential consequence of bl ocking in the relief valve. Management of Change: Temporary or permanent changes to any protective device should be evaluated through the MOC process. Procedures: The procedures (and associated training) for diagnosing a faulty relief valve should include requirements to: o Check on other pressure gauges in the system, accounting for difference in head (height of fluid) o Replace the pressure gauge to ensure it is working correctly o Check pressure control system for correct functioning o Obtain senior management approval before isolating any relief system, via structured and formal temporary bypass procedure

7.3 Verbal–Linguistic Intelligence | 95 Figure 7.2 Simple Poster About Response to a Person-Down Incident 7.3 Verbal–Linguistic Intelligence Written incident reports and bulletins are the most common ways that findings from process safety incidents are captured and communicated. Writing the report may be as instructive to the members of the investigation team as it is to the people who later read that report. Larkin suggests several techniques to improve verbal-linguistic learning (Larkin 2018): • Write in simple, direct sentences. This can be measured via an option in MS Word spell-check. Target a Flesch-Kincaid reading level of 8 if possible. • Keep line length short. The 4.5-inch line length in this book is slightly longer than the 3– 4 inches Larkin recommends. • Keep paragraphs short. Short paragraphs hold readers’ attention better than long ones. • Use a sans-serif font. The font in this book is one example. Serif fonts can reduce the efficiency of learning from written messages.

240 Other IS considerations dealt with primarily at the design stage include: Designing vessels to withstan d the maximum temperature and pressure to which they are likely to be subjected in operation. Designing control valves to limit the maximum flow/temperature/pressure that downstream equipment can withstand. Use of seal-less pumps and designing pump spillback loops to reduce risk associated with deadheading. Selection of materials of construction that are compatible with the process fluids to be handled and the chosen design conditions, in order to minimi ze corrosion/erosion potential rather than relying on inspection programs to identify and correct. Minimizing the size of vessels, piping and other equipment to limit the potential release quantity. Minimizing the number of flanged and threaded joints and hoses/flexible connections which present increased leak potential. Addressing issues associated with reactive chemicals, such as eliminating the potential for wate r contamination of the process if a water-reactive ma terial is being used. In most cases, an assessment of alternatives will involve consideration of cost—both initial an d ongoing, such as inspections, testing and maintenance—and other fa ctors, including relative risk reduction using other risk management approaches, such as emergency relief/mitigation system vs. higher vessel design pressure. For complex, technically challenging, or inherently higher risk projects, a formal risk comparison should be carried out using tools such as ranking methodologies or quantitative risk assessment. Ankers (Ref 10.2 Ankers) describes a software application for identifying inherently safer process options, with an emphasis on early hazard identification. Other methods are discussed later in this chapter.

APPENDIX E - CLASSIFYING PROCESS SAFETY EVENTS USING API RP 754 3RD EDITION 489 Community Impact. LOPCs that result in an officially declared community evacuation, including precautionary evacuation, or comm unity shelter in place are Tier 1 events. Direct damage from fire/explosion. Direct damage cost includes the costs to repair equipment or replace it and the costs of clean up and environmental reparations. The fire or explosion must be a result of an LOPC. Direct cost greater than or equal to $100,000 USD are classified as Tier 1 and cost greater than or equal to $2,500 up to $100,000 USD are classified as Tier 2. Releases from engineered pressure relief and upset emissions from permitted sources. Engineered pressure relief devices include pressure relief device (PRD), rupture disks, Safety Instrumented System (SIS) devices, or manually initiated emergency de-pressure devices. LOPC from these devices are excluded from Tier 1 or Tier 2 classification if the release is as design and proven safe. Safe release ca n be defined as the maximum concentration of release below ½ the LEL for flammables, or ERPG-3 for toxics at potentially occupied locations around the release. However, if the amount of discharge is greater than or equal to the threshold quantity (described in Section E.2.3) in any one-hour period and results in any one or more of the following consequences: rainout; discharge to a potentially unsafe location, an on-site shelter-in-place or on-site evacuation, excluding precautionary on-site shelter-in-place or on-site evacuation; and/or public protec tive measures (e.g. road closure) including precautionary public protective measures. These cr iteria are also applicable for a permitted or regulated source. E.2.3 Release Quantity Criterion If the discharge contains one of the four conseq uences described in Section E.2.2, the tier of the release is determined by the re lease quantify described as follows. 60 Minute Release. To have a PSE Tier 1 and 2 classification, the acute TQ threshold (described next) must be exceeded any 60-minut e window of the release. For a steady state release, the release amount is normalized to the amount rele ased over the 60-minute period. Source models must be developed for non-steady state releases. Acute release above the th reshold quantity (TQ). API classifies materials in eight Tier 1 and Tier 2 threshold release categories. The rele ase categories recognize the relative potential hazardous consequences of release; the higher the potential consequence, the lower the TQ. The category determining the TQ is listed as toxic (toxic inhalation hazard, or TIH Zone), flammability (boiling point and flash point) or co rrosivity. When a material can’t be classified by these characteristics (in this order), the United Nations Dangerous Goods (UNDG) Packing Group is used. UN Packing Groups can usually be found in section 14, Transportation, of a materials SDS sheet. (Refer to Section 7.3.2) Indoor v. outdoor release. The Tier 1 and 2 release categories are further subdivided for outdoor and indoor release. The lower indoor quantity accounts for a potentially greater hazardous consequence associated with indoor release. Table E.2 shows the relationship between Tier 1 and 2 outdoor and indoor threshold quantities (TQ).

82 Human Factors Handbook Figure 8-3: An example decision fl ow chart for unresponsive casualties (Based on a compilation of several resources) 8.3 Navigation For larger job aids, such as manuals and lo ng procedures, it is important to support navigation to help the reader find the information they want. Ways to support navigation include: Headings Headings should be large and stand out from the surrounding text. This will help people to identify information. Th ey can also be color-coded in terms of information, or to indicate the level of the heading. Indents and tabs The first page of each sections can be indented or indicated by a tab. Color-coding To help the reader find and identify re levant information color can be used to code and group text. For example, sub-sect ions can be color-coded as per Figure 8-4. As some people are color blind, it is important also use other means of distinguishing information, for example, by the use of icons (see section 8.6). Some examples are also shown in Figure 8-4.

A.3 Index of Publicly Evaluated Incidents | 207 Section 3: Selected Causal Factors (Continued) Facility Siting—Secondary Findings C1, C14, C24, C52, C53, C63 HA10 J109, J129, J150, J157, J233, J239 S4, S7, Gap in a Standard—Primary Findings C4, C9, C22, C29, C30, C42, C43, C44, C45, C56, C64, C65 D7 J84, J91, J117, J232 S17 Gap in a Standard—Secondary Findings C6, C34, C38, C39, C54, C70, C72, C74 D19, D42 HB3 J6, J40, J81, J206, J207, J218, J227, J237 S15, Human Factors—Primary Findings A7, A11 C34, C46, C60, C71 J1, J11, J13, J14, J50, J52, J55, J63, J65, J66, J75, J76, J81, J105, J107, J110, J132, J134, J140, J157, J158, J165, J166, J171, J174, J176, J177, J178, J179, J180, J181, J182, J185, J187, J188, J189, J190, J196, J197, J200, J204, J213, J217, J230, J234, J251, J252, J253, J254, J271 S2, S15 Human Factors—Secondary Findings A2, A4 C10, C12, C13, C27, C38, C49, C51, C68, C76 D9, D32 HA3 J23, J28, J30, J35, J40, J41, J43, J45, J48, J61, J97, J108, J115, J128, J130, J131, J145, J147, J148, J149, J167, J192, J194, J221, J233, J236, J237, J256, J257, J270 S5 Reactivity Hazards—Primary Findings C1, C3, C7, C12, C24, C32, C36, C49, C61, C62, C68, C73 J1, J5, J6, J7, J21, J24, J25, J30, J31, J32, J35, J38, J42, J43, J47, J51, J56, J66, J68, J69, J71, J72, J74, J77, J78, J83, J90, J92, J97, J98, J99, J100, J101, J104, J112, J113, J116, J118, J120, J121, J123, J124, J125, J126, J129, J130, J131, J132, J133, J135, J142, J143, J144, J145, J146, J147, J148, J149, J150, J151,

249 4.During the inherent safety revi ew, at the design development stage, identify potential human factors/ergonomics issues that should be addressed by the design team. 5.Document the review and follow-up items. Figure 10.3: Inherent Safety Review Process During the systematic review of the process flow schematic the team will examine some of the following questions: Can safer chemicals be used (i.e ., non-toxic/non-flammable or non-volatile reactant)? Can quantities be reduced (particularly intermediate requiring storage)? Can potential releases be redu ced via lower temperatures or pressures, elimination of equipment or by using seal-less pumps?

Piping and Instrumentation Diagram Development 188 How can we resolve this problem? The problem used to be resolved by application of a stuffing box and gland. In that solution, a small space is provided around the hole and donut shaped “packing” is placed inside of that. The packing in the first generations of stuffing box was a ribbon of felt. Therefore, felt from one side prevented the leakage from the other side; it provided enough freedom for the shaft to rotate. The problem with a stuffing box and packing is that there is usually some leakage from the system. The other problem is that it needs frequent checking and inspec - tion to make sure the packing is healthy. This solution is still available for some pumps where the rotation speed of the shaft is not very high or the ser - vice is not very critical. It is interesting to know that valve stems are still handled by stuffing box and packing. The second solution is a “mechanical seal” which is a more robust system. From a very simplistic viewpoint, a mechanical seal is two blocks of graphite sliding on each other. Both of these blocks could be in the form of donuts. One donut is fixed in the pump casing and around the hole and the second block of graphite is attached to the shaft at the penetration point. During operation of a fluid mover, these two blocks are sliding on each other. The features of the graphite blocks are: (i) they have very smooth surfaces and (ii) they inherently have some lubricity features because of the nature of graphite. Therefore, on the one hand, because of the very low roughness of the graphite blocks surfaces, there is almost no leakage from the system. On the other hand, because of the oily feature of graphite blocks, the shaft can easily rotate in the hole. In reality, a mechanical seal is more complicated than the above and has several other elements. The two sliding blocks plus other mechanical parts around them is named a mechanical seal. These days, there are more complicated mechanical seals that are constructed of different materials and sometimes graphite blocks are replaced by ceramic blocks, however, the fundamental operation remains the same. Mechanical seals are a very critical part of every fluid mover and need frequent inspection. However, a mechanical seal is not something that can be seen on a P&ID so why should we care about that in P&ID development? Even though there is no footprint of a mechanical seal on a P&ID each mechanical seal should have a supporting system and the supporting system should be shown on the P&ID. The supporting system for a mechanical seal is a system to flush, cool down the sliding blocks, and also lubricate those parts. Basically, the problems associated with mechanical seals are because there are sliding blocks, there is heat generation in sliding block surfaces, and this generated heat should be dissipated. By providing some sort of lubrication, the generated heat can be decreased. The supporting system that provides sealing, cooling, and lubrication for a mechanical seal is named a “seal flush plan. ” There are plenty of different arrangements to provide cooling, flushing, and lubrication for mechanical seals. The concept of sealing systems is applicable for all fluid movers. However, here we continue our discussion only for pumps. More discussion on sealing systems of gas movers is in Section 10.8. There are different arrangements for sealing systems of pumps, which are standardized for the sake of decreas - ing the price and also facilitating fabrication and inspec - tion. There is a standard generated by the American Petroleum Institute under the name of API‐682 that introduced more than 30 different plans as supporting systems for mechanical seals. Usually, the mechanical engineer in each project decides on the selected type of seal flush plan based on standard plans stated in stand-ards or a combination of a few of them, or even totally non‐standard arrangements. Deciding on a specific mechanical seal flush plan is done through the flow chart that is provided in the standard. In standard API‐682 each seal flush plan is specified by a number. Therefore, the seal flush plan could be introduced by seal flush plan number 13 or seal flush plan number 15 by referring to that specific standard. Each seal flush plan needs a stream of fluid for the purpose of cooling, flushing, and lubrication. This flow stream can be provided from the process fluid, “the fluid that is being pumped, ” or from an external source. The external source fluid can be plant water or some other type of fluid. The final destination of each flushing/cooling/lubricating fluid could be the process fluid or other destinations. To make sure that we have enough flow of Figure 10.26 Loose shaf t casing touch.

98 INVESTIGATING PROCESS SAFETY INCIDENTS 6.3 LEADING A PROCESS SAFETY INCIDENT INVESTIGATION TEAM An effective incident investigatio n management system, as described in Chapter 4, depends on many factors, driven by managem ent’s commitment, support and actions. A management sy stem that provides strong team leadership and organizational support will help the investigation team to succeed in understanding what happened, determining causal factors and root causes, developing plans to prevent recurrence, and sharing learning both inside and outside the company. The selection of the team leader for an incident investigation will depend on a number of factors related to the incident, including: 1. Its actual (or potential) severity and complexity; 2. Its health, safety, environmental, or business interruption implications; 3. The anticipated complexity of the investigation. Various approaches for determining the scope and size of the incident investigation are described in Section 6. 5 and Figure 6.1, and the choice of team leader will be a function of the scale and type of incident. However, the general abilities of the investigation leader will be similar and should include the following comp etencies and qualities: Leadership abilities and experience, including process safety experience Communication skills with all levels in the organization and other stakeholders (verbal, written and presentation) Problem solving/ logical and systematic thinking Objectivity Planning and organization/ administration Commitment to safety Technical incident investigation skills Conflict management skills and experience Ability to handle information with confidentiality and sensitivity The selection and training requirements for the team leader and the rest of the investigation team are detailed in Section 6.4. Investigation team leaders should be identified and trained for the appropriate investigation types or tiers to which they have been assigned. Ideally, the team leader should be independent of the incident itself, although this may not always be possible or practical, particularly with a lower tier investigation. For

Evaluating the Prior PHA 49 • Logic errors and inconsistencies in the analysis • Failure to document hazards • Improper application of the risk tolerance criteria 3.2.1 Application of Analysis Method(s) Chapter 5 of the CCPS book Guidelines for Hazard Evaluation Procedures [2] specifies how to apply various PHA methodologies and the results for each method. Chapter 6 in that book covers factors for deciding which method to use in a particular situation. The discussion of PHA methods in this section is limited to the most commonly used core PHA methods: What-If/Checklist, HAZOP, and FMEA. It focuses on their use in the PHA revalidation step (which is the most common form of PHA activity through- out the “operate and maintain” portion of the facility life cycle). It excludes their use for various hazard analysis activities conducted early in the life cycle before the initial PHA (e.g., during preliminary hazard identification or design review activities) or late in the life cycle (e.g., during mothballing, de- commissioning, or demolition activities). In practice, management must ultima tely decide which analysis method provides sufficiently rigorous informatio n to support their risk judgments and resource allocation decisions. Process complexity and process hazards are often cited as the two main reasons for using more structured methods. Example – Improper Core Analysis Method A PHA due for revalidation is being reviewed. The reviewer observes that the core analysis is primarily an industry checklist specific to truck/railcar unloading of a hazardous chemical feedstock. The core methodology used in this study (checklist alone) was likely inadequate. While the checklist performed is useful, it is supplemental to the insights that would be gained using the What-If/Checklist combination or HAZOP techniques.

8 • Emergency Shutdowns 145 to people and equipment. Fires can prevent emergency responders from being able to access the area af fected by the fire. Explosions can destroy the emergency response equipment due to damaging overpressures, such as fire monitor s and fire water supplies, as well. An example of an explosion severely damaging the emergency response equipment is provided elsewhere [81]. For the purposes of this guideline, there are four general types of abnormal situations that warrant quick, emergency shut-down, usually in this order based loosely on magnitude of the loss event: 1. When the Safety Instrumented System (SIS) or Emergency Shutdown System (ESS) are activa ted (before a potential loss event); 2. When the SIS or the ESS are activated (after loss event); 3. When there is a loss event (but no activation of the ERP); and 4. When a loss event activates the ERP. These four types of emergency shut-downs are shown Figure 8.1. Once the shut-down has been completed, there are two general conditions for the process equipment’s end state. Both of these end states are described next: the first is when the emergency shut-down ends at the normal shut-down end sta te (Section 8.4.1); the second is when the end state is different from the normal shut-down end state (Section 8.4.2).

D.1 The HRO Concept |279 Redundancy : B ack-up systems, cross-checking of safety- critical decisions and the “buddy system” in which staff observe each other to catch and head-off errors . Deference to expertise : Responsibilities are clearly defined for norm al operation and decision-making is hierarchical. However, in emergencies, decision-m aking shifts to individuals with expertise regardless of their position in the organization.Empowerment : Management-by-exception is practiced. Managers focus on strategic, tactical decisions and only get involved with operational decisions when required. Preparedness : Well-defined procedures for all possible unexpected events. HROs effectively anticipate potential failures. HRO leaders intentionally engage with front-line staff to remain sensitive to the challenges of day-to-day operations. They remain attentive to what CCPS term s “Catastrophic incident warning signs” CCPS (Ref D.3), trivial signals that may be early indicators of emerging problems. They take warning systems, as well as performance m etrics seriously, and are slow to dism iss them or explain them away. HROs are reluctant to oversimplify. While they understand that simplicity in design is good, they also know that their operations are inherently complex. Therefore, deep understanding is required to adapt to day-to-day challenges. In such complex systems, they recognize the need to understand the systemic causes of incidents, rather than placing the blam e on the operator. HROs have a “just” culture and foster a sense of personal accountability for safety. They have systems that m ake it easy to report near misses and incidents without fear of punishm ent, and they give all employees stop-work authority. They follow-up incident investigation by implementing corrective actions. • • • •

15.6.4 Simplification The Bhopal facility relied on end-of-p ipe monitoring and control systems attached to the storage v essels, the reliability of which required they be Figure 15.12. Alternative route for carbaryl production (Ref 15.4 Crowl) in good working order. Reliance on such systems can create a dual problem: the safeguards may not be available when needed, and their existence may provide a false sense of security for process operators who may ignore initial warning sign s, such as a pressure increase. Many of these issues are brought into focus when one analyzes a system through use of the simplification principle. 15.7 EXAMPLE: INHERENTLY SAFER PROCESS FOR PRODUCTION OF TRIALKYL PHOSPHATE ESTERS The traditional route to manufacturi ng alkyl phosphate esters involves reacting phosphorus with chlorine to produce phosphorus trichloride, oxidizing the phosphorus trichloride to phosphorus oxychloride, then reacting phosphorus oxychloride with an excess of the desired alkanol, 421

308 INVESTIGATING PROCESS SAFETY INCIDENTS Table 13.3 Example Checklist for W ritten Reports 13.5.2 Avoiding Common Mistakes For improved quality of written incident investigation reports, the incident investigation team should follow these guidelines: 1. Avoid jargon specific to the proc ess that the intended reader may not understand. One good guide is to ensure that the written report is understandable to the inten ded reader who does not have detailed knowledge of the sp ecific process involved. 2. For increased readability and comprehension, limit the use of abbreviations and acronyms. Most of these can be avoided. Define each abbreviation and acronym used. 3. Decide on the selected reader’s le vel of technical competence and then be consistent in writing to that level. Assume the reader has a certain minimum knowledge of the chemical process industry. 4. Avoid intermixing opinions, speculation, and other judgements when presenting the factual findings. Th e report should convey the factual basis for the causal factors and root causes. The investigation team may have to make judgements and identify probable and possible causes when data are insufficient fo r a definitive determination. The report should clearly indicate when judgments were made. Separating CHECKLIST FOR W RITTEN REPORTS Intended reader/user identified and technical competence level chosen Purpose of report identified Scope of investigation specified Summary/abstract length is no longer than one page Summary/abstract answers what happened, why, and general recommendations Background—describes process, investigation scope Sequence of events—clearly describes what happened and timeline Findings—factual findings are presented Causal factors – what happened is determined Root Causes—identifies multiple and underlying causal factors Recommendations—describes specific action for follow-up Other—necessary charts, exhibits, information Content agreed to by team members Distribution identified

EQUIPMENT FAILURE 203 precooled solvent medium above its flash point. Because the Teflon® coating on the centrifuge basket had been worn away, ignition of the fl ammable mixture could also have been due to metal-to-metal contact between the basket an d the bottom outlet chute of the centrifuge, leading to a friction spark. A static discharge mi ght also have been responsible for the ignition. Since the incident, the company has required use of nitrogen inerting when centrifuging flammable liquids at all temperatures. (Drogaris) Lessons learned include monitoring the oxygen concentration in conjunction with inerting and sealing the bottom outlet to minimize air en try. Because the ignition source was uncertain (static discharge, frictional heat), this incident illustrates why it often is prudent to assume an ignition source when designing for flammable materials. In his book, Lessons from Disaster: How Organizations have No Memory and Accidents Recur , Trevor Kletz is quoted as saying “Ignition source is always free.” (Kletz 1993) Example 2. An explosion occurred in a dust collector connected to the process vents for polyester plastic extrusion equi pment. The explosion was safely vented. The bags were off their cages and charred, as show n in Figure 11.18. The dust colle ctor housed 144 cages. During the investigation it was found that one of them was not grounded. A check of an identical unit revealed problems with the grounding between the cages and the tube sheet, as shown in Figure 11.19. It was also found that the type of bag used had been changed, but the need for a grounding strap was not understood, and that grounding/bonding checks were not part of the asset integrity program. The type of bag used was changed to ones with an improved grounding strap to ensure grounding of the metal cage support inside the bag, and the procedures were fixed to include conductivity checks (Garland). Figure 11.18. Damage to dust collector bags (Garland)

PROCESS SAFETY AND MANAGEMENT OF ABNORMAL SITUATIONS 17 Example Incident 2.1 – BP Amoco Polymers ( cont. ) The molten pre-polymer entering the polymer catch tank was about 620 °F (327 °C), resulting in thermal decomposition in the absence of cooling. This decomposition caused polymer foaming and expansion, which flowed into the unheated conne cted piping and solidified. The vessel relief, vent, and drain lines were found to be full of solid polymer when the unit was dismantl ed, which explained the presence of significant pressure in the polymer catch tank that was not apparent from checking pressure gauges and opening drain valves. The blocked locations are shown in the shaded areas in Figure 2.5. Lessons Learned in relation to abnormal situations: 1) For management/engineers: Polymer decomposition wa s not fully understood, identified, or adequately discussed in HIRA studies and operating procedures. The polymer catch tank design was unsuitable for the dirty service of waste polymer. 2) For supervisors / operators / technicians: Operations and maintenance personnel were unaware of hazards regarding polymer thermal decomposition. Absence of a formal maintenance procedure for cleaning the polymer catch tank. A startup procedure to test the extruder was not followed. Operational readiness review (a ka pre-startup safety review) should be conducted before every startup.

208 | 6 Where do you Start? surveyed, compared to verbal surveys where interviewers might interpret responses differently. Written surveys can also collect information from a wide group of people in less time than individual interviews, and if conducted anonymously they can help solicit more honest and accurate inform ation. Finally, employees may hesitate to ask an interviewer to repeat questions that they could easily reread if questions are written. If the survey questions are multiple choice, avoiding open- ended questions, the responses can be statistically analyzed, especially if they are conducted online. However, the multiple- choice form at does not allow respondents to convey emotion, which is better detected via face-to-face interviews. In face-to-face interviews, interviewers can also ask follow-up questions when they sense there is more behind an interviewee’s response. The best way to elicit the depth of the concern is to use a scale for responses from strongly agree to strongly disagree. Therefore, it m ay be valuable to use the multiple-choice on-line survey initially, and then develop focused follow-up questions to be explored through interviews with targeted groups. Perform Individual Interviews A meaningful culture interview requires an examination of values and behaviors. Culture interviews therefore resemble interviews perform ed in PSMS audits, especially in that they: Follow question protocols to be developed and followed, B enefit from developing rapport between interviewer and interviewee, Should include thanking the interviewee for their participation; and Should be documented. However, interviews m ay be more challenging than audits because they involve assessing feelings rather than objective facts. Culture interviews also differ from audits in that they should involve a selection of people representing a range of job functions • • • •

116 INVESTIGATING PROCESS SAFETY INCIDENTS other distractions. In most cases, the witnesses are not trying to provide false data; they are usually trying to provide an account of what happened as best as they can recall it. In some cases, witnesses may be emotionally upset after incidents that were particularly serious. These human performance characteristics are often at the root of apparent inconsistencies and conflicts generated from comparin g witness testimony. Figure 7.3. Illustration of Human Observation Limitations Although humans have a remarkable capacity to observe, interpret, recall details, and then articulate this information, humans are not computers. No single witness has a co mplete view or comprehension of the entire occurrence; each person experiences a unique perspective. Discrepancies in descriptions of th e incident may be due to different perspectives or even differe nt experiences of the indi vidual witnesses. In one way, this concept could be comp ared to each witness seeing an instantaneous vertical slice or “snapshot” view of a large, moving, panoramic occurrence. All incoming information is processed and filtered by the brain as part of the cognitive comprehensio n process. The information is again processed and “filtered” as it is ar ticulated and transm itted to others.

253 Figure 10.4. Applicability of inherently safer design at various stages of process and plant development.

EVIDEN CE IDEN TIFICATION , COLLECTION & M ANAGEM ENT 153 Table 8.3 Examples of Electronic Data • Backup of data from control system such as distributive control system (DCS) and programmable logic controller (PLC). Data includes historic trend data, set points, measured values, trends, event logs, etc. • Configuration files from control system, including range, alarm settings, units, etc. • Data from interlocks/ Safety Instrumented Systems including event logs, activations, overrides, etc. • Any electronic records that replace paper systems as detailed in Table 8.1, such as maintenance records, permit to work, MOC, etc. • Security camera video (on site and from neighboring sites) • Email records from operations, maintenance, and management • Data from personal electronic devices (see Chapter 7) • Telephone and text records • Gate/building entry/exit records • Newscasts showing footage of the incident. 8.2.5 Position Evidence and Data Position data is the last of the five data types and is often linked to people data (See Chapter 7) and physical data . Position data may help answer the following typical questions. • What failed first? • Where did the fire start? • Where was the pressure the highest? • How far did an object travel? • Where was the witness at each point during the incident? • How far apart are the two items? • Which gaskets failed and which did not? • Is the distance between the scratches the same as the distance between the protruding bolts? Examples of position data ar e provided in Table 8.4 below.

10 • Risk Based Process Safety Considerations 200 did not exist or were not understood and implemented. An effective emergency management program includes regular emergency response plan drills and debrie fing sessions. Additional emergency management guidance is provided in other publications [14, p. Chapter 18] [21, p. Chapter 11]. 10.3.4 Additional reflections on the asset integrity and reliability element As noted in Section 10.3.3, som e incidents occurred during the transient operating mode were due, in part, to weaknesses in the asset integrity and reliability program. Many of these incidents occurred due to compromised equipment that did not perform as expected— the equipment was not fit for use. No matter when the equipment failure occurred, the failure of equipment that was properly designed, fabricated, installed, and ope rated is due to it not being maintained through a robust ITPM program on both the preventive and mitigative engineering controls. (More discussion on the overall equipment life cycle was provided in Chapter 9.) Strong leadership support of resources to manage ITPMs is crucial, as inadequate maintenance can be attributed to and have led to some of the worst industrial incidents, with some noted specifically duri ng start-ups due, in part, to unreasonable maintenance sc hedules [101, pp. 124-125]. Unfortunately, some incident re ports noted that the Process Hazard Analysis (PHA) had simply “failed to identify” the causes of the incident, and thus, the “unexpected” or unanticipated failure occurred and there were insufficient safegu ards in place to protect personnel, the environment, or property from harm. Since several incident reports noted the HIRA as a weak element without considering what a PHA Team should assume during its review, this element appeared to be a much larger contri butor than it probably is . This viewpoint tended to skew the results more to the in effective HIRA RBPS element as shown in Figure 10.2. No PHA team can anticipate every combination of

13.8 Embed and Refresh | 177 approached the group and noticed a puddle on the floor. It smelled like hexane. “What should we do?” one worker asked. “What do you think we should do?” he replied. “They weren’t serious about shutting down to fix it, were they?” another worker asked. “One way or another, you’re going to find out,” Rakesh replied. “But if I were you, I’d shut down and fix the leak.” This scenario played out several times over the next few months. Typically, the problem was a flange that had been improperly torqued during construction or the installation of the wrong gasket. One time a supervisor told the workers to hang a bucket under a leak and keep running. When Prasad saw the bucket, he reprimanded the supervisor and observed him closely for the next few weeks. Six months after restart, Rakesh deployed the new hazard analysis procedure. It included a short on-site version for use in issuing work permits, a more in-depth version for use in MOCs, and a detailed version for formal PHA. Several pilot sessions were held for each version, during which Rakesh sought input from the workers and supervisors. One year following restart, the president of Medjool, Inc., paid a surprise visit to the Chana plant. He and Samir had arranged for a production stand- down during the visit to celebrate all of Chana’s improvements. During a lunchtime feast, the president walked around with a videographer. He asked some employees to say a few words about their commitment to following procedures, analyzing hazards, or process safety in general. He asked others to talk about what working at Chana used to be like. The videographers quickly stitched together the video clips into a “Before and After” video which was played later in the afternoon. The following week, a series of posters showing different workers and their safety quotes appeared around the plant. These were greeted with much celebration.

5 • Facility Shutdowns 84 5.6.1 Case study with no process safety-related incidents C5.6.1-1 – Dolph in Energy Limited (DEL) [44] Project’s Years : 2003-2010 Cause of the process shut -down: A major facility shutdown The project covered seven year s from concept development and preliminary engineering to full operations, involving one onshore sour gas processing plant and two offshore platforms, each with twelve wells and a separate 80 km (50 mile), 0.9 m (36 inch) underwater supply line from the platform to the onshore sou r gas plant. The project also included the 364 km (226 mile) underw ater pipeline from the onshore plant export gas plant to receiv ing facilities through an onshore distribution pipeline/network. There were two major project phases associated with the commissioning and safe start -up of the equipment (Table 5.1). Phase one ended with a Warranty Shutdown in 2009 that lasted thirty days; phase two, a Maintenance Shutdown in 2010, lasted eleven days. At the project’s peak manpower times, there were 986 people in the first phase and 272 people in the second phase. The num ber of work permits exceeded 1,000 for phase one and 300 permits fo r phase two, with 520MM hours contractor time logged for phase one and 79MM hours logged for phase two. The first phase invol ved coordinating the shutdown- related efforts with the units that were not a part of the shutdown. These Simultaneous Operations (SIMOPS) were safely executed due to rigorous communications protocols established through the project’s Operations Shutdown Team (Ops SD Team). Tracking of i ncidents during the shut-down (or start-up) : There were no process sa fety incidents, no lost -time injuries, and no medical treatment cases during the seven -year effort (Table 5.2).

194 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS screen would have made it easier to spot no flow downstream of the debutanizer. While the DCS system allowed different levels of alarm prioritization, this feature was not used effectively. Consequently, alarms that should have had different priorities as signed were presented in the same place, leading to excessive alarms and making it di fficult for operator s to understand the relative importance of one alarm versus another. This is what we now refer to as “alarm management”, including rationalization and prioritization as well as how the alarms are presented in the HMI. A key action from the investigation wa s to identify safety-critical alarms and to present them to the operators so that they are distinguishable from less important, operational alarms. Fu rthermore, the report states that ultimate plant safety should not rely on an operator response and should thus require an automated response of suitable integrity, based on an appropriate hazards and risk analysis. 7.2.8.2 Abnormal Situation Recognition At the time, the operating staff were handling several abnormal situations, including a fire on the crude unit and process upsets caused by the power supply interruptions. However, it is apparent that they did not spot the crucial problem associated with the hi gh level in the debutanizer arising from its closed outlet valve. Chapter 3, section 3.1.2.1, refers to the contribution instrument failures can have to abnormal situations. These types of failures can make diagnosis difficult, and in this case, problems were identified with two valves as well as anomalies with level indication. The valve FV-385 feeding the deetha nizer appears to have closed when it was manually adjusted to 36% open by the operator. The outlet valve from the debutanizer, FV-436 closed automatically to maintain level but then stayed closed, even though the indicator on the DCS showed that it had opened. The level indicator on the debu tanizer reached 79% and stayed there throughout the incident. Many of the level gauges on the plant were of the differential pressure type and these can be inaccurate, particularly if the density of the fluid changes. This issue

270 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Figure 13.3. Logic diagram for consequence models for volatile hazardous releases (CCPS 1999)

Containers 153 A big mistake is sometimes made by piping modelers in that the containers are modeled based on their P&ID symbol rather than the container datasheet. Possibly the only exception to this general rule is for nozzles on a tank shell or on a tank roof. If a nozzle should be on a tank roof it should be drawn on the tank symbol in the P&ID and if a nozzle should be on a tank shell it should be drawn on the tank shell in the P&ID. Here lack of room to show a nozzle on a shell is not a good excuse to put a shell nozzle on the roof of the tank. Tanks have roofs and shells, and their nozzles could be on the roof or on the shell. We like to install the nozzles in a location that is easy for monitoring and inspection. It is easy to recognize that nozzles on the shell are more accessible than nozzles on the roof. Plant operators don’t like roof nozzles because they need to take steps up to be able to inspect them. Therefore the general rule is to put all the nozzles as much as possible on the shell of a tank. However, there are some nozzles that cannot be installed on the tank shell. They are for example vent nozzles, PSV nozzles, VSV nozzles, blanket gas and VRU nozzles (discussed at the end of this chapter) and Nozzles for none‐contact type level sensors. The none‐contact type level sensors will be discussed in Chapter 13. If there are nozzles to be placed on the shell, it is best to put them on the lower part of shell, which is more accessible. If there are nozzles to be placed on the roof, generally they are grouped into one area to be accessible by installing only one catwalk over the roof. However, these requirements cannot be met in all cases. For example overflow nozzles are always located at the top of the tank’s shell. Contact type level instruments should be placed in the middle of the shell where the liquid level is. Table  9.7 summarizes the features of shell and roof nozzles.The general location of some nozzles is listed in Table 9.8. 9.9.3 Nozzle Ele vation Versus Liquid Levels In non‐flooded liquid containers it is important to note the relative elevation difference between nozzle el evation versus liquid level in the container. Table 9.7 Nozzle loca tions for tanks. Nozzle on shell Nozzle on roof Schematic Features ●More accessible ●Chance of liquid leakage ●More prone to clogging ●Less accessible ●Leak‐proof against liquid ●Could be more prone to corrosion Examples By default choice OthersTable 9.8 Nozzle loca tions for tanks. Nozzle Name Location Process fluid nozzles Preferably below LLLL Instrument nozzles Lower than LLLL, except level sensor Manway Different locations Overflow nozzle On the top of shellThief hatch On top or roof Pressure protection nozzleOn top or roof Vacuum protection nozzleOn top or roof Free vent nozzle On top or roof Clean‐out doors On the bottom of shellDrain nozzles On the bottom of shell Truck‐out nozzles On the bottom of shellHand hole Within the reach of operator Drain and vent size On the bottom of shellSting nozzle On the bottom of shell Steam‐out nozzle On the shellPurge nozzle On top or roof View or inspection port Within the reach of operatorMedia adding/removal On the bottom of shellCathodic protection On the bottom of shell

52 INVESTIGATING PROCESS SAFETY INCIDENTS 4.1.2.1 N otification Notification should follow the company protocol to report details of the incident internally or externally to spec ific individuals or organizations. The circumstances of the incident and the progress of the investigation should also be communicated via the compan y’s incident reporting system and/or the incident management/emergency response system. The term report can have several meanings. Sometimes, the term could mean a verbal initial notification or communication to alert the organization that an incident has occurred. The term also refers to the final, formal written incident investigation report. The term report still causes confusion when discussing regulations. For example, US OSHA requires a notification report within 8 hours following a fatality incident or within 24 hours for in-patient hospitalization, amputation, or eye loss. The reporter should make a written documentation of any verbal communication, noting the time, person involved, extent of information disclosed, and any special instructions or requests made by US OSHA at the time of notice. US OSHA further requires a written report for each work-related injury or illness that is severe enough to be recordable (29 CFR 1904, OSHA, 2000). Many regulatory agencies require immediate notification . However, the application of the term is inconsiste nt. Some jurisdictions require formal notifications for certain levels of injuries within a specified period of time. Other jurisdictions require immediate notice when certain quantities of hazardous materials are released. Regulatory requirements vary by co mmunity, locale, state, and country. The specific extent (number of agencies), format, and timing of all external notifications should be id entified beforehand, including contact information, and incorporated into the incident investigation and/or other applicable management systems. W ith th is information readily at hand, the proper notifications may be made quic kly and accurately when an incident occurs. Records of notifications and any follow-up communications should be preserved. Internal notifications, s ometimes called alerts or flash reports, are trigger mechanisms for starting specific portions of the incident investigation management system and for decision ma king. Obviously, medical treatment of injured personnel and stabilization of the incident site always takes priority over other activities if th ere is a conflict regarding the use of available resources during the early stages of an incident. These notification alerts may

YYJW INVESTIGATIN G PROCESS SAFETY INCIDENTS Table 15.3 Example Categories for Incident Investigation Findings 332 Table 15.4 Recommendations Review Checklist 335 Table 15.5 Example Follow-Up Checklist 336 Table 16.1 Questions for Identi fying Learning Opportunities 344

4 • Process Shutdowns 53 Projects that succeed have a clear framework and a clear execution plan, which is tracked an d monitored at each stage of the project. These reviews occur and are managed by a project “gatekeeper” when executing a project, the project-related risks, the process safety risks, the safety and occupational health risks, the environmental risks, and the minimum regulatory compliance constraints, which should be unde rstood and managed at all times. As is illustrated in Figure 4.2, th e stages in a large capital project life cycle are: 1) Appraise; 2) Select ; 3) Define; 4) Detailed Design; 5) Construction (includes Fabricatio n and Installation); 6) Start-up (includes Commissioning); 7) Operation; 8) Project Phase Out; 9) Small Projects and Management of Change (MOC) efforts; 10) End of Life Project Initiation; and 11) End of Li fe. Each of these stages, including the distinction between the different Front-End Loading (FEL) stages 1, 2, and 3, is described in more detail elsewhere [31, p. 9]. Since many severe incidents have occurred during the construction, commissioning, and operation stages, resulting in significant impact—fatalities, injuries, environmental harm, and property damage—it is essential that the process safety hazards are understood and their risks are managed during every stage. Hazards studies and risk reviews may be requ ired by the company during each of the life cycle stages. Construction activities include fabrication, installation, quality management, and pre-commissioning, with operational readiness activities required to prepare for commissioning, start-up, and ope ration. The commissioning stage concludes with the facility and its project-related documentation handed over to the operations group for normal operation. In addition, the capital project’s operation stage may require additional test runs to confirm that the sp ecific equipment or process unit performance specifications have been met before the handover to the

184 Human Factors Handbook Figure 16-1: Examples of error-likely situations More information can be found in Table 16 -2, which also offers some tactics for managing these three types of situations. It can be used as a checklist to identify error-likely situations. Chapters 2 to 4 discussed types of human error and their causes – slips, lapses, and mistakes. Task-specific conditions can contribute to potential errors and mistakes, such as unclear instructions, task complexity, and inadequate task experience. A set of error conditions is given in Table 16-3 to use in task-specific error assessment, with matching tactics for error management. It can be used as a checklist to identify high risk situations. Shift handover with defective equipment or plant in abnormal state Communicating between physically separated individuals or teams during a long duration task Starting up a modified process for the first time Performing a long duration task where precision, such as mass balance calculations, is required Monitoring a process for a long time with low task demands or no active tasks to be done Performing a long duration task where the correct sequence of actions is necessary A highly committed, task focused, and proficient team working hard to perform a challenging task with a tight deadline Attempting to isolate and purge a storage tank and pipework with out-of-date or incorrect information Trying to meet a tight schedule of production or maintenance while understaffed

2.10 Lear n to Assess and Advance the Culture |71 cause of an incident (Ref 2.33). For exam ple, the response to a corrosion-caused release should not stop with just replacing the corroded pipe. Was the corrosion rate excessive? If so why? Does this suggest a processing problem that must be addressed? Why was accelerated corrosion not detected? Bear in mind that m ost failures and near misses are rarely a unique event that can be shrugged off. Instead, fix the root causes to potentially uncover a broader more invasive problem . B e sensitive to leading indicators Form al leading indicators (see Section 3.1) can help identify norm alization of deviance and other developing problem s with the PSM S and process safety culture. In the post m ortem of nearly all undesired events, the investigations revealed that the information needed to detect, prevent, or mitigate the events in question were available to the organization but they were ignored or not understood. Near-m isses are potent leading indicators that should not be ignored, because they highlight conditions that are more likely to cause an incident. See Appendix D for a discussion of how high reliability organizations act on leading indicators and near-misses. Assess the culture Part of assessing the process safety culture of an organization involves form ally measuring it periodically. Organizations should establish carefully considered and mutually agreed upon key performance indicators that should be collected, reported, and analyzed by organization management on a periodic basis. See Section 4.5 for a m ore detailed discussion or process safety culture metrics. See section 4.2 for a discussion of the relationship between process safety culture m etrics and com pensation. Know the subcultures

5 • Facility Shutdowns 75 5.3 Projects requiring a process unit or facility project-related shutdown This section describes brownfield project-related considerations that may apply to facility shutdowns in addition to those discussed for a process shutdown (see Chapter 4, Se ction 4.3). (Greenfield projects will be discussed in Chapter 9.) Facilit y shutdowns also rely on a robust protocol for effectively managing al l the steps in the project: 1) planning, 2) preparing the equipm ent, 3) executing the work, 4) commissioning, and then 5) safely starting back up. These steps were described in Chapter 4, Section 4.4., with the two transient operating modes for a facility shutdown discussed in this chapter: 1. The shut-down beforehand (steps 1 and 2 listed above; Type 5, Table 1.1), and 2. The start-up afterwards (step s 4 and 5 listed above; Type 6, Table 1.1). It is essential for larger projects to use a systematic and disciplined approach to manage the project’s process safety risks using these project life cycle stages. Although these stages apply to both small and large projects, a facility shutdown should have a capable project manager who can effectively lead the complex project’s cross- functional Project Management Team (PMT) through each of the project’s stages, especially during handovers between groups (see Chapter 4, Section 4.3.3). The effe ctive project manager can help the company achieve success for both th e project-related scope (on time and on budget), as well as the process safety risk s (no incidents). This section discusses some additional considerations which may apply to a complex project: additional brownfield-related project considerations (Section 5.3.1), retrofit and expansion projects (Section 5.3.2), control system upgrades asso ciated with the project (Section

170 Human Factors Handbook Table 15-1: Principles of shift design Principle Guidance Maximum hours per day As fatigue increases quickly after eight hours, the duration of each shift should be limited, such as no more than 12 hours per 24 hour period. Rapid forward rotating shifts As few people adjust to night shifts, rapid forward rotating shifts (such as two day shifts followed by three or fewer consecutive night shifts) minimize the adverse impact of working nights. Working nights should be followed by rest days to allow people to recuperate. Regular rest breaks Rest breaks help people recover during a shift. Minimize early starts Early starts, such as 06:00, disrupt sleep and should be minimized. Limit consecutive shifts Fatigue accumulates – people need rest days.

86 INVESTIGATING PROCESS SAFETY INCIDENTS The determination of potential seve rity can be complicated. It is recommended that personnel responsible for classifying incidents based on potential severity be trained in the classification methodology. CCPS’s earlier document provided guidance on how to determine potential severity of a loss of primary containment (LOPC) of a hazardous material (flammable and toxic) (CCPS, 2011). See Appendix G fo r an extract from this publication addressing the potential chemical impact of Tier 1 process safety incidents. i. API Recommended Practice 754 The American Petroleum Institute (API) developed a similar guide for process safety performance indicators , incorporating input from CCPS, and subsequently revised and published a second editi on of the recommended practice (API, 2016a). The purpose of this document is to identify leading and lagging indicators in the refining an d petrochemical industries to drive improved safety performance. API prop oses indicators for use at both corporate and a site levels. In addi tion, API has published a guide for reporting process safety events (API, 2016b). Other industry organizations, including the European Chemical Industry Council (CEFIC, 2016), International Council of Chemical Associations (ICCA, 2016), and International Association of Oil & Gas Producers (IOGP, 2011), have adopted API RP 754, sometimes with minor variat ions. As discussed in Section 5.2.1.i, CCPS aligned its guidance wi th API RP 754 in 2018. Although API RP 754 is intended for standardized reporting of process safety events (i.e., incidents), some co mpanies have used it as a classifying tool for the purpose of determining the type and depth of investigation. As in the case of CCPS guidance, companies may wish to consider potential severity when determining the type of investigation, and small companies may reduce direct cost criteria as appropriate for their operations. ii. Logic Tree A few companies use a logic tree approach to determine incident classification and the type of investigat ion to conduct. An example logic tree is shown in Figure 5.1; it contains simple questions requiring yes/no answers. In this example, actual and potential serious injury and fatality incidents would receive a formal investigation. An unsafe act or behavior with injury and/or fatality potential would have an informal investigation, although at management discretion, it may receive a formal investigation. An unsafe act or behavior with no injury or fatality impact would not be investigated and would be recorded for trend analysis on ly, unless there is a trend worthy of investigation.

General Rules in Drawing of P&IDs 25 unavoidable in some cases. However, pipe crossing (or better, pipe clashing) in field shows a mistake in the design. A crossover could be shown in P&IDs in two forms, a “jump” or a “jog” (Figure 4.6).The decision on which line should be “manipulated” and which remains intact will be based on the prevailing line rule shown in Figure 4.7. For example, in crossing a horizontal process line and a vertical process line, the horizontal process line remains intact (Figure 4.8). 4.2.2 Equipmen t Crossing Equipment–equipment crossing is not allowed on a P&ID (Figure 4.9).Jump Jog Figure 4.6 P&ID repr esentation of pipe crossing. Most prevailing lines Horizontal process lines > Vertical process lines > Instrument signalsLeast prevailing lines Figure 4.7 Lines prev ailing rule. Figure 4.8 Examples of line crossings . Figure 4.9 Equipmen t–equipment crossing is not allowed.

Piping and Instrumentation Diagram Development 160 9.12 Blanketed Tanks When the goal is to store volatile liquids in a container, a floating roof tank can be used. The other option is to use a fixed‐roof tank with a blanking or padding system. Blanketed tanks can be used if the liquid of interest is volatile but not very volatile.A blanketing system provides a positive pressure over the surface of the liquid inside the tank in order to mini-mize the escape of volatile liquids, and also to provide a safe atmosphere inside the tank (Figure 9.25). To implement a blanketing system two questions need to be answered: what type of blanketing gas should be used, and what type of blanketing system should be designed? Blanketing gas should be an inert gas to minimize the chance of flammability and corrosion. The most com-mon blanketing gases are nitrogen, natural gas, and car - bon dioxide. Nitrogen is the best and the most expensive gas for blanketing purposes. However, we can’t always afford to use nitrogen as the blanketing gas. The most common use of nitrogen gas as a blanketing gas is in the food industry, where the products are expensive and need to be in contact with a food‐grade gas. In the oil industry, the most common type of blanketing gas is nat - ural gas, which is mainly composed of methane. Methane can also be considered as an inert gas. You may be sur - prised that natural gas is “inert” since it is very flamma-ble. However, the fact is that natural gas is only flammable when it is mixed with air. Natural gas, when used as a blanketing gas, is always enclosed in the top space of the tank and it is not in contact with air, so there is no chance of fire and it can be considered as an inert gas in this application. The last type of blanketing gas is carbon dioxide (CO 2). Carbon dioxide can form acidic vapors when it is in contact with water. These acidic vapors are corrosive. Therefore CO 2 is generally useable in applica- tions where the equipment is already corrosion resistant. Cold vent Automatic drain valve Figure 9.23 Cold v ent stack. To Atm.To VRU In - Situ treatment - Scrubber - Vent condenserEx - Situ treatment Figure 9.24 Differ ent methods of handling vent vapors.Figure 9.25 Blanketed tank .Table 9.11 Differ ent types of free vents. Straight type Gooseneck type Ell type Vent type Pros The least expensive Mid‐price Less troublesome Cons Risk of getting in the rain and debris Risk of dripping of condensate on the roof The most expensive

18 Guidelines for Revalidating a Process Hazard Analysis • Revalidations that are more freque nt are needed to meet company loss prevention goals. • If a major project is completed late in the revalidation cycle, it is often more effective to revalidate the PHA as part of the project hazard analysis effort rather than performing dozens of MOCs for the project and then Updating or Redoing the PHA in its entirety at a later time. • MOCs, pre-startup safety reviews (PSSRs), and operational readiness (OR) reviews are intended to maintain the integrity of original safety features designed into the process, and to ensure that any new hazards are prop erly managed. However, the potential for overlooked and uncon trolled hazards exists, and the potential for such oversights incr eases with the number of process modifications. Some companies choose to conduct a revalidation based upon the cumulative or simultaneous number of changes (e.g., during a unit shutdown). • Some companies have establishe d frequencies for revalidating PHAs based upon a risk categorization (e.g., high, medium, low); the higher the risk, the more frequent the revalidation. • Companies with multiple proces ses may decide to stagger the revalidation schedules for those pr ocesses to balance revalidation efforts from year to year. • Significant incidents or an unfavorable incident trend in a process might give reason to question the adequacy of the prior PHA, prompting an expedited revalidation. Similarly, incidents at another company site, or even outside the company, may provide reasons to re-evaluate the prior PHA. • Mergers or acquisitions may pr ompt management to quickly reconcile quality or protocol disparities in PHAs or to conduct new baseline PHAs based on current company standards. • Companies might have a concern that PHAs conducted early in the development of the facility PSM program may not have been conducted rigorously or effectivel y, or the team may have lacked sufficient process knowledge/experience to identify and evaluate all hazards. Thus, the revalidation schedule for the facility might be accelerated. In summary, multiple factors influenc e the PHA revalidation schedule for a company. The decision should be made on a process-specific basis to address needs unique to a process, facility, or company.

204 | 6 Where do you Start? culture supports both. Support your case with strong exam ples, including near-m isses that could have caused significant property loss and casualties. The free publication The Business Case for Process Safety (CCPS 6.1) m ay help you argue the financial case. Selected questions from Appendix F may prove useful in highlighting culture gaps that can resonate with leadership and enable some early successes. You m ay need to make your case at m any levels of the organization. 6.2 ASSESS THE ORGAN IZATION ’S PROCESS SAFETY CULTURE B efore any changes or im provem ents to the process safety culture can occur at a facility, the existing culture must be assessed. This assessment should take place prior to the initial culture improvement effort, and before subsequent improvement cycles. Ideally, the assessment should address all culture core principles. If budget or staff is limited, triage of the principles m ay be perform ed to narrow the focus of the assessment to those that m ay provide the greatest im provement for the least cost and effort. Since most evidence of process safety culture exists outside of hard operational and financial data, assessing the culture largely consists of interviewing, observing, and surveying people as they go about their duties. To obtain an accurate picture, a large and diverse set of em ployees m ust be interviewed, from senior leadership through middle m anagement and supervision to hourly personnel. Responses and observations will likely differ by level, reflecting a diverse set of opinions on the status of the culture. Understanding the differences in perspective by level can be as inform ative as the individual responses. Culture assessments should consider the com ponents highlighted in figure 6.1and discussed in the following paragraphs. The master culture assessment protocol and follow- up questions presented in Appendix F can be used as a starting

10 • Risk Based Process Safety Considerations 195 Pillar II (24%): Hazard Identification and Risk Analysis, Process Knowledge Management; Pillar III (37%): Asset Integrity and Reliability, Operating Procedures, Management of Change, and Emergency Management. This overall trend is observed: When Pillar I’s elements were weak, Pillar II’s elements were compromised, and when Pillar II’s elements were weak, the effectiveness of the elements in Pillar III were compromised. Each of these three pillars and their eight corresponding elements, those contributing most often to incidents during transient operating modes, are discussed in more detail in the next few sections. Figure 10.2 Summary of CCPS RBPS Weaknesses from Incidents Occurring During the Transient Operating Modes Note: The incidents used to generate Figure 10.2 are provided in the Appendix. Cumulative 70%Pillar II Pillar IVPillar IPillar III Pillar II Pillar IPillar III Pillar III Pillar III Pillar IV Pillar IV Pillar IV

246 | 7 Sustaining Process Safety Culture understand more deeply the potential impact on the process safety culture. Measure : Review process safety culture metrics and PSMS m etrics and ask how these are being impacted by culture. Assure M ulti-Level Support A sound process safety culture requires support across the entire organization in order to succeed. However, most of the problems associated with upper level support are not related to resistance but instead to multiple other com peting and pressing priorities. For the culture to flourish, the support from senior leadership should be steady and consistent (See Chapter 3 on Leadership). Also required for success is support from the lower levels of the organization, without whom the culture will be only publicity. Lack of support from the bottom is usually due to lack of understanding or lack of information, i.e., communications problems, as well as problem s with m utual trust. Plan for Succession As discussed in Chapter 3, unplanned successions in key positions can create vulnerabilities. This goes beyond operational and process safety professionals. Indeed, senior and m id-level leadership positions are also key in this regard. Succession planning not only results in better continuity, it also tells the organization that their leaders are developing them . This creates better m orale and engagement … and better future leaders. Onboard New Em ployees Every new employee represents a vulnerability to the culture. This m ay be due to lack of knowledge about cultural expectations or may be from external cultural influences. The new employee orientation program should provide the core expectations for working with the culture. Then, new employees should be •

Chapter No.: 1 Title Name: <TITLENAME> p04.indd Comp. by: <USER> Date: 25 Feb 2019 Time: 12:32:25 PM Stage: <STAGE> WorkFlow:<WORKFLOW> Page Number: 357 357 Piping and Instrumentation Diagram Development, First Edition. Moe Toghraei. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.Companion website: www.wiley.com/go/Toghraei_PID Part IV Utilities Utilities are auxiliaries of equipment and instruments and are required for their functioning. As utilities may be needed by a large number of items in a plant, as an economic decision, they are mainly based on air and water. Ambient air free and water is abundant. However there plenty of cases that should deviate from the basic air and water as the utility of choice. These are mainly fluids but one important exception is electricity. The fluids are used as utility to transfer material or to transfer energy. The examples of utilities for the goal of material transfer are potable water, plant water, etc. The examples of utilities for the goal of energy transfer are instrument air, cooling water, etc.

210 8.68 Negron, R., Using ultravio let disinfection in place of chlorination. The National Environmental Journal, 48-50, 1994. 8.69 Norman, D., Turn Signals are the Facial Expressions of Automobiles, Addison-Wesley Publishing Company, 1992. 8.70 US Occupational Safety & Health Administration (US OSHA), 29 CFR 1910.1200, Hazard Communication Standard, 2012. 8.71 Petroski, H. (1985). To Engineer Is Human: The Role of Failure in Successful Design. St. Martin's Press/Vintage Books, 1985/1992. 8.72 Rand, G., Petrocelli, S. , Fundamentals of Aquatic Toxicology. Hemisphere Publishing, 1985. 8.73 Sanders, R. E. Management of Change in Chemical Plants Learning From Case Histories, Butterworth-Heinemann, 1993. 8.74 Stankiewicz, A., Re-engin eering the Chemical Processing Plant: Process Intensification (pp. 261-308), CRC Press/Marcel Dekker, Inc., 2003. 8.75 Tyler, B., Thomas, A., Doran, P., and Greig, T., A toxicity hazard index, Chemical Heal th and Safety, 3, 19-25, 1996. 8.76 United Nations, Global Harmonized System of Labeling and Classification of Chemicals (GHS), 4th Edition; United Nations, New York and Geneva, 2011. 8.77 U.S. Chemical Safety an d Hazard Investigation Board, Investigative Report, LPG Fire at Valero – McKee Refinery. Washington, DC: Report No. 2007-05-I-TX, 2008. 8.78 U.S. Coast Guard (USCG), CIM 16616.6A, Chemical Data Guide for Bulk Shipment by Water, 1990. 8.79 U.S. Department of Energy (USDOE), AEGLs, ERPGs, or Rev. 21 TEELs for Chemicals of Concern 2005, DKC-05-0002, 2005. 8.80 U.S. Department of Heal th and Human Services, Agency for Toxic Substances and Disease Regi stry (ATSDR), www.atsdr.cdc.gov/

6 | 1 Introduction Dow Plant example shows that we can—by learning from incidents and embedding what we learn in the culture. 1.2.1 The Theory of Root Cause Correction Most process safety hazards require multiple barriers to control them. No barrier is 100% reliable. Every barrier has a “probability of failure on demand,” represented in Figure 1.3 as holes in slices of Swiss cheese. •Hazards are controlled by multiple protective barriers. •Barriers may have weaknesses or “holes.” •When the holes align, a hazard may pass through all the barriers, with the potential for adverse consequences. •Barriers may be physical or engineered containment, or procedural controls dependent on people. •Holes may be latent/incipient or opened by people. Figure 1.3 The Swiss Cheese Model (Adapted from Reason 1990) Preventive barriers control process deviations and keep hazards within process equipment. When releases do occur, mitigating barriers prevent or reduce the consequences for workers, the facility, the public, and the environment. If we do not manage our barriers properly, it’s as if we increased the size of the holes in the Swiss cheese. The probability of failure on demand—the probability that hazards will pass through the barriers and cause harm—will increase. We manage process barriers by adhering in a disciplined way to our PSMS and applicable standards. For example, the safe work practices element of RBPS (CCPS 2007) controls the following barriers: • control of ignition sources • hot work

LESSONS LEARNED 343 Other sources of information may include insurers and other government agencies such as the Na tional Transportation Safety Board (NTSB) for railcar events, OSHA, etc. Further references are provided in the References section. 16.1.3 Cross-Industry Many organizations tend to recognize only those incidents that have occurred in similar operating environments or in similar processes within the same industry sector. For example, the chemical industry currently does not have a common platform to exchange incident information with the oil and gas sector, pulp and paper, or other industries. Given that the hazards, equipment, and processes used may be very similar in these industries, there is a significant need and opportunity to share lessons across geographical or industry boundaries. Lessons learned from enti rely different industries (such as the airline industry) may also be relevant to the chemical processing industry since there are common touchpoints (human fact ors, prestart checks, etc.). These opportunities should not be overlooked. Material for learning from incidents can be obtained from a number of sources, as detailed in 16.1.2. 16.2 IDENTIFYING LEARNING OPPORTUNITIES A well-written investigation report and associated recommendations should be structured in such a way that the le ssons learned are clearly identifiable. However, this is not always the case and management must play a role in ensuring that the learning opportunities are identified an d communicated to appropriate personnel and across organizational boundaries. How this is organized will depend on the type and size of company or facility. Nevertheless, at least one me mber of staff, in a management and/ or safety function, should be responsible for this activity; receiving incident reports from within the organization as well as seeking ou t information from external events from sources such as th ose listed in 16.1.2. This person or group should then extract the key learni ng that may be relevant to one or more of their facilities and prepare a communication aid to allow this learning to be disseminated to staff.

3.2 Characteristics of Leadership and Management in Process Safety Culture |95 broad functions at the plant, and it is advantageous to keep the functions together in organizations by specialty. For this reason, PSMS managers often have no direct control over all the elements. The m anager may have an indirect relationship with some functions, while others m ay operate totally independently. Facility leaders should create a culture where collaboration and coordination break down the silos, so they can ensure that all skill areas fulfill their process safety responsibilities and foster better integration of PSM S elements. Avoid the “Flavor-of-the-M onth” As will be described in Section 3.4 the consistency of the process safety message is important. Com peting and changing values make them seem tem porary, rather than core to the organization, while changing goals m ay come to be considered optional. Leaders should deploy goals thoughtfully, strategically and systematically. The Organization for Econom ic Cooperation and Development (OECD) summarizes this succinctly (Ref 3.20): “The CEO and other leaders create an open environment where they: Keep process safety on their agenda, prioritise it strongly and remain mindful of what can go wrong, Encourage people to raise process safety concerns or bad news to be addressed, Take every opportunity to be role models, promoting and discussing process safety, Delegate appropriate process safety duties to competent personnel whilst maintaining overall r esponsibility and accountability, Are visibly present in their businesses and at their sites, asking appropriate questions and constantly challenging. the organisation to find areas of weakness and opportun-• • • • •

Overview of the PHA Revalidation Process 7 Note that different complementary analyses can be used to address the same topic or area of conc ern. For example, facility si ting issues affecting PHA results can be identified and evaluated in multiple ways, including: (1) the consideration of siting during the core analysis, (2) a facility siting checklist review as a complementary analysis, and/or (3) a detailed analysis in accordance with recognized and generally accepted good engineering practices (RAGAGEPs) applicable to a process. For example, the American Petroleum Institute (API) has developed recommended practices for assess ing facility siting topics in subject processes [5] [6] [7]. The Chemical Indust ry Association (CIA) provides guidance for performing occupied building risk assessments (OBRAs) [8]. Such RAGAGEP analyses are often quantitative (either consequence-based or risk-based) and supported by checklists, calculation sheets, and/or models of varying complexity. Company policy may require such assessments, but if they are not explicitly included in PHA requirements , they are not complementary analyses for the purposes of this book. Examples of complementary analyses are listed in Table 1-1 along with CCPS or other industry publications where additional information can be found. Table 1-1 Complementary Analysis Examples Complementary Analysis Information Source What-If Analysis* CCPS book Guidelines for Hazard Evaluation Procedures [2, pp. 100-106] Checklist Analysis* CCPS book Guidelines for Hazard Evaluation Procedures [2, pp. 107-114] FM Global Checklists Industry publications, guidelines, and recommended practices (RPs) Facility Siting Study API RP 752 [5], 753 [6] , 756 [7] CCPS book Guidelines for Siting and Layout of Facilities [9] CIA OBRA [8] Human Factors Analysis CCPS book Guidelines for Preventing Human Error in Process Safety [10] Hierarchy of Hazard Controls Analysis (HCA) CCPS book Inherently Safer Chemical Processes: A Life Cycle Approach [11] Damage Mechanism Review (DMR) API RP 571 [12] API RP 970 [13] Dust Hazard Analysis (DHA)* CCPS book Guidelines for Combustible Dust Hazard Analysis [14], National Fire Protection Association (NFPA) Standard 652 [15] *can be a core analysis or a co mplementary analysis in a PHA

270 | Appendix B Other Safety and Process Safety Culture Frameworks Staff have non-technical knowledge and skills related to hum an factors, team perform ance and error managem ent techniques. Authority to m ake decisions lies with the most qualified employees. Contingencies are in place to fill vacated roles with com petent staff. Com bat the Norm alization of Deviance Attributes and descriptors indicating the presence of normalization of deviance include: The organization fails to implement or consistently apply its PSM S across the operation; regional or functional disparities exist.Procedures, policies, and safeguards are routinely circumvented to get the job done. The organization fails to provide adequate or effective systems, processes, and procedures for work being perform ed. The organization fails to provide necessary financial, hum an, and technical resources. Impracticable rules, processes and procedures, which m ake compliance and achievem ent of other organizational outcom es mutually exclusive. Employees find workarounds in response to operational inadequacies. The organization fails to provide employees with effective m echanisms to resolve operational inadequacies. Operational changes are im plemented without m anagement of change or PHA/HIRA. Rules and operational procedures are not followed. Extended time between reporting of process safety issues and their resolution.• • • • • • • • • • • • •

148 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS Table 5.9 Learning from Abnormal Situation Incidents Common Tools and Methods Strengths Weaknesses Near-Miss Analyses A review of the learning from previous near- misses can be used to help prevent the occurrence of more significant incidents in the future. Rules should be established to define a “near-miss” within an organization so corrective actions can be consistently applied. Previous incidents (internal and across industry) May be used to identify abnormal situations that could cause an incident. Requires maintaining an available list of incidents to share with personnel, unless learning is embedded into procedures. Fault Trees or Bow Ties Can be used graphically to depict paths for abnormal incidents to occur and safeguards to prevent. Personnel must be trained to construct the graphical model to reflect correct content and results. Process Metrics Can be used as Leading Indicator of process upsets or non- acceptable situations. Metrics should be meaningful and not a target disguising underlying process safety issues. Most companies and facilities have a procedure for investigating incidents and accidents. However, an abnormal situation may have been prevented from becoming a major event because it was stopped by another barrier (procedural or phys ical). Unless the company has a culture and procedure for identifyin g and investigating “near-misses”, including getting to the root cause and not simply concluding it was due to “human error”, a learning opport unity may be lost. The facility may not be so lucky next time when that other barrier might not be performing as well.

384 INVESTIGATING PROCESS SAFETY INCIDENTS Plot Plan (1 of 2)

Piping and Instrumentation Diagram Development 316 The capacity of the pump is controlled by a flow loop. A flow sensor on the discharge line sends a signal t o the flow controller, which, in turn, sends a deviation signal to the VSD on the electromotor of the PD pump. 15.7.1.2.3 PD P ump Control by Stroke Adjustment A third type of PD pump control is only for reciprocating pumps. In the reciprocating type of PD pumps, a mem-ber does a forward and backward movement repeatedly in order to achieve the pumping. The third method of controlling the capacity is appli- cable only to reciprocating pumps: adjusting the span of the member’s reciprocating movement, or simply adjusting the stroke. This can be done by attaching a servomotor to the “stroke‐adjustment lever” of the pump. The signal from the flow controller can increase or decrease the stroke length, which will adjust the flow rate. This control is shown in Figure 15.43.15.7.2 Gas M over Control Systems There are two main types of gas movers: ●Dynamic. These can be either axial or centrifugal. ●Positive displacement. These are either reciprocating or rotary. Control systems for gas movers are very similar to pump control systems; both need to have capacity con trol. Whereas we don’t need minimum protection for PD pumps, we do need it for PD gas movers. Minimum pro-tection for pumps is referred to as “minimum flow con-trol, ” and for gas movers, it is called “surge protection. ” Again, for centrifugal gas movers, there are two control systems needed: capacity control, and anti‐surge control. 15.7.2.1 Capacity Control Methods for Gas Movers There are five main types of control for gas movers. These are: recirculation pipe (or spillback pipe), speed control, suction throttling, discharge throttling, and a combination of control methods based on the structure of the gas mover. Table  15.10 shows which methods are technically applicable for which types of gas mover. ●Recirculation. This is a general control method that can be used for any type of gas mover. In this FC VSD MTFE Figure 15.42 PD pump con trol by VSD.FC FT FE Figure 15.43 PD pump con trol by adjusting stroke length. Table 15.10 Con trol methods for gas movers. Dynamic Positive displacement Axial Centrifugal Reciprocating Rotary Recirculation General control method Speed control via VSD General control method Suction throttling Discharge throttling Special control Applicable Applicable Applicable Applicable

355 The guidance contains more detaile d questions and review items to apply during each of these project phases. The guidance does recognize that IS may not be applicable for a ll facilities and project phases in the same way or in the same sequence. IS should be examined as early as possible to minimize the impact on the project. For existing processes, the guidance notes that the consideration of IS is limited to scenarios where a Major Chemical Accident or Release (MCAR) could reasonably occur. CCHS de fines a MCAR as an incident that meets the definition of a Level 3 or Level 2 incident in the community warning system incident level clas sification system defined in the hazardous materials incident notifica tion policy, as determined by the department; or results in the releas e of a regulated substance and meets one or more of the following criteria: (1) Results in one or more fatalities; (2) Results in at least twenty-four hour s of hospital treatment of each of at least three persons; (3) Causes on - and/or off-site property damage (including clean-up and restoration activities) initially estimated at $500,000 or more. On-site estimates are performed by the stationary source. Off-site estimates are performed by appropriate agencies and compiled by the department; or (4 ) Results in a vapor cloud of flammables and/or combustibles th at is more than 5,000 pounds (Ref 14.4 CCC ISO). Facility PHAs can be used to identify which scenarios have consequences (without crediting any ac tive safeguards) that rise to the MCAR level. In most chemical/process ing industry PHAs that are part of formal PSM programs, this is do cumented in PHA worksheets by describing the consequences so that offsite and onsi te impacts that result from major releases are clearl y identified. This is also usually supplemented by a qualitative or quan titative consequence or severity ranking that clearly indicates the severe consequenc es that are the hallmark of a MCAR. Facilities can perform one of the fo llowing methods to ensure that ISS are considered and documented for the covered processes: An independent ISS analysis that is done in addition to a PHA. Checklists or guideword analys is contained in the guidance document that incorporates ISS can be used to accomplish this analysis.

8.7 NASA Space Shuttles Challenger, 1986, and Columbia, 2003 | 115 trained properly for a fire involving ammonium nitrate, they would have immediately evacuated all personnel and civilians from the facility and beyond. In all 15 people died, of which 12 were emergency responders. Hundreds more were injured in this incident. 8.7 NASA Space Shuttles Challenger, 1986, and Columbia, 2003 The 1986 Space Shuttle Challenger incident is a textbook case of failed safety culture primarily driven by tremendous pressure to produce. On an unusually cold day in January 1986, the Space Shuttle Challenger broke apart and burst into flames 73 seconds after launch. The incident occurred because two O-rings in the solid rocket booster, affected by the cold, burned through, allowing exhaust gases to burn through a liquid hydrogen tank. According to the Rogers Commission Report (Rogers 1986), the ambient air temperature at launch was 2 oC (36 oF) as measured at ground level approximately 300 meters (1,000 ft) from the launch pad—8 oC (15 oF) colder than for any previous launch. The previous evening, the National Aeronautical and Space Administration (NASA) and the solid rocket booster manufacturer had a lengthy phone call discussing the concern about launching in cold weather. The manufacturer’s engineering leadership recommended a delay in the launch, but after pushback from NASA, the manufacturer’s management team reversed the recommendation and provided their approval to launch. In such a sophisticated and complex process, overruling even one no-go condition should have been unacceptable. However, after delaying the launch for a year, NASA was eager to go. NASA used the data in Figure 8.7 (top) to justify the decision to go ahead with the launch. This graph showed the number of incidents versus temperature, but it was an incomplete picture. Had NASA viewed this data along with the number of successful launches without incident, as in Figure 8.7 (bottom), they might have decided to delay the launch. The complete data showed that all launches at temperatures below 19 oC (65oF), 6 oC (10 oF) above the recommended minimum launch temperature), incurred burn incidents. In the face of pressure to produce, the poor safety culture coupled with the lack of complete understanding of risk cost the seven astronauts aboard the Challenger their lives. See Appendix index entries S9 and S10

396 INVESTIGATING PROCESS SAFETY INCIDENTS Logic Tree (8 of 9)

16 Guidelines for Revalidating a Process Hazard Analysis 1.6 PHA REVALIDATION CYCLE PHA revalidations are performed periodi cally throughout the life cycle of a process. However, as illustrated in Figu re 1-3, they are not performed in isolation. Revalidations consolidate th e information gathered by other PSM elements in an operating plant and ensure the PHA provides a current assessment of the process risks. The PHA revalidation cycle (or schedule) is the time between the previous PHA and the PHA revalidation. During this time, the facility gathers operating experience on the process. This experience can be internal (e.g., through MOC or incident investigation) or external (e.g., regulatory changes or RAGAGEP revisions). Some facilities with mature PSM proce sses prefer to maintain a “living,” “evergreen,” or “continuous” PHA, which can be an effective way to merge the PHA and MOC processes. For these facilitie s, the relevant PHA documentation is revised whenever a change occurs for an y reason (e.g., process technology improvement, recommendation implementation, capacity expansion, and incident response). However, such revisions to portions of the PHA are usually Figure 1-3 PHA Creation and Revalidation Cycle

9 • Other Transition Time Considerations 181 Working at heights; Working near or over water; Presence and/or removal of overhead/underground/subsea pipelines and utilities; and Location of temporary facilities, (e.g. trailers) associated with deconstruction activities. If the end-of-life project is for demolition or deconstruction, potential issues may involve, but are not limited to [31, p. 24]: Proximity of neighboring facili ties and buildings may require dismantling and prohibit toppling/explosives, Deconstruction of some equipment for future re-use, Partial decommissioning of operating facility, Presence of asbestos and PC Bs in older facilities, Simultaneous operations with adjacent facilities, Vibration that might impact adjacent operations, Underground cables and piping, and sewers, including unknown o Locations, and o Connections to other adjacent facilities, Environmental remediation. These issues require care ful planning to perform decommissioning safely and efficientl y, and personnel should be fully knowledgeable of the hazards and the appropriate safety measures selected to mitigate the hazards and reduce the activity’s risks. Some companies may perform a formal “Hazards of Decommissioning (HAZDEM)” review before decommissioning, as well, to help ensure that the decommissioning activities are incident-free. When compared to deconstruction and dismantling activities, it is important to recognize that there may be greater hazards to personnel or other equipment during demolition activities. Proper planning and strong oversight and guidance with a decommissioning team project manager is essential to avoi d injuries and incidents.

274 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Table 13.1. Typical discharge scenarios Liquid Discharges • Hole in atmospheric storage tank or other atmospheric pressure vessel or pipe under liquid head • Hole in vessel or pipe containing pressurized liquid below its normal boiling point Gas Discharges • Hole in equipment (pipe, vessel) containing gas under pressure • Relief valve discharge (normally vapor only) • Boiling-off evaporation or volatile vapors from liquid pool • Generation of toxic combustion products as a result of fire • Vapor phase from a pressure vessel leak above the liquid level (subsequent boil-off will also contribute to the release Two-Phase Discharges • Hole in pressurized storage tank or pipe containing a liquid above its normal boiling point • Relief valve discharge (e.g. due to a runaway reaction or foaming liquid) Figure 13.5. Selected discharge scenarios (CCPS 1999) (Fryer and Kaiser 1979)

Plant Process Control 311 In option A, we basically adjust the pump’s flow using a control valve on the discharge piping of the pump (dis - charge throttling). In option B, we adjust the pump’s flow by adjusting the rotation speed of the shaft. These are two completely different methods of pump control. By using a VSD, we are changing the location of a pump curve, while by using the control valve we are changing the system curve. In the majority of cases, we use a VFD (variable frequency drive) as our VSD of choice for rotary machines in plants. Some people use the analogy of controlling the speed of a car for pump capacity control. We can see this in Figure 15.33. Option A for controlling a car’s speed is like pushing the gas pedal to the limit and pushing the brake pedal at the same time whenever we want to reduce speed. Option B for controlling a car’s speed is like adjusting the speed by adjusting the position of the gas pedal.Table 15.8 Options for c entrifugal pump capacity control analogy. Controlling the speed of a car Controlling flow rate Pushing the brake pedal whenever you want to decelerate while always pushing the gas pedal! Controllling signal MotorC Pinching back the valve whenever you want to decrease the flow rate (without touching the pump’s motor) Releasing the gas pedal whenever you want to decelerate Controllling signal MotorC Decreasing the flow rate of the pump (by adjusting its motor RPM) whenever you want to decrease the flow rate Whenever there is more than one control method for controlling a piece of equipment, both of them can be implemented through a “split‐range” control loop. For example, the capacity of a pump can be controlled by a split range between a VFD and a control valve.

Piping and Instrumentation Diagram Development 292 The temperatures of the skin of tubes and also the fire box should be monitored. If the fired heater is the area with stringent environ- mental regulations and/or the efficiency of firing is very important the flue gas components can be monitored by process analyzers. The examples are carbon monoxide and the oxygen content of the flue gas. A typical monitoring system around a fired heater is shown in Figure 14.36.TISkinSkinCO O2 PIFuel ga s Pilot gasTITITITIAl AlPC TC FC FCFC SD SD PI SD SDFigure 14.36 Monit oring of fired heaters.

34 | 2 Core Principles of Process Safety others to fulfill their process safety responsibilities. Without trust, a leader assigning a critical process safety task accepts that it may not be done. And without m utual trust between managers and their experts, deference to expertise simply cannot happen. In the absence of trust, em ployees may dismiss a m anager’s statem ents about safety as not serious. Managers m ay seek blame for errors, rather than seeking root causes. Workers may be reluctant to report near-m isses and incipient safety problems. Managers may withhold funding for safety because they believe workers are being lazy. Managers may second-guess the experts and end up m aking poor decisions. When trust is lost, a culture can be seriously dam aged. Trust can be destroyed m uch faster than it can be built. And, rebuilding trust after it is lost can take as long as building it in the first place. CCPS cites lack of trust as a key warning sign of potential catastrophic incidents (Ref 2.12). Table 2.1 summarizes some indicators of trust and mistrust within an organization. Table 2.1 Indicators of Trust or Mistrust Trust Indicators Mistrust Indicators Personnel willingly volunteer “It’s not my job” (Ref 2.13) Leadership and peers trusted to make right decisions Problems and concerns hidden from m anagement Problems reported without fear of reprisal People feel they are missing essential information Employees satisfied with problem resolution Unusual friction between groups, shifts, facilities, corporate, etc. Auditing and checking welcomed Cliques Comradery across organization Recklessness tolerated (Ref 2.13) Bad ideas challenged when proposed Conflict with contractors, neighbors, labor, press, etc. (After Ref 2.12 except as noted)

Plant Process Control 313 ●Pumps in a circulating, closed system: they may not need a minimum flow control loop because the flow in such pumps is fairly constant. However, the following examples of pumps may require minimum flow control: ●Large pumps of 5 hp or mor e ●Pumps on the main stream. The rule of thumb for providing minimum flow con- trol is as follows: ●Power < 5 hp: no minimum flow line is required, as men tioned before. ●5 hp < Power < 10–20 hp: continuous minimum flow line. Inste ad of going to the expense of installing a control valve, we can put a restriction orifice (RO) on this line (Figure  15.35). By doing this, we are always recirculating a portion of flow, even in cases where flow to the pump is higher than the minimum flow and we don’t really need recirculation. Thus, we are con-tinuously wasting energy. We know that, but the pump is so small that installing an expensive control loop for them is hard to justify. (You can picture an RO as a “frozen” control valve with a specific opening size.) ●10–20 hp < Power < 15–30 hp: ON/OFF minimum flow line. Thi s is a cheaper option than installing a control valve. In this case, the flow controller will just provide an ON/OFF function (Figure 15.36). ●Power > 35 hp: minimum flow line with a control valve. This is the most complicated, most expensive option, but it is the most common method of controlling mini-mum flow in centrifugal pumps. This type of control can be done with at least two different arrangements: with a flow loop, and with a pressure loop, as shown in Figure 15.37. We have the option of controlling by flow sensor or by pressure sensor. Usually we go with a flow loop. When there are two (or more) pumps in parallel and only one of them is operating and the rest are spares, a minimum spillback pipe from the common header works well.Senses more than 100 m3/hr so closes the va lve Senses less than 10 0 m3/hr so opens the va lveMinimum flow: 100 m3/hrFV FVFC FCFIC FICFT FTFX FX0 m3/hr 80 m3/hr165 m3/hr 20 m3/hr165 m3/hr 80 m3/hr Figure 15.34 Minimum flo w control. RO Figure 15.35 Minimum flo w control by RO.FC Figure 15.36 Minimum flo w control by switching valve.

100 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION the cells on the chart can be clicked to find more information about specific predicted reactions. General hazard statements, predic ted gas products, and literature documentation for the selected pair of chemicals are shown at the bottom of the chart. Other functionality to understand reactivity are: chemical information, physical properties, synonyms, a reactive groups tab, absorbent incompatibilities, and a pl ace to start to look for compatible materials of construction. Finally, the Helphelp function is very useful to understand how to create mixtures and understand their reactivity. The Reactive Chemicals Checklist contained in Appendix D is adapted from a CCPS Safety Alert A Checklist for Inherently Safer Chemical Reaction Process Design and Operation . (CCPS 2004) It provides a checklist addressing chemical reaction hazard identification and reaction process design. Figure 5.8 is a chemical reactivity worksheet showing the reactivity of sulfuric acid and sodium hydroxide, a strong acid / strong base pair. The worksheet illustrates that the acid and base are incompatible and hazardous reactivity issues are expected. The hazard summary lists the hazards of mixing including corrosive reaction products and gas evolution that might result in pressurization. The potential gases tab lists ac id and base fumes, and nitrogen oxides. This view of the CRW also summarizes the intrin sic chemical hazards and NFPA 704 ranking. Figure 5.8. – CRW for strong acid strong base pair (CCPS) Documents The following books and resources are available for helping to understand reactive chemical hazards. Bretherick’s Handbook of Re active Chemical Hazards, 7th Edition. B r e t h e r i c k ’ s i s a 2 - volume set of all reported risks such as explosio n, fire, toxic or high-energy events that result from chemical reactions gone astray, with ex tensive referencing to the primary literature. (Bretherick & Urban 2006) CCPS Essential Practices for Managing Chemical Reactivity Hazards. This book provides technical guidance to help small and large companies to identify, address, and manage chemical reactivity hazards. Appendix C has the flowchart developed for this book that guides

26 Guidelines for Revalidating a Process Hazard Analysis August 2017. Compliance is a general obligatio n, so this standard has no specific list of chemicals or quantities. The Canadian PSM standard requires an initial risk assessment (PHA) that must be revalidated at least every five years. It is sometimes argued that specific la ws or regulations in one jurisdiction can create obligations elsewhere. This was particularly evident in the prosecutions following the 1998 explosion at the Longford gas plant in Australia [25, pp. 167-174]. It was argued that the site operator’s performance of retrospective HAZOPs (as regulations requ ired elsewhere) created an obligation for them to be performed for facilities in Australia, even though there were no specific Australian PHA requirements at the time. (Australia enacted specific regulations on Major Hazard Facilities in 2000). Therefore, organizations should strongly consider being globally cons istent in how they perform PHAs and revalidations. 2.1.2 Specific RAGAGEPs With respect to process safety, companies around the world usually rely on their compliance with RAGAGEPs to show th at they have fulfilled their general obligations. For example, the adequacy of overpressure protection is one of the most important factors in risk judgme nts faced by any PHA team. Compliance with the American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code (or an international equi valent, such as the European Union Pressure Equipment Directive) is typica lly considered a very reliable safeguard against vessel rupture due to overpres sure. Examples of related issues a revalidation team should consider are whether: • Process changes have been appropriately managed such that no new scenarios might exceed the design basis for the pressure relief devices. • Reductions in maintenance staffing or inspection/test frequencies for the pressure relieving devices or protected equipment have affected code compliance. • Mechanical integrity of the protected equipment has been maintained within acceptable limits for its pressure rating. Boiler and pressure vessel codes are uniquely important in that many jurisdictions require compliance by law; however, organizations typically have more flexibility in their approach to compliance with other RAGAGEPs. Multiple RAGAGEPs might apply to a process being evaluated in a PHA, so it would be unreasonable to expect that an y initial PHA team or revalidation team

EQUIPMENT FAILURE 229 Mechanical engineers often have the lead role in identifying failure rates in determining reliability. Failure rates can be used in reliab ility-centered maintenance and can be translated into frequencies of failure and probability of fa ilure on demand which are key values used in semi-quantitative and quantitative risk analysis. A technique for inspection, borrowed from pr ocess safety, is the use of risk-based inspection (API RP 580). This combines corrosi on mechanisms predicting corrosion rates and consequence models predicting potential outcomes to prioritize sections of piping or process vessels for greater or less inspection attention. RBI has been shown to enhance safety and reduce inspection costs by focusing attention on the most important areas. Instrument, Electrical and Control engineers will often be tasked with designing Safety Instrumented Systems (SIS) and ensuring that the required reliability and probability of failure on demand is achieved. They may often be aske d to develop procedures for safety instrument system test protocols, instrument calibrations and testing, control loop response capabilities etc. Similar to the mechanical en gineer’s role in asset reliabilit y, instrumentation failure rates may be used to support semi quantitative and quantitative risk analysis. All engineers can support asset integrity by being observant when walking through the facility. Watch and listen for vibrating equipment and report concerns to your supervisor. You may see or hear something that is not be ing monitored by maintenance inspections. A new engineer can benefit from reviewing the CSB investigations and videos relevant to this chapter as listed in Appendix G. Tools Tools that may be used in understanding how equipment fails and supporting asset integrity to prevent that failure include the following. API Standards and Practices . API has created many standards and practices that address the design and life cycle integrity of various types of equipment. These are available at www.API.org . CCPS Guidelines for Asset Integrity Management. This book includes details on failure modes and mechanisms and testing an inspection programs for various types of equipment. (CCPS 2016) CCPS Guidelines for Improving Plant Reliability through Data Collection and Analysis . This book provides guidance on how to collect, and use with confidence, process equipment reliability data for risk-based decisions. It provides the techniques to gather plant Reliability-Centered Maintenance (RCM) – A systematic analysis approach for evaluating equipment failure impacts on system performance and determining specif ic strategies for managing the identified equipment failures. The failure management strategies may include preventive maintenance, pred ictive maintenance, inspections, testing, and/or one-time changes (e.g ., design improvements, operational changes). (CCPS Glossary)

64 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Types of Fires Each liquid and vapor fire type has its own charac teristics. These fire types are listed in order of increasing consequence severity. The type of fire is dependent upon the flammability characteristics determined by proce ss safety information of the fuel. Pool fire - The combustion of material evaporating from a layer of liquid at the base of the fire. (CCPS Glossary) Jet fire - A fire type resulting from the discharge of liquid, vapor, or gas into free space from an orifice, the momentum of which induces the surrounding atmosphere to mix with the discharged material. (CCPS Glossary) Flash fire - A fire that spreads by means of a flame front rapidly through a diffuse fuel, such as a dust, gas, or the vapors of an ignitable liquid, without the production of damaging pressure. (CCPS Glossary) Fireball - The atmospheric burning of a fuel-air cloud in which the energy is in the form of radiant and convective heat. The inner core of the fuel release consists of almost pure fu el whereas the outer layer in which ignition first occurs is a flammable fuel -air mixture. As buoyancy forces of the hot gases begin to dominate, the burning cloud rises and becomes more spherical in shape. (CCPS Glossary) Pool fires involve a pool of liquid. This may be from a spill of a flammable liquid on the ground or water, from the top of a hydrocar bon tank, or where a spilled liquid has followed grading to a drain. The location of a pool fire may be influenced by grading, berms or dikes, and containment booms. Firefighting foams may be used to blanket a pool fire to control, and potentially, extinguish it. Note direct applicatio n of firewater to burning pools is dangerous as this can spread the fire and water will not exting uish the flames as it sinks below the surface. The radiation from a pool fire can cause harm to people and equipment depending on the duration of exposure. Additionally, the products of incomplete combustion can be toxic.

90 | 3 Leadership for Process Safety Culture Within the Organizational Structure imperative, also help drive good process safety culture. Conversely, poor morale often accom panies a poor safety culture. Good morale means more than general happiness in an organization. Additionally, a sense of pride and satisfaction perm eates the organization, creating a sense of wanting to belong and of not wanting to disappoint their peers and leaders. An old naval m axim says that while the ship’s Captain delegates responsibility for nearly every duty to others, the m orale of the crew is the one responsibility that cannot be delegated. The same holds true in industrial facilities. Morale is the responsibility of senior leadership and cannot be delegated. Morale can be particularly hard to maintain during organizational changes such as new management, downsizing and changes in ownership. In addition to learning new roles, policies, and procedures, workers m ust deal with uncertainty about what the future holds for them personally. Leaders, especially new leaders, should pay particular attention to morale during such changes. Understand Process Safety vs. Occupational Safety Research at the Wharton School (Ref 3.19) revealed that there is no statistical correlation between the rates of process safety incidents and the rates of occupational injuries and illnesses. This can be easily understood by considering that occupational safety deals with how workers are protected, while process safety deals with how processes and equipment are designed, operated, and m aintained. Many leaders still do not understand this, despite numerous recent incidents that highlighted the issue. Process safety leaders should resist the temptation to infer that good occupational safety results m ean that the PSMS is functioning well. Leaders should discuss indicators of process safety performance

72 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS the MOC process may be utilized as the initial step to return decommissioned equipment to service. Example Incident 3.14 – Reboiler Decommissioning A distillation column had two reboile rs; one was relatively new and the other near end-of-life. The column only required one reboiler to be on-line. The decision was made to decommission the older reboiler as it had some steam leaks and a future capital investment (CAPEX) plan included provision for its replacement. A detailed decommissioning procedure was written and approved including some specific inspection requirements to ensure that the reboiler was properly cleaned. The reboiler was then safely disconnected from the column and bl ind flanges were installed on the column nozzles. The reboiler tubes were to be water flushed to remove any residual chemical traces and solids. The tubes were then to be dried and checked for mois ture. In addition, a physical inspection was to be conducted on each tube to ensure they were clean. The responsible personnel ch ecked for moisture as required and, finding none, concluded that the reboiler was dry and the physical inspection of each tube wa s not necessary. This fact was not shared with the plant management. Six months later, the newer rebo iler experienced a mechanical problem and since the new CAPEX re boiler had not yet been received, the decision was to return the remo ved reboiler to se rvice. However, the old reboiler failed its pressu re test and several tubes were corroded and had some small holes. The investigation found that a small amount of solids had remain ed in some of the tubes and reacted with water to form acid. Ov er the six-month period, the acidic solids had created pinhole leaks. The personnel responsible for the inspection admitted to not conducting an inspection of all the tubes and stated that the reboiler tested dry and an inspection of a few tubes did not show any solid residue.

150 Polychlorinated biphenyl s (PCBs) Environmental Poly cyclic aromatic hydrocarbons Environmental Reactivity of Types of Compounds . Many additional hazards derive from the hazardous reactivity of combinations of chemicals. The open literature contains numerous lists of th e reactivity of different types of chemical combinations. Table 8.3 pres ents examples of combinations of compounds that are known to be reactive. More complete discussions and lists of highly energetic chemical interactions are published by CCPS (Ref 8.13 CCPS 1995), Yoshida (Ref 8.84 Yoshida), Medard (Ref 8.58 Medard), FEMA (Ref 8.40 FEMA), and Bretherick (Ref 8.9 Bretherick). A complete review of the topic of chemical reactivity is also available from (Ref 8.13 CCPS 1995) (Ref 8.20 CCPS 2003a). The Interaction Matrix. The chemical interaction or reaction matrix is a recognized useful hazard identifica tion tool. These matrices simply list the materials onsite or in a give n operation on both axes of a table/matrix with the intersection no ted in the body of the matrix to indicate where the two materials can react adversely with each other. Notes are used to clarify the condit ions of the reactions and hazards presented by them. Since a matrix is a two-dimensional tool, it cannot directly show combinations of more th an two materials. Table 8.3 is an example of chemical reactivity matrix. Matrices are shown in CCPS Essential Practices for Managing Chemical Reactivity Hazards (Ref 8.20 CCPS 2003a). A general method for producing a plant/site–specific chemical compatibility matrix is given by the American Society for Testing and Materials (ASTM) International (Ref 8.7 ASTM), and an example matrix for a process is presented by Gay an d Leggett (Ref 8.41 Gay). Necessary utilities, such as inerting nitrogen, and the materials of construction should be listed as components in a matrix, as should the operator and other populations that are impacted by the process. Tests or calculations may be appropriate if the effect of an interaction identified in a matrix is unknown as described below. The U. S. Coast Guard (Ref 8.78 USCG) has published a comprehensive general chemical compatibility matrix for use in determining the possible reactions between different types of

3 Normal Operations 3.1 Introductio n This chapter covers the transient operating modes associated with the normal operations associated with the everyday production of materials. Although there are engineering controls designed for day-to-day operation (discussed in Section 3.2), administrative controls are needed, as well, to effectively manag e safe start-ups and shut-downs. Since procedures are essential fo r being able to operate a process consistently and safely, especially during the transient operating mode, this chapter includes a brief overview of procedures in Section 3.3. This chapter continues with di scussions on how normal shut- downs and start-ups are performed safely (Section 3.4 and Section 3.5, respectively). Then some lessons learned from incidents which occurred during normal shut-downs and start-ups are covered in Section 3.6. This chapter concludes wi th a discussion on the applicable RBPS elements for the normal shut-downs and start-ups in Section 3.7. The next two chapters on process and facility shutdowns tie in to this chapter to complete Part I of this guideline on normal operations. 3.2 The normal operation Process safety applies to all mode s of operations, including normal operations. For this guideline, normal operations are defined in Table 2.1 as: The operating mode when the process is operating between its start-up and shut-down phases an d within its normal operating conditions. Guidelines for Process Safety During the Transient Operating Mode: Managing Risks during Process Start-ups and Shut-downs . By CCPS. © 2021 the American Institute of Chemical Engineers

246 10.5.4 Inherent Safety Review Team Composition The composition of the inherent safe ty review team will vary depending upon the stages of the developme nt cycle and the nature of the product/process. The team composition is generally four to seven members selected from the typical skill areas checked in Table 10.1. Knowledgeable people, with an approp riate matrix of skills, are required for a successful review effort. The industrial hygienist/toxicologist and chemist play key roles on the team to ensure understanding of the hazards associated with reactions, chemicals, intermediates, and products, and to explain hazards, es pecially where there may be choices among chemicals or processes. Often, an organization will strive fo r the elimination of a specific toxic material from a given process, such as chlorine from a water or wastewater treatment proc ess. Alternatives will also have other hazards and risks that require an informed choice. The industrial hygienist, chemist, and safety engineer play an important role in generating the information needed to sele ct between alternatives. 10.5.5 Inherent Safety Review Process Overview As previously stated, good preparatio n is very important for an effective inherent safety review, particularly for a new process. Preparation for the review is summarized in Figure 10.2 and includes the following background information: 1.Define the desired product. 2.Describe optional routes to ma nufacture the desired product (if available), including raw materi als, intermediates, and waste streams. 3.Prepare simplified process flow diagram and tentative layout. Include alternative processes 4.Define chemical reactions. Desired and undesired Determine potential for runawa y reactions/decompositions. 5.List all chemicals and materials employed Develop a chemical compatibility matrix. Include air, water, rust, byproducts, contaminants, etc. 7.Define physical, chemical, and toxic properties.

-*45
0'
'*(63&4




























































xxi Figure 10.24 Incident Sequence 249 Figure 10.25 Complete Causal Factor Chart for Fish Kill Incident 250 Figure 10.26 Top of the Predefined Tree 251 Figure 10.27 First Question of th e Human Performance Difficulty Category 252 Figure 10.28 Human Engineering Branch of the Tree 253 Figure 10.29 Analysis of the Human Engineering Branch 254 Figure 11.1 Common Human Factors Model 263 Figure 11.2 Example of Poor Pump and Switch Arrangement 264 Figure 11.3 Incident Causation Model 272 Figure 12.1 Incident Investigation Recommendation Flowchart 279Figure 12.2 Layers of Safety 287 Figure 12.3 Bow-Tie Barrier Method 288 Figure 12.4 Example Recommendations and Assessment Strategies 292 Figure 14.1 Flowchart for Implementation and Follow-up 316 Figure 16.1 Example Safety Alert 349 Figure 16.2 CCPS Process Safety Beacon 350 Figure 16.3 ICI Safety Newsletter No. 96/1 & 2 351 Figure 16.4 ICI Safety Newsletter No. 96/7 352 Figure 16.5 Learning Event Report
&YBNQMF 353 Figure 16.6 1SPDFTT
4BGFUZ
#VMMFUJO
&YBNQMF 354

10 • Risk Based Process Safety Considerations 190 involved in developin g and sustaining parts of the RBPS elements. Element 5—Stakeholder Outreach : The company benefits from dialogues with the public, emer gency responders, professional groups, and regulatory agencies by sharing information on the plant’s process hazards, the methods used to safely manage risks, and how to respond sa fely if a loss event occurs. 10.2.2 Pillar II: Understand Hazards and Risk” The “Understand Hazards and Risk pill ar is the technical foundation of a risk-based approach, establishing the technology framework that is essential when developing how the risks are managed (Pillar III) and monitored (Pillar IV). Everyone should understand what the process hazards are and how the risks have been assessed. The two elements within this pillar are: Element 6—Process Knowledge Management : This element focuses on the recorded information, that includes verified, accurate, up- to-date, technical documents, engineering drawings and calculations, and specifications used to fabricate and install the process equipment. This informat ion is accessible to everyone working with the hazardou s materials and energies. Element 7—Hazard Identificati on and Risk Analysis (HIRA) : The technical information is used in th is element to help identify the process hazards and their potenti al consequences. Each company has a risk tolerance level, helping each plant assess their risks using qualitative or quantitative ap proaches. The risk analyses are used to establish the engineering and administrative controls needed to help safely manage the risks. 10.2.3 Pillar III: Manage Risk The “Manage Risk” pillar contains the elements needed for safe operations, using and applying the technical knowledge and risk analyses from Pillar II. Once the process hazards and risks are

252 | 7 Sustaining Process Safety Culture On a continuing basis, plans, budgets, personnel, and leadership should be considered with the intent of sustaining culture and PSM S perform ance. With succession planning and continuous professional development, the organization should build a cohesive team allowing the culture to sustain even as the organizational structure and personnel change. This all takes com mitment to process safety, leadership, and continual improvement, along with all the other culture core principles, which needs to be renewed from time to time. Process safety culture needs to remain strong during economic downturns and upturns, and when a facility or company is about to be closed or sold. These are the sternest tests for the leaders of process safety culture: to keep everyone’s focus on doing the right thing as they have been, despite the stresses that occur during these often-wrenching changes. Process safety culture requires real leadership. When setting process safety goals, some companies favor priority goals, stating “Safety First” or “Nothing is More Im portant than Safety.” Others prefer goals that are more concrete, targeting “Zero Incidents,” or “X% Reduction in incidents,” sometimes with a time target. Still others, despairing of ever reaching or staying at zero, favor continuous reduction goals with slogans like “Drive to Zero.” Ultimately, any of these will do. Continuous improvement to zero process safety incidents is possible. Even if there is a tem porary setback in performance, it m ust be taken as a sign that culture efforts m ust be redoubled, not abandoned. However, even if zero incidents are attained, the process safety culture journey continues. The improvement of the culture should be relentless. Good luck on your process safety culture journey, and thank you in advance for your leadership.

2 • Defining the Transition Times 24 Figure 2.1 The anhydrous ammonia release incident at start-up. [18]

Piping and Instrumentation Diagram Development 88 6.7 Pipe Route Generally speaking, the real pipe route cannot be seen on the P&IDs. There are cases, however, that a special route should be considered for some routes based on engineering considerations. Such special pipe routes should be formally communicated with the Piping group (to develop proper pipe models and isometric drawings) to make sure they will be implemented in the plant during construction. Because this route cannot be shown on the P&ID, such special requirements for pipe route should be captured on the P&IDs in the Notes. A few of these special requirements are sloped, no liq- uid pocket, no gas pocket, free draining, vertical, hori-zontal, and minimum length or distance. 6.7.1 S lope The slope on pipes can be important for horizontal pipes if a liquid or a two‐phase flows with a liquid component flowing inside of them. There is no slope for vertical pipes, and slopes on gas flows are not important. Generally speaking, horizontal pipes are used without a slope for different reasons, including the difficulty in handling the pipe elevation and pipe supports. Therefore, by default no horizontal pipe will ever have a slope.There are two features that should be specified on a pipe symbol: the direction of slope and the slope magnitude. The slope direction is the direction of flow (direct slope) or against the flow (reverse slope). The slope direction is shown on a pipe with a triangle symbol in Figure 6.45. The other feature of a slope is its magnitude. From theoretical point of view, a slope can be specified by a percentage, by a fraction, or by the angle. However, angles (e.g. in degree) are not used to specify slope. A slope is specified either by a percentage or more practi-cally by a fraction based on standard lengths of a pipe. The slope magnitude can be specified on the slope sym-bol on the P&ID as 0.5% or 2/12. Here, 12 feet is the standard length of pipes (Figure 6.46). Level signal From LT-1159 From wash water tank To wash water tankPID-300-1001 Min. flow PID-300-1001To wash water pre-heate rWash water PID-300-1003Wash water Symmetric piping Seal flush plan 21WAT-A A-8 /uni2033-3015 3/4/uni2033 3/4/uni2033 2/uni2033 1/uni2033 1/uni2033 1/uni2033 1/uni20331/uni2033 3/4/uni2033M M 3/4/uni20331/uni20331/uni20331/uni20331/uni2033 1/uni20336×4 6×4 6×4 300-P-130 300-P-1408×6 8×66×4 TSS TSS3/uni20334×34 ×3 FOFCIS OS FC 130 FV 130LC 131 LV 131 FE 130FE 140 PG 130 PG 141PG 140 PG 131WAT -A A-4 /uni2033-3014WAT -A A-4 /uni2033-3017 WAT-A A-6 /uni2033-3018WAT-A A-6 /uni2033-3016WAT-A A-8 /uni2033-3016PID-300-1001 Figure 6.44 An e xample of symmetrical piping. Figure 6.45 Slope symbol on pipes. 0.2% Figure 6.46 Slope symbol and slope magnitude on pipe .

304 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Many software packages are available to facilitate consequence analysis. Some are available for free as they support modeling requir ed by governmental regulations. Others are proprietary and typically offer greater capabilit y, ease of use, and technical support. These models are sometimes deceptively easy to use. Care should be taken in understanding the limitations of the model and in gathering the be st data for use in the model. Every model has uncertainties. Selecting data that will yield a cons ervative, and still realistic, result is the best approach. Other Incidents This chapter began with a description of the DPC Enterprises toxic release. Other incidents relevant to consequence analysis include all th e incidents listed in Chapters 4, 5, and 6. Exercises List 3 RBPS elements evident in the DPC Enterprises L.P. chlorine release incident summarized at the beginning of this chapter. Describe their shortcomings as related to this accident. Considering the DPC Enterprises L.P. chlorine release incident, what actions could have been taken to reduce the risk of this incident? List the stages of a consequence analys is in the order they are conducted. What hole size and leak duration should be used in a source model? Why? An LNG storage tank leaks into the diked area around the tank. Describe the expected transport and consequence effects. A refinery HF acid unit has a release of isobutane and HF acid. Describe the expected transport and consequence effects. What 2 atmospheric conditions (wind direct ion, wind speed, and atmospheric stability) would you select to represent the conditions illustrated in the wind roses in Figure 13.15?

186 Guidelines for Revalidating a Process Hazard Analysis Topic/Question Y/N MOC No. Have personnel been relocated such that responding to equipment deficiencies (e.g., shutting down a pump with a leaking seal) may take a significantly longer time? Human Factors Changes Has the operator training pr ogram deteriorated since the previous PHA? Have maintenance practices deteriorated since the previous PHA? Have inspection and test practices deteriorated since the previous PHA, particularly for standby safety systems (e.g., interlocks, relief valves)? Have chains of command or delegation of authority for critical positions deteriorated? Have control displays changed such that information necessary to diagnose and respond to upset conditions is not readily accessible? Have operator communication systems changed? Are equipment and piping system labels being maintained? Are color coding systems for piping and components being maintained? Have emergency alarm or notification systems changed? Have safe work practice, such as those listed, deteriorated since the previous PHA? • Confined space entry • Entry into process area • Lockout/tagout • Line breaking • Lifting over process equipment Have fire detection or suppression systems been modified such that they require a different operator response? Have facility modifications made alarms difficult to see or hear? Have facility modifications resulted in the potential to overwhelm an operator with alarms during upset conditions?

440 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Supply appropriate information to the contra ctor to ensure that the contractor can safely provide the contracted services. Contractual relationships can extend many laye rs as sub-contractors engage in their own contracts with other sub-contractors. It can be difficult, and very important, to ensure that process safety (and other topics such as qua lity) expectations are communicated through the layers of contractors. As an example, Figure 21.4 depicts the primary contractual relationships involved in the Deepwater Horizon incident. Workforce Involvement Personnel, at all levels and in all positions in an organization, have a role in the safety of the organization’s operations. Workforce involvement provides a system for enabling the active participation of company and contract workers in the design, development, implementation, and continuous improvement of process safety. Those personnel directly involved in operating and maintaining the process are those most exposed to the hazards of the process. These workers are also frequently the most knowledgeable people with respect to the day- to-day details of operating the process and maintaining the equipment and facilities. They may be the sole source for some types of knowledge gained through their unique expe riences. Workforce involvement provides management a formalized mechanism for tapping in to this valuable expertise. This proactive engagement would illustrate at least two positive things, the right people are involved in the review, and the workforce, down to the operating staff, is able to provide candid views without fear of adverse consequences. Workforce involvement is specifically mentio ned in the U.S. OSHA PSM and U.S. EPA RMP regulations (termed “employee participation”) as well as other regula tions regarding process safety. This is a requirement for the presence of people with “experience and knowledge specific to the process being evaluated” at a ha zard identification study of a covered process. Proactive companies will expand that to include workers directly involved in maintenance and operations at these reviews and will encourage th eir honest input. Subject matter experts such as process engineers, mechanical engineers, mate rial engineers, etc. should be relied on for technical information and other process safety information. Operators and mechanics should be relied on for evaluating the understanding of the process, the clarity and efficiency of procedures, and an understanding of what is being done in the field vs. what engineering and management think is being done. Some opportunities to involve the work force include as members of a hazard identification study team, in writing operating procedures and safe work procedures, during management of change reviews, and on incident investigation teams. Many companies implement a stop-work authority which allows anyone to stop an activity, without fear of repercussion, if they believe the activity is potentially dangerous. This can be seen as an ultimate work force involvement activity.

318 | Appendix E Process Safety Culture Case Histories odors offsite and exceeding the perm itted level. The process would also periodically generate significant noise at a decibel level and frequency that was very irritating beyond the fence line. One Earth Day, the plant dutifully held an open house to show off their state-of-the-art facility and show how they had all but elim inated their process and office waste. The neighbors were not as interested in this and their questions quickly turned to the site’s odor and noise. One neighbor asked, “What is happening in the plant when em issions increase?” The technical m anager hesitated, and then with encouragement of the plant manager explained that it happened when the gas rate was high. The neighbor then asked the same question about noise. The technical manager explained that it happened when the gas rate was low. Another neighbor then asked, “So if you know why there are the odors and noise, why not avoid those conditions?” The technical manager explained that they were trying to, but it was not a direct correlation and they had not yet figured it out, but they would keep looking for a solution. The meeting ended am icably, and a few days later, the technical m anager received a call from the neighbor saying, “Your gas rate is low.” The technical m anager signaled through his office window and the operator turned up the gas. The neighbor, noticing the immediate reduction in noise replied, “Thank you.” Over the following months, with neighbor input, the plant got the gas rate under control and achieved the lowest rate of emissions in the com pany. This exam ple shows how the linkage between culture and the PSMS elem ent of stakeholder outreach. What positive culture attributes did the plant m anager exhibit in being open with the com munity? Establish an Imperative for Safety, Provide Strong Leadership, Ensure Open and Frank Communications, Foster Mutual Trust.

Piping and Instrumentation Diagram Development 14 may say: “for this P&ID sheet there is a need to expend 100 hours from the IFR to IFC. ” It is important to know the time required to develop a P&ID sheet comprises five time spans: 1) Time t o develop the technical content of P&ID. 2) Time t o get the drawing drafted. 3) Time r equired by the reviewers and approvers to check and sign the P&ID.4) Time r equired to implement the reviewers’ comments on the P&ID. 5) Time t hat a “Document Control group” needs to offi- cially issue the P&ID sheet. The first item is the technical component of the man‐ hours, which relates to developing the technical content of P&ID.

216 | 6 Where do you Start? providing feedback to facility personnel. Critical judgments should be avoided. Watch the tim e. Try to complete the interview on tim e. If extending the time would be beneficial, ask, “This is taking a bit longer than we planned, would another 10 minutes be okay?” Re- schedule if necessary. Additional techniques to help build rapport: Maintain eye contact . This connotes interest in and attention to the interviewee, and allows the interviewer to m ore easily read body language. Maintain a comfortable distance . Sitting close to the interviewee can m ake them uncom fortable while sitting too far may create emotional distance. The com fortable distance between people varies widely around the world and even within countries. Interviewers should understand this before starting interviews. Mirror the interviewee. Approximately matching the tone, tem po, and body position of the interviewee can foster rapport between the interviewer and interviewee if it is done subtly and does not look calculated. Business cards. The interviewer may present a business card as part of the introduction process. This can help establish them as independent while also conveying that they are not hiding anything. If a card is presented, it should be done casually, saying for example, “If you have any further thoughts, just call me.” When interviewing senior managers, and in cultures where business cards are traditionally shared, a more formal presentation of cards may be appropriate. Interviewer r eactions . Interviewers should avoid positive or negative reactions to what the interviewee has said, whether verbal or non-verbal. Positive feedback about the level of sharing m ay be helpful where warranted, including nodding and sm iling, but all negative feedback should be • • • • •

CASE STUDIES/LESSONS LEARNED 205 This was changed in August 1992 to a “matrix” system where the plant manager role was eliminated and the senior operators were appointed as team leaders, reporting to an area manager. Staff had to apply for their new job roles and the new system was not fully understood. Further details of these ch anges are provided in an IChemE Loss Prevention Bulletin article, Failure to manage organisational change – a personal perspective (Lynch/IChemE 2019). 7.3.5 The Incident PREPARATION The site shift manager and area manager discussed the removal of residue from 60 Still Base on Thursday, 17 September 1992. Over the following weekend, the contents were distilled to reduce the level of volatile MNTs in the usual way. This was completed and the still base was allowed to cool before the remainin g liquid material was pumped out to storage. At 09:45 on Monday 21 September, the area manager instructed staff to apply steam to the heater batte ries, in order to soften the sludge, and advised that the temperature wa s not to exceed 90 °C. By 10:15, a skip had been obtained to hold th e waste sludge, and a scaffold was erected beneath the manhole cover. A permit to work was issued by the team leader, initially to a fitter, although he then left for an earl y lunch. A second permit was then provided for the operators to remove the manhole cover, which took about 30 minutes. Some of the materi al was scooped out of the vessel and found to be gritty, with the co nsistency of “soft butter”. It is understood that the material was no t tested further, and the area manager assumed it was a thermally stable tar. RAKING Two process technicians then started to remove material using a 2.5m long metal rake that was “found” on the ground nearby. After about an hour, (at approximately 12:50) a 2m length of sludge had been removed and the team decided to make an exte nsion for the rake so it could reach further back into the still base. At th e same time, the fitter returned from lunch and noted that the inlet to the still base had not been sealed off.

4.5 Process Safety Culture Metr ics |149 Does the scope and boundaries of the PSMS cover all hazardous materials or processes or are som e om itted because they are not covered by regulation? Is the facility using recognized good engineering practices to control hazards and maintain equipm ent and safeguards? In the USA. and other countries that follow the PSM regulation, are RAGAGEPs identified and followed? Are people with the right experience, skills, and perspectives being assigned to HIRA/PHA, MOC review and incident investigation teams? Gaps indicate weakness in ability to understand and act upon hazards and risks. Is process safety knowledge up-to-date? Measurements can include num ber of completed M OCs where process safety knowledge has not yet been updated, the length of time following MOC to update process safety knowledge, and the length of tim e since the last update. In more advanced cultures, have the risks determ ined through HIRA/PHAs been used properly to determine the levels of effort and evaluation of other RB PS elements? A key aspect of RB PS is that processes and operations with higher risk should receive greater attention in the asset integrity effort, have a m ore detailed and higher level of approval in MOC and Operational Readiness, more specific training, etc. Likewise, lower risk processes and operations m ay receive less attention, but of course should not be ignored. Evidence that the permit-to-work process is insufficient. A review of com pleted permits should show whether job safety analyses were performed with adequate risk analysis and appropriate isolation and safeguarding were perform ed. Evidence of inadequate MOC. This can be determined by com paring change orders to MOC documents to look for changes claimed to be replacements-in-kind that were not • • • • • • •

OVERVIEW OF RISK BASED PROCESS SAFETY 47 described above. More formal training needs analysis helps underpin a necessary training program. This should include process safety hazards and how to participate in or interpret risk analysis studies, as appropriate. Formal testing of knowledge and skills is an important part of this element to assure that participants have understood the material. It includes on-the-job task verification. RBPS Element 13: Management of Change Management of Change (MOC) is a system to identify, review, and approve modifications to equipment, procedures, raw materials, processing conditions, and people or the organization ((CCPS, 2013b), other than replacement in kind. This helps ensure that changes are properly assessed (for example, for potential safety risks), authorized, documented, and communicated to af fected workers. Documentation includes changes to drawings, operating and maintenance procedures, training material, and process safety documentation. The objective is to prevent or mitigate incidents prompted by unmanaged change. Many past incidents have been due to changes that were not properly assessed, and which defeated existing safeguards or introduced new hazards. For example, the Piper Alpha platform was modified after start-up to include gas recovery, but without sufficient risk assessment of the higher risks this introduced (Broadribb, 2014). The MOC is a formal process involving similar tools to the initial hazard identification and risk assessment. A newer aspect of MOC is the recognition that organizational change can also create process safety issues and a specific Management of Organizational Change (MOOC) procedure has been developed (CCPS, 2013b). Well operations are subject to changes as the process of drilling is dynamic. Unexpected geological conditions may be encountered that make the initial well plan invalid and require an update. This should be the subject of an MOC review to ensure new hazards are not introduced by the change. RBPS Element 14: Oper ational Readiness Operational readiness evaluates the process be fore start-up to ensure the process can be safely started. It applies to restart of facilities after being shut down or idled as well as after process changes and maintenance, and to start-up of new facilities. An important aspect is to verify that all barriers identified in design reviews and captured in a risk register and action track ing system have been implemented and/or any outstanding actions are approved for later close-out. An aspect for upstream in onshore remote locations and offshore in GOM is the need to shut down and de-man during severe weather events (hurricanes) and hence to restart more frequently than is the normal for process facilities. Similarly, startup of not normally manned operations requires special attention.

79 | 6.2 Seek Learnings Each reader and leadership team might make different prioritizations than just described. That is perfectly acceptable. The order in this example was colored by the authors’ experience, and each reader’s experience is almost certainly different. The bottom line is that any effort to learn from external incidents should be based on company goals and priorities. This will help deliver the greatest value for the effort. 6.2 Seek Learnings Having been guided by the company’s process safety improvement goals, the evaluator seeks internal and external incident reports that have the potential to provide important learning. How do you identify the most relevant incidents external to your company? Start with the Appendix of this book. You can search for publicly reported incidents based on the management system elements and culture core principles that appear to have failed in each incident. You can also search on selected causal factors. Additionally, the electronic version of the index enables you to search based on type of equipment and industry, and you can search on multiple parameters. These features allow quick identification of incidents with findings relevant to the improvement goals. Other free public sources of incident reports exist, including those mentioned in Chapter 2. Conference and journal papers can be obtained, sometimes for a small fee. A simple Internet search may also reveal additional useful case histories. Some countries have a policy of retaining government investigation reports outside of the public domain but provide ways to request copies. If the report is not in your primary language, various free translation services available online are generally are good enough to allow you to decide whether it would be valuable to obtain a more formal translation. Finally, though many incident reports are in text format, several public investigation organizations produce reports as videos and screen presentations (such as PowerPoint®). Once you have identified the most relevant incident reports to study, find a place where you can maintain a steady focus, get comfortable, and read.

RISK MITIGATION 341 three manual buttons: one button to activate the double block, another to activate the bleed, and a third to activate the purge. On the day of the incident, the operator chose to shut down the reactor using the three manual buttons on the control panel. The activati on of these three buttons was equivalent to the activation of the manual shutdown button or the automatic shutdown. The first step was to close the process air valves to the reactor. The second step was to open the air bleed after the air to the reactor was blocked in. The third st ep was to activate the timed nitrogen purge. The operator pushed the first two buttons, but mistakenly did not push the inert gas purge button. The standard operating procedure, or SOP for this critical step was not followed by the operator. Failure to initiate the inert gas purge allowed the flammable contents of the reactor, including the catalyst, to enter the air sparge r system. The air sparger pipe contained an inventory of reactor contents for about a day. As the reactor was started up on November 14 and approached start-up temperature, an explosion occurred in the air sparger inside th e reactor. Oxygen was available because the reactor had not been purged, fuel was availabl e from the reactor contents, and the ignition source was probably the catalyst that was plated on the inside of the air sparger. 1987 was a time when industry was converti ng from pneumatic control to computer control. The use of push buttons was common. As a result of the Pampa incident, Celanese implemented a detailed risk assessment and ri sk management methodology to identify and mitigate risks including the ones described at Pampa. Another aspect of the process safety management system includes rigorous controls around safety instrumented systems designed to mitigate similar hazards. Lessons Process Safety Competency. The shutdown system activated an indicator light when the shutdown started and another light when the shutdown and purge were complete when either the automatic system or the one button manual system was activated. When the three-button manual shutdown was used, the system gave no st atus feedback. In order to detect the lack of inert gas purge, the next shifts would have had to detect the absence of the purge from the computer activity log printed in another room. Current knowledge of human performance recognizes that one should not design a system in which a single error can lead to potential catastrophic consequences. Also, there should be obvious feedback systems for operations for critical actions. Hazard Identification and Risk Analysis. The independent, manual shutdown buttons were identified in a Process Safety Review prio r to the event as a potential source of human performance issues which could adversely impact shutdown, but changes were not recommended. The consequences of not doing the purge were well understood, but because the scenario was well known, the review team underestimated the likelihood of the error. The initiating event was that the operator neglected to start the inert gas purge cycle during shutdown.

6 • Recovery 110 Figure 6.4 Example abnormal situ ation fault detection model. 6.4.3 Response to expected deviat ions with successful recovery The process is designed to make quality products during normal operations of a continuous proc ess; for batch pr ocesses, quality product is extracted and separated, as needed, from the batch vessel at the end of the production run. Suc cessful recovery ef forts result in continued production, exhibiting the recovery behavior as was illustrated in Figure 6.3 for a cont inuous process. When everything is working well and as designed, it is a good and safe production day at the facility—in other words, the control of its hazardous materials and energies have been effectively managed. 6.5 Incidents and lessons learned When there are established, effectiv e, normal recovery procedures, no process safety-related incidents sh ould occur during this time. When abnormal situations repeat, they are likely to be leading indicators of Lower Repair CostsHigher Good Equipment Condition PoorFailure BeginsPerformance Degradation Sensed Condition Indicators Visible or Audible Indicators Failure

317 R Designing for Operation The INSET Toolkit instead recomme nds a multi-attribute decision analysis technique to evaluate the overall inherent SHE aspects of the various process options. The INSET Toolkit is particularly interesting as an Inherent SHE measurement tool because it represents the combined consensus expertise of a number of companies and organizations, and because it is intended to consider safety, health, and environmental factors in one set of tools (Ref 12.10 Hendershot 1997). Gupta and Edwards (Ref 12.8 Gupta 2003) also propose an ISD measurement procedure for use in differentiating between two or more processes for the same end product. Palaniappan, et al. (Ref 12.15 Palaniappan 2002; Ref 12.14 Palani appan 2001) have developed an i- Safe index, which includes five supplementary indices: Hazardous Chemical Index (HCI), Hazardous Reaction Index (HRI), Total Chemical Index (TCI), Worst Chemical Index (WCI) and Worst Reaction Index (WRI). Lastly, CCPS (Ref 12.2 CCPS 1995) presents a large range of decision aids for risk analysis and decision making in industry, including cost-benefit analysis, voting methods, weight ed scoring methods and decision analysis, so that critical risk decision s can be made in a more consistent, logical, and rigorous manner. 12.2 SUMMARY This chapter outlines how various tools can be used to support the implementation of inherent safety. Inherent safety considerations are particularly important in conducting hazard reviews (i.e., a preliminary hazard analysis), because of the im portance of considering inherent safety early in the design sequence when changes can most readily be made. Qualitative PHA techniques, su ch as the HAZOP, What-If, and Checklist methodologies, as well as consequence-based methods such as the Dow Fire and Explosion Index , can be adapted for use in IS reviews. In addition, consideration of the human-machine interface should be made as early as possible in the desi gn phase of a process. In the next chapter, the discussion will turn to conflicts wh ich may arise during the application of inhere ntly safer design.

393

164 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION Valves. Control valves, emerge ncy isolation valves, check va lves, unit isolation valves and other types of valves. In addition, the failure position of valves on loss of instrument air or electrical power is shown. Safety systems. Pressure relief valves, de pressuring systems, connections to flare systems. Figure 10.3. Example piping and instrumentation diagram (AIChE 2019)

2 • Defining the Transition Times 20 Table 2.2 Definitions for the transient operating modes. A transient operating mode: the time when the process equipment is being shut-down, stopping the equipment and taking the equipment from its normal operating conditions to end normal operations. For a Continuous Process (open, steady-state systems): A planned series of steps to stop the process at the end of normal operations, taking the process equipment from its normal operating conditions to an idle, safe, and at-rest state. For a Batch Process (closed, unsteady-state systems): A planned series of steps to stop the process at the end of normal operations, taking the process equipment from its final operating conditions to a clean, idle, safe, and at-rest state in preparation for the next batch. A transient operating mode: The time when the process equipment is being restarted (or started after being commissioned), taking the equipment to the normal operating conditions for normal operations. For a Continuous Process (open, steady-state systems): A planned series of steps to take the process equipment from an idle, safe, and at-rest state to the normal operating conditions. For a Batch Process (closed, unsteady-state systems): A planned series of steps to prepare for and begin a batch process, taking the process equipment from a clean, idle, safe, and at-rest state to the normal operating conditions. A transient operating mode: The time that may require procedures in addition to the normal shut-down procedures for stopping the equipment in preparation for a planned project or maintenance shutdown. Note: If other groups are involved in the planned shutdown, such as engineering, maintenance, or contractors, special permits and handover procedures must be in place beforehand, as needed, before performing the shutdown-related activities.3Shut-down designed for a planned shutdown1Shut-down or Normal shut-down 2Start-up or Normal start-up

Table 26-3: Human Factors investigation tools Investigatory tools Description of tool Root Cause Analysis Root Cause Analysis seeks to: • Determine what happened. • Determine why it happened. • Identify what to do to reduce the likelihood of reoccurrence. Root Cause Analysis looks at physical, human, and organizational causes. It consists of five steps: 1. Define the problem. 2. Collect data e.g., the impact. 3. Identify possible causal factors. The “Five Whys” tec hnique can be used here, where the question “Why did this happen?“ is asked five times to explore all possible causes. 4. Identify the root causes. 5. Identify and implement solutions. Fish Bone Diagrams Fish Bone Diagrams consist of the following steps: 1. Define the problem/effect. 2. Identify major factors involved e.g., equipmen t, process, people, materials, environment. 3. Identify possible causes. 4. Analyze the diagram.

5 • Facility Shutdowns 83 in the project’s life cycle. In additi on, if a facility shutdown contains specially-protected equipment (e.g., it has been mothballed during the shutdown period since it was not being worked on), the measures implemented to preserve the equi pment should be addressed when pre-commissioning the equi pment (Section 5.5.2). 5.5.2 Recommissioning mothballed equipment The recommissioning of mothballed equipment will depend upon how long the equipment has been moth balled and how well the prevention techniques, if any, were maintained (Section 5.3.4). In all cases, a multi-discipline project team should be assigned to inspect and test the mothballed equipment to determine its integrity. The preservation approaches used will need to be validated (e.g., rotating the motors) or reversed (e.g., coating the equipment with oil), as needed. For mothballed equipment that had been in operation and was shutdown, the recommissioning team should perform an Operations Readiness Review (ORR) before the operations group restarts the equipment and its associated processes [14]. 5.6 Incidents and lessons learned Details of some facility shutdown-rel ated incidents are included in this section. The incident summary is provided in the Appendix.

Piping and Instrumentation Diagram Development 6 In a PFD, we see a only as one process step, although such a symbol brings to mind the concept of a tank. Also in some PFDs, there is only one symbol for different types of pumps including centrifugal or PD type (and this could be confusing for those who are not familiar with a PFD). When we show a symbol of a pump on a PFD, we are basically saying, “the liquid must be transferred from point A to point B, by an unknown type of pump that will be clarified later in the P&ID. ” A pump symbol in a PFD only shows “something to transfer liquid from point A to point B. ” Each symbol on a PFD has a general meaning and does not refer to any specific type of that equipment (Table 1.1). The other aspect of the concept of a PFD is that its lines do not necessarily represent pipes; they only repre-sent streams. This means that one line on a PFD could actually represent two (or more) pipes that go to two (or more) parallel units. Therefore, one main activity during P&ID development is deciding on the type of a piece of equipment. A general symbol of a heat exchanger on a PFD should be replaced with a specific symbol of shell and tube heat exchanger, a plate and frame, or the other types of heat exchangers when transferred to the P&ID. Sometimes the type of process elements is decided when sizing is done, but in other times this decision is left to P&ID developers. This activity is commonly overlooked because in many cases P&ID development starts with a set of go‐by P&IDs rather than blank sheets.Regarding text information, the main difference is in equipment callouts, which will be explained in detail in Chapter  4. Here, however it can be said that an equip-ment callout is a “box of data” about a piece of equipment shown on a PFD and on a P&ID. In a nutshell, the difference between equipment call- outs on a PFD and on a P&ID is that on the PFD the data are operational information, while on the P&ID, the data are mechanical information. As a PFD is mainly a pro-cess engineering document, all its information are for normal operations of a plant. The P&ID is a document that illustrates the capability of the equipment. Table 1.2 shows the differences between one‐item callouts on a PFD and on a P&ID. As can be seen in Table 1.2, the numerical parameters in a PFD callout are generally smaller than the corre-sponding parameters in a P&ID callout because on a PFD, only the “normal” value of a parameter is men-tioned, whereas in a P&ID the “design” value of the parameter is reported. There are, however, debates on the parameters that should be shown on a PFD callout. From a technical point of view, a PFD callout should not have the sparing philosophy, and also the reported capacity should be the capacity of all simultaneously operating parallel units. But not all companies agree on this. Table  1.3 summarizes the differences between PFDs and P&IDs.Tank Heat exchangerM M MM MPFD symbol PumpSymbolTable 1.1 PFD symbols compar ed with P&ID symbols.

Piping and Instrumentation Diagram Development 390 pipes with partial flow. In such cases if the budget is not enough, only the lower portion of the pipe (up to a certain angle) could be insulated. For valves there are different types of insulation extent. A valve could be insulated completely or partially. The complete insulation of a valve means insulating its body and its bonnet. As the main bulk of fluid is inside of the “body” of a valve the main heat insulation happens by insulating only the body of the valve. However, if the completeness of insulation is very important, it could be decided to insulate the bonnet of the valve too. The extent of valve insulation could be mentioned as an acro-nym beside or below valves as “B” (for body insulation) or “B&B” (for body and bonnet insulation). Obviously a valve that has insulation on its bonnet is more difficult for inspection and repair. Where there are heat tracers below the insulation it is easier to decide on the extent of insulation; as we have expended money on heat tracing, it is better to keep the equipment as insulated as possible. Table 18.9 lists priority parts for insulation for several types of equipment.18.3.4 Summary of I nsulation Table 18.10 shows the summary of requirements in dif - ferent insulation systems. 18.4 Utility Stations Utility stations (US) are pretty similar to potable water fountains that you may see in malls and other public areas. The difference is that US provides utilities more than only potable water. USs are boxes which contain different utilities for process plants during inspection, maintenance, and repair. They may also be used during normal operation but in short period of times. Whenever an operator needs to use any of these utilities, he needs to connect a piece of hose to the connection inside of the US and then open its valve. The utility in each US could be a steam, air, water, nitrogen gas, electricity, etc. All these utilities, except electricity, are named “process utilities. ” A schematic of a utility station is shown in Figure 18.9.Figure 18.10 shows two different ways of showing utility stations on P&IDs. On the left‐hand side of Figure 18.10 the utility streams in USs are plant air (A), steam (S), and plant water (W). On the right‐hand side of Figure 18.10 a US with an addi-tional utility stream of nitrogen gas (N) is shown.Table 18.10 Differ ent requirements in insulation systems. InsulationHeat tracingOther considerations Heat conservationYes No No Personal protectionYes (thin) No No Winterization Yes Yes Yes Figure 18.9 A utility sta tion on a P&ID.Table 18.9 Insula tion priority of items. Coverage of insulation (if insulation needed) Priorities Pipes Whole pipe but left out the flangesNot applicable Valves Partially Priority 1: body Priority 2: bonnet Vessels Whole vessel Tanks Partially Priority 1: wallPriority 2: roof Pumps Whole pump Not applicable Compressors Whole compressor Not applicable Drivers Electric motors: no, they generally need coolingSteam turbine: full insulationNot applicable Heat exchangersShell and tube: full insulationPlate: generally no insulationSpiral plate: full insulationAerial cooler: No insulationNot applicable Safety devicesPartially Priority 1: inlet and outlet pipesPriority 2: PSV body (but not rupture disk)

40 Human Factors Handbook Examples of cognitive bias are as follows: 4.6.2 Supporting cognition – where to find more information People can develop and effectively apply their cognitive skills by use of techniques such as “20 second scans”. People are trained and directed to pause before starting a task and to scan the wo rk site to identify anything unusual or unexpected. They should then think whether it is safe to start work or do they need to seek help or change their pl an of action. This pause reduces the possibility of “task focus” and reduces the possibility that a wish to “get the job done” causes someone to not see unexpected hazards. Chapters 16 and 17 cover how to identify tasks that may be prone to error and how to pre-empt these errors. Chapter 18 looks at ways to help people develop non-technical skills and detect errors and mistakes made by other people. Cognitive performance can also be aided by developing individual psychological skills, such as situation aw areness. More information can be found in Chapters 20 and 21. Chapter 21, Section 21.5 explores how to develop self- awareness and a quick-thinking response to task performance. Confirmation bias • After a person identifies a possible cause of a process upset, he/she then looks for information to support their opinion, while ignoring information that suggests a different cause. • The person goes with their original opinion, despite some information suggesting they are wrong. Authority bias • An operator has too much confidence in the opinion of those in authority, while not trusting their own feelings. • The person goes with a senior engineer’s decision despite feeling that the judgment is wrong.

80 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION What must be present for a dust explosion to occur? Find the values for Upper and Lower Flammable Limits in air, Upper and Lower Flammable Limits in oxygen and the Limiti ng Oxygen Concentration in Figure 4.18. Estimate the Lower Flammable Limit for a va por mixture of 0.4 mole fraction Methane and 0.6 mole fraction Ammonia. What is your reference for the pure chemical Flammable Limits? Show your work. What was the source of ignition in the BP Te xas City explosion and in the Imperial Sugar Dust explosion? How might these sources of ignition have been controlled? How are electrical sources prevented fr om becoming a source of ignition? Figure 4.18. Methane flammability diagram (For Problem 4.9) (Wiki) References API, American Petroleum Institute, Washington, D.C., www.api.org. API 520 “Sizing, Selection, and Installa tion of Pressure-Relieving Devices” API 521 “Pressure-Relieving an d Depressuring Systems” API RP 752 “Management of Hazards Associ ated with Location of Process Plant Permanent Buildings” API RP 753 “Management of Hazards Associ ated with Location of Process Plant Portable Buildings” API RP 756 “Management of Hazards Associ ated with Location of Process Plant Tents”