text
stringlengths 0
5.91k
|
|---|
9 • Other Transition Time Considerations 166
Decommissioning includes decon struction, when the equipment
or process unit is dismantled an d individual components of the
equipment or individual equipmen t from the process unit may be re-
used, and demolition, when the equipment or process unit is dismantled for scrap and or material recycling. When a project calls for
equipment to be decommissioned, transported to another site and recommissioned, the challenges brou ght about are a combination of
those noted in this chapter plus th e transportation-related issues.
As was noted earlier, if a company has decided that the equipment
will be partially dismantled or dismantled-in-place when it has
mothballed its equipment or processes, it is essential that the equipment’s condition is assesse d and then properly addressed
before attempting to reuse it. As shown in Figure 9.2, these
mothballing “steps” would be captur ed in stage 7A, with the proper
Operational Readiness Review (ORR) pe rformed in stage 7B before re-
commissioning the equipment and beginning operations again in stage 5. Incidents can occur if the equipment is not fit for its intended service.
Due to the different hazardous materials being handled, planning
for the decommissioning project, including hazard identification and risk management, should be in place to reduce the potential for
incidents that may cause injury and environmental damage. This
includes establishing a project management team, including consultants and contractors experienced in decommissioning, to manage the stages of the project. De commissioning and end-of-life of
equipment or facility may prompt higher focus on cost savings and
subsequently project-related short cuts. These short cuts inevitably
raise the risk profile and careful at tention should be undertaken to
avoid this mind-set.
As was noted earlier, handovers between engineering, operations,
maintenance, and the specialized decommissioning contractors |
INVESTIGATION M ANAGEM ENT SYSTEM 75
and perform queries of incident da ta to spot systemic trends. Additionally,
the management team’s endorsemen t of the incident investigation
management system is importan t when introducing a new or revised system.
4.3.1 Initial Implementation— Training
Implementation of a new or revised management system often begins with
presenting training for the four grou ps described earlier in this chapter.
1. Management
2. All employees in a position to notice and report all incidents
(including near-misses)
3. Incident investigation team members
4. Incident investigation team leaders
4.3.2 Developing a Specific Investigation Plan
The incident investigation management system should include guidance on
how to develop a specific investigation plan for an incident. The specific plan
should include leader and team se lection, a designated mechanism for
documenting the team activities, del iberations, decisions, communications,
and a record of documents requested, received, or issued. The objective of
the investigation plan should not be limited to identifying physical causes but extended to underlying management system issues.
The primary
objectives of a process safety incident investigation plan
should include:
• Identification of the physical causes—process and chemistry
• Identification of the PSM-re lated multiple root causes,
• Identification of recommendations to prevent recurrence, and
• Assistance in interpreting the re commendations or auditing their
implementation as needed
Figure 4.2 offers a typical checklist to use during the planning stage of an
investigation of a major complex incident. Low complexity incident
investigations do not always call for a formal plan.
|
xxviii GUIDELINES FOR MANAGING ABNORMAL SITUATIONS
Reliability
Centered
Maintenance A systematic analysis approach for evaluating
equipment failure impacts on system performance
and determining specific strategies for managing
the identified equipment failures. The failure
management strategies may include preventive
maintenance, predictive ma intenance, inspections,
testing, and/or one-time changes (e.g., design
improvements, operational changes).
Risk Based
Inspection A risk assessment and management process that is
focused on loss of containment of pressurized
equipment in processing fa cilities, due to material
deterioration. These risks are managed primarily
through equipment inspection.
Safe Operating
Limits Limits established for crit ical process parameters,
such as temperature, pressure, level, flow, or
concentration, based on a combination of
equipment design limits and the dynamics of the
process.
Safety
Instrumented
System A separate and independent combination of
sensors, logic solvers, final elements, and support
systems that are designed and managed to achieve
a specified safety integrity level. A SIS may
implement one or more Safety Instrumented
Functions (SIFs).
Safety
Management
System Comprehensive sets of po licies, procedures, and
practices designed to ensure that barriers to
episodic incidents are in plac e, in use, and effective.
Situational
Awareness The conscious dynamic reflection on the situation by
an individual. It provides dynamic orientation to the
situation, the opportunity to reflect not only the
past, present and future, but the potential features
of the situation. The dynamic reflection contains
logical-conceptual, imaginative, conscious, and
unconscious components which enables individuals
to develop mental models of external events. |
PROCESS SAFETY INCIDENT CLASSIFICATION 157
CCPS 2011, “Process Safety Leading and Lagging Metrics…You Don’t Improve What You Don’t
Measure”, Center for Chemical Process Safety, New York.
CCPS 2019, “Process Safety Metrics Guide for Selecting Leading and Lagging Indicators
Version 3.2”,
https://www.aiche.org/sites/default/files/docs/p ages/ccps_process_safety_metrics_-_v3.2.pdf.
IOGP 456, “Process safety – recommended prac tice on key performance indicators”,
International Association of Oil and Gas Producers, London, U.K.
NASA 2008, “That Sinking Feeling: Total Loss of Petrobras P-36”,
https://sma.nasa.gov/docs/default-sourc e/safety-messages/sa fetymessage-2008-10-01-
lossofpetrobrasp36-vits.pdf?sfvrsn=c4a91ef8_4 .
PSIE, CCPS Process Safety Incident Evaluation App, https://www.aiche.org/ccps/tools.
UN, “Globally Harmonized System Of Classifi cation And Labelling Of Chemicals (GHS) Sixth
Edition”, United Nations, New York and Geneva, 2011, https://www.un-
ilibrary.org/transportation-and- public-safety/globally-harmoni zed-system-of-classification-
and-labelling-of-chemicals-ghs-six th-revised-edition_591dabf9-en
|
CONSEQUENCE ANALYSIS 275
Liquid Discharge. For liquid discharges, the Bernoulli equation is used. The driving force
for the discharge is normally pressure, with th e pressure energy being converted to kinetic
energy during the discharge. Liquid head can also contribute to the driving force. For pipe
flow, the mass flux through the pipe is constant and, for pipes of constant cross-sectional area,
the liquid velocity is constant along the pipe as well. In all cases, frictional losses occur due to
the fluid flow.
Gas and liquid discharge equations contain a discharge coefficient which will affect the
discharge rate. A discharge coefficient (often 0.6 – 1.0) is applied to account for irregular hole
shapes compared to idealized circular sharp- edged holes. All discharge rates will be time-
dependent due to changing composition, temper ature, pressure, and level upstream of the
hole. Average discharge rates are case-depende nt, and intermediate calculations may be
necessary to model a particular release. The mass flow rate of two-phase flashing discharges
will always be bounded by pure vapo r and liquid discharges calculations.
Gas Discharges. Gas discharges may arise from several sources: from a hole at or near a
vessel, from a long pipeline, or from relief valves or process vents. Different calculation
procedures apply for each of these sources. The majority of gas discharges from process plant
leaks will initially be sonic or choked flow. The sonic discharge equation is used combined with
an estimate of the discharge coefficient. For ga s discharges, as the pressure drops through the
discharge, the gas expands. For gas discharges through holes, the mechanical energy balance
is integrated along an isentropic path to dete rmine the mass discharge ra te. A simple rule of
thumb for many pure materials is that the gas mass discharge rate is 10% of the liquid mass
discharge rate for the same conditions and hole size.
Two-Phase Discharge. When released to at mospheric pressure, any pressurized liquid
above its normal boiling point will start to flash and two-phase flow will result. Two-phase flow
is also likely to occur from depressurization of the vapor space above a volatile liquid,
especially if the liquid is viscous or has a te ndency to foam. For co nsequence modeling, the
discharge models must be selected to maximize the mass flux.
Flash, Evaporation, Aerosol, and Pool Spread Models
A discharge can be in the form of a gas, two-phase, or a liquid as shown in Figure 13.4. Figure
13.6 illustrates this point further. Aerosol formatio n is also possible for case B if the release
velocities are high. |
EQUIPMENT FAILURE 193
Heat Exchange Equipment
Overview. Heat exchange equipment is used to contro l temperature by transferring heat from
one fluid to another. Heat transfer equipment includes heat exchangers , vaporizers, reboilers,
process heat recovery boilers, co ndensers, coolers and chillers. Much of what is stated in this
section will also apply to heating/cooling coils in a vessel such as a reactor or storage tank.
Failures in heat transfer equipment can lead to loss of temperature control, contamination
of the fluids, or loss of containment. Temperatur e is frequently a critical process variable, so
failure of this equipment due to fouling, plugging , or loss of the heat transfer fluid supply can
lead to undesired consequences. A HIRA is need ed to assess the consequence. Heat exchange
equipment can see thermal stress due to temper ature gradients. This can lead to loss of
containment. The Longford fire and explosion in Chapter 12 is an example of this failure mode.
Heat exchanger failure modes include the following.
Fouling due to cooling water quality, low velocity, and microbiological fouling (both
aerobic and anaerobic)
Erosion
Stress corrosion cracking
Weld failures
Tube to tube sheet failures (r oll failure, seal weld failure)
Poor fluid distribution at the baffles
Heating or cooling media velocity not designed properly
Non-condensable material accumulation
Hazards associated with heat exchange equipment include:
Overpressure. due to blocked inlet/outlet streams
Inadvertent mixing of chemicals
Accelerated corrosion of downstream equipment, e.g.
acid gas leaks to cooling water can form acid which rapidly corrodes carbon steel
lack of oxidant in cooling water sy stem can cause rapid algae fouling
lack of oxygen control in a steam system can cause steel cracking
Chemical release from cooling towers
Water reactivity for heat exchangers us ing steam or water for heat transfer
The consequence of heat exchanger leaks depends on the nature of the process, the
direction of the leak (process side to utility or vice versa), and the fluids involved. Failure to
keep the fluids separate due to tube leaks can result in reactive chemical incidents (see
example 1), or release of a toxic or flammable ma terial into the low-pressure side where it can
escape elsewhere, such as at a cooling water tower. A tube rupture can result in a shell rupture
if the tube side is operating at significantly higher pressure than the shell. This risk is typically
mitigated with shell side pressure protection th at takes into account the tube side pressure.
Example 1. A plant had an explosion in the outlet piping of an oxidation reactor which
ruptured a 0.9 m (36 in) pipe (see Figure 11.12). The explosion was caused by the reaction of |
212 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION
Figure 11.27. Heater an d adjacent column at NOVA Bayport plant
(CCPS d)
Example 3. After a shutdown for maintenance, a hydr ogen reformer in an ammonia plant was
being restarted. In the normal start-up procedure at the plant, nitrogen gas is passed through
the primary reformer and a heating rate of 50 °C (122 °F) per hour is maintained at reformer
outlet. This nitrogen flows in a closed loop, that is, it is recycled back into the reformer. This
cycle continues until the temperature of 350 °C (662 °F) is obtained at the reformer outlet. To
increase reformer outlet temperature, more burners are ignited.
Because of an emergency shutdown, sufficient nitrogen inventory was not available at site
for startup. At least 8 to 10 more hours were re quired for nitrogen inventory makeup. To save
production loss, the startup procedure was initia ted. Furnace firing was started in the absence
of nitrogen gas, and reformer outlet temperatures were monitored for a 50 °C (122 °F) per hour
heating rate. Reformer outlet temperatures we re not increasing, so the firing rate was
increased. During this period, many alarms a ppeared on the control system for convection
zone temperatures. The alarms were inhibited to avoid any inconvenience to the control panel
operator, because he was busy with the steam drum level control. Since no changes occurred
in these outlet temperatures, the firing rate was further increased, and 56 of 72 burners were
fired. This represents about 70% of the heat in put, without any fluid flow through the reformer.
The board operator instructed the plant operator to have a physical check of the reformer. The
operator found that the reformer tubes were melting down inside the furnace.
The furnace was being fired and reformer outlet temperatures were being monitored
without introduction of any nitrogen through th e reformer. Because of the absence of any flow
through the reformer, its outlet temperature did not increase and the increase in heat with no
process flows resulted in high-tube temperatures and finally melting of the tubes. (Ramzan)
|
5.3 Process-Related Element Grouping |187
procedure through workflow management. A potential downside
of E-MOC is that it can seduce participants to act on their own
rather than m eeting with the full group of individuals involved in
the M OC, thereby reducing open and frank com m unication.
Therefore, leaders should m ake a special effort to encourage
com munication in the MOC process.
E-M OC system make it easier to proliferate the number of
approvers. More is not always better. When too many approvers
are listed, each may think that another one will catch any errors,
so they give the MOC a cursory review and approve it. If all
approvers take this approach, important issues will be m issed.
The MOC for process safety can be combined with the change
m anagement systems for other considerations, such as quality
and environm ent. This can make good sense for purposes of
efficiency and for cross-fertilization of ideas in the MOC reviews.
Care should be taken when doing this that process safety does
not get lum ped into occupational safety, for the many reasons
discussed throughout this book.
Emergency MOCs may be required from tim e to time to keep
the facility running when some component fails. With 168 hours
in a week during which approvers are likely to be in the plant less
than 60 hours, emergency M OCs are likely to occur on off-shifts
and require verbal approval. In a strong process safety culture,
emergency MOCs should be rare, and occur only when the risk of
not making the emergency change outweighs the risk of making
the change. If approval was given verbally, the proper
documentation of the verbal approval should be done as soon as
the approver returns to work. However, when em ergency MOC is
implemented, the process should be returned to its original state
as soon as possible.
Tem porary MOCs may also be required from tim e to time for
product or process trials or repairs. Temporary M OCs should be
planned and scheduled, and should not be conducted on an |
8.2 Bhopal, Madhya Pradesh, India 1984 | 109
equipment was faulty. And the uninformed action of emergency responders
who told residents to flee rather than shelter in place had fatal consequences.
Figure 8.3 Bhopal Scrubber (left) and Flare (right) (Source: Dennis
Hendershot, reproduced with permission)
The failures that led to the Bhopal disaster should be ingrained in our
memories. Perhaps the most important takeaway is that every barrier must
be maintained and functional. Furthermore, any process that requires as
many barriers as those found in the Bhopal plant should prompt decision-
makers to consider inherently safer design strategies, especially designs that
minimize the amount of hazardous materials present. The plant exercised this
strategy with its storage of phosgene, producing it on an as-needed basis and
storing a minimal amount. However, it did not do the same for MIC, which is
17 times more toxic than phosgene.
Another key takeaway from Bhopal is the importance of considering all
stakeholders, including the surrounding community. The public emergency
response plan was nonexistent. The company did not inform local police and
the public of the best actions to take in the event of an incident. On that fateful
evening, the police told people who were safely sheltered in place to evacuate,
sending them out into the toxic cloud.
Bhopal was a wake-up call that drove many countries to enact new laws
to improve process safety and enhance emergency preparedness. It also
|
194 Guidelines for Revalidating a Process Hazard Analysis
Q T R
Are calculations, charts, and other documents available that verify
siting has been considered in the layout of the unit? Do these siting
documents show that consideration has been given to:
• Normal direction and velocity of wind?
• Atmospheric dispersion of gases and vapors?
• Estimated radiant heat intensity that might exist during a fire?
• Estimated explosion overpressure?
Are appropriate security safeguards in place (e.g., fences, guard
stations)?
Are gates located away from the public roadway so that the largest
trucks can move completely off the roadway while waiting for the
gates to be opened?
Where applicable, are safeguards in place to protect high structures
against low-flying aircraft?
Are adequate safeguards in plac e to protect employees against
exposure to excessive noise, cons idering the cumulative effect of
equipment items located close together?
Is adequate emergency lighting provided? Is there adequate
redundant backup power for emergency lighting?
Are procedures in place to restrict nonessential or untrained
personnel from entering areas deemed hazardous?
Are indoor safety control systems (e.g ., sprinklers, fire walls) provided
in buildings where personnel will frequently be located, such as
control rooms and administrative buildings?
Are evacuation plans (from buildings, units, etc.) adequate and
accessible to personnel?
Are evacuation drills routinely conducted?
|
180 Human Factors Handbook
Table 16-1: Example of locks removed on wrong blinds
Example of a failure in task planning: Energy Institute – locks removed on
wrong blinds
What happened?
Four locked blinds under hazardous en ergy control (HEC) were removed from
the transfer line under the coke drums.
The blinds should have been left in pl ace for a confined space entry isolation
to the heater.
Three of the blinds were found hanging from the cables with the locks and
tags attached. A cable had been cut to remove the fourth blind.
The product could have leaked through the valves, entering the tubes inside
the heater.
Why did it happen?
The permitted scope was too broad. It covered two jobs and 11 different
blinds, which were generically referred to as “blinds”.
Unclear job plan.
Lack of communication.
Lack of clarity around removing locked blinds. A workaround allowed the same
crew to remove locked blinds when a hydro blind was leaking.
What did they learn?
It is important to define the field coordinator’s and operator’s roles and
responsibilities clearly to ensure blind verification is effective.
It is important to ensure the task desc ription includes a specific blind count
and blind locations.
The onboarding should be updated to ensure new staff know that locked
blinds should not be removed.
A procedure should be developed and implemented for hydro blind
management.
(adapted from [65])
16.3 Human Factors and task planning
Many tasks have specific safety requirem ents, such as operational turnarounds
and maintenance tasks. For example, maintenance tasks commonly require
isolation of specified sections of pipework and vessels, purging, and blinding at
specific points. A reliable method is need ed for task-specific safety assessment,
communication of safety critical info rmation, and task authorization. |
Conducting PHA Revalidation Meetings 147
Example 7 – Excessive Conservatism
Risk analysts tend to be conservative. Thus, revalidation teams have a tendency
to recommend additional safeguards, even when the risk is categorized as
tolerable. Given finite resources, the resolution of any recommendations to
reduce tolerable risks further will cons ume resources better spent on resolving
recommendations to reduce elevated risk s. Furthermore, a recommendation to
reduce one particular risk may be more th an offset by increases in other risk(s).
For example, adding a redundant pressure relief valve may marginally decrease
the risk of vessel rupture due to overpressure. However, the additional relief
valve might fail open under normal operating conditions, significantly increasing
the risk of a loss of containment when no overpressure exists. Thus, the study
leader should challenge recommendations to add more safeguards when risk is
already tolerable.
Example 8 – Excessive Optimism
Team members tend to be optimistic in their risk judgments, particularly when
human performance gaps are involved. In their experience, workers make very
few serious errors, and when they do , they detect and correct the error
themselves, or the engineered controls ac t to minimize any harm. If a particular
revalidation team is knowledgeable of LO PA, it may be difficult for them to
accept that a particular human error could happen more than once in ten years.
A facilitator should watc h for phrases like “no one would ever close the wrong
valve around here.” A review of past inci dents or a discussion of these types of
events with operators might reveal these, or similar, events do indeed happen
occasionally. The team might then consid er a more detailed supplemental risk
assessment using Human Reliability Anal ysis techniques where applicable.
This general tendency towards optimism also extends to equipment
performance. Many LOPA teams work hard to get a loss scenario to show a
tolerable risk (sometimes by taking cr edit for too many basic process control
system [BPCS] safeguards such as IPLs , double counting IPLs, or ignoring
common cause failures between IPLs). If such issues are noted, the leader should
challenge the team to follow their r evalidation charter and recommend risk
reduction measures whenever they dete rmine that elevated risks exist, as
estimated using company-approved methods.
|
176
Figure 8.4: Operating Ranges and Limits
to gases resulting in a pressure rise. The mandatory action temperature
limit is set low enough to allow time for the response to prevent loss of containment from high pressure. Anothe r example, an upset in feed to
a tank could lead to an overflow on high level. The control system design
provides enough time for the tank le vel control to sense the upset and
to take corrective action on the flow into or out of the tank before it overflows. For such a tank, the maxi mum setpoint for the level must be
reduced to allow adequate response time.
Basic Process Control Systems (BPCS) and Safety Instrumented Systems
(SIS) . There are few chemical plants that are so robust that an active
control system is not required. Using both active and passive controls
can assure product yield and qua lity and maintain safe operating
conditions. This type of control sy stem is known as a basic process
control system (BPCS). The BPCS acts to alarm and moderate a high or
low operating condition within the ne ver exceed limits. However, when
a high risk that is considered to be intolerable cannot be lowered with
existing control systems or other layers of protection, a SIS is provided to rapidly shutdown or otherwise automatically place the process in a
|
82
boards. The website of the Canadian Centre for Occupational Health and
Safety (CCOHS), Canada’s national center for occupational health and
safety information, provides guid ance for selection of alternate
chemicals. This includes hazard as sessments for alternative chemicals
which should include consideration of the following issues in order to
minimize risk (Ref 4.1 CCOHS):
•Vapor pressure
•Short-term health effects
•Long-term health effects
•Skin toxicity
•Sensitization of the respiratory system
•Cancer-causing potential and reproductive effects
•Physical hazards (e.g.,flammability)
The Organization for Economic Cooperation and Development
(OECD), an international organizati on dedicated to promoting policies
that will improve the economic and social well-being of people around the world, has also developed a Substitution and Alternatives
Assessment Toolbox (SAATOOLBOX), wh ich includes information and
resources on chemical substituti on and assessment practices and
practical guidance on how to cond uct assessments, including case
studies. Case studies are descriptions of alternative or chemical hazard assessments that have been conduc ted by manufacturers, academic
institutions, NGOs or government bodies, and include evaluation of
alternatives for: (Ref 4.24 OECD)
Flame retardants
Plasticizers (phthalate-free)
Solvents
Nonylphenol ethoxylates for detergents
The REACH (Registration, Evaluation, and Authorisation of
Chemicals) regulation in Europe, adopted in 2007, requires manufacturers and importers of cert ain chemicals above a one metric
ton threshold to gather information on the properties of their chemical
substances, and to register the inform ation in a central database. It also
calls for the progressive substitution of the most dangerous chemicals
(referred to as "substances of ve ry high concern") when suitable |
23. Working with contractors 305
Common mobilization activities include:
• Instructing contractors on safety procedures.
• Familiarization with client specific documentation.
• Communicating mandatory safety rules.
• Highlighting anything that is unique to the site or company, and that is
different to practices in the rest of the industry.
• Providing a site orientation and induction, including process hazards.
• Briefing contractors on simultaneous operations.
• Asking if they have performed thes e operations before and checking
what training and support they need.
• Emphasizing a “don’t hesitate to speak up” approach and reminding
contractors that they are working in an open challenge culture.
• Communicating key behavioral expectations, such as stopping when a
risk is found, or if a work instru ction cannot be completed properly.
• Sharing fatigue risk management requirements, such as adopting the
client’s fatigue risk manage ment policy and requirements.
Sufficient time should be allowed with in schedules to enable briefing of
contractors who are less familiar with the site and safety management procedures.
Supporting open communication
If contractors are reluctant to disagree, this may be mitigated by:
• Communicating that there will be no adverse repercussions to express
challenges or disagreement.
• Stating that contractors must report any unsafe condition or event, or
potential risk.
• Demonstrating reporting mechanism for incidents, accidents, near
misses and unsafe acts.
• Actively listening to contractors.
Operational readiness review
An explicit “readiness to commence wo rk” review can help confirm that the
contractors:
• Are equipped with work instructions.
• Are aware of nearby or simultaneous activities.
• Have certificated equipment, tags, and locks etc.
• Have clearly defined roles.
• Are aware of stop and hold points.
• Are aware of applicable emergency plans.
• Have a named person/role to report to. |
253
combinations of equipment failures and human performance issues that can result in an
incident. FTA is well suited for analyses of high ly redundant systems. For systems particularly
vulnerable to single failures that can lead to incidents, it is better to use a single-failure-
oriented technique such as FMEA or HAZOP Study. FTA is often employed in situations where
another hazard evaluation technique (e.g., HAZOP Study) has pinpointed an important incident
of interest that requires more detailed analysis.
The fault tree is a graphical model, as shown in Figure 12.6, that displays the various
combinations of equipment failures and human perf ormance issues that can result in the main
system failure of interest (called the Top Even t). This allows the hazard analyst to focus
preventive or mitigative measures on significant ba sic causes to reduce the likelihood of an
incident.
Fault Tree Analysis is a deductive technique that uses Boolean logic symbols (i.e., AND
gates shown as a flat arch, OR gates shown as an arrowhead in Figure 12.6) to break down the
causes of a top event into basic equipment fa ilures and human performance issues (called
basic events). The analyst begins with an incide nt or undesirable event that is to be avoided
and identifies the immediate causes of that even t. Each of the immediate causes (called fault
events) is further examined in the same manne r until the analyst has identified the basic
causes (shown as circles in Figure 12.6) of ea ch fault event or reaches the boundary established
for the analysis (shown as a diamond shape in Figure 12.6). The resulting fault tree model
displays the logical relationships between basic events and the selected top event.
Top events are specific hazardous situations th at are typically identified through the use
of a more broad-brush hazard evaluation techni que (e.g., What-If Analysis, HAZOP Study). A
fault tree model can be used to generate a list of the failure combinations (failure modes) that
can cause the top event. These failure modes are known as cut sets. An important qualitative
outcome of an FTA is the minimal cut set (MCS) is a smallest combination of component failures
which, if they all occur or exist simultaneously , will cause the top event to occur. For example,
a car will not operate if the cut set “no fuel” and “broken windshield” occurs. However, the MCS
is “no fuel” because it alone can cause the Top event; the broken windshield has no bearing on
the car’s ability to operate.
Fault tree analysis can be quantified. Where frequency data are available for the basic
events, the resultant frequency of the top event can be calculated using Boolean algebra or
arithmetical approximations. This data may be used in quantitative risk assessment.
Fault-Tree and Event Tree are suited to provide estimated of
scenario likelihood or frequency. HAZARD IDENTIFICATION |
OVERVIEW OF RISK BASED PROCESS SAFETY 49
Example Incident: Piper Alpha
The Piper Alpha incident in 1988 is a ri ch source of lessons learned (see prior
highlight box under the element of Hazard Identification and Risk Analysis). This
aspect of the incident relates to how the control room was disabled immediately
by the first explosion incident. This was where emergency management should
have occurred. Due to fires and smoke, personnel retreated to the accommodation
module to await further instructions and evacuation. The accommodation was not
smoke resistant and most of the fatalities occurred from smoke inhalation in that
location. A few personnel jumped into the sea and survived the fall and were
rescued. Neighboring facilities continued to pump hydrocarbons towards the
Piper Platform even after they were aware it was on fire as they did not have
permission from shore management to stop the flow.
Several important lessons were learned. The emergency management center
should be protected against potential incidents and a backup control center should
exist in a separate and distanced location to enable emergency management if the
first center is lost. There should be temporary safe refuges for personnel to escape
to prior to evacuation. These locations mu st be protected against fire and smoke
and allow transit to evacuation points – usually the lifeboats. Further, emergency
response should be delegated to operational personnel on adjacent facilities – a
type of stop work authority, without the need for shore-based management
approval.
RBPS Application
Emergency Management sets out a comprehensive approach to emergency
management which ensures that all aspects of emergency preparedness are
carefully assessed, and that emergency response is practiced so that everyone
knows what needs to happen.
3.2.4 Pillar: Learn from Experience
This pillar addresses how to learn from experience – from incidents and near misses,
from leading and lagging metrics, and from audit reports. The aim from all of these
is to identify deeper systemic causes and to implement corrective actions that solve
the wider issue, not just the specific in stance. The final element of management
review leads to continual improvement. This goes beyond prescriptive requirements.
Good process safety should seek improvements by learning from experience, going
beyond compliance which can degrade into a tick-the-box mentality.
RBPS Element 17: Incident Investigation
Incident investigation is the process of reporting, tracking, and investigating
incidents and near misses to identify root cau ses so that corrective actions are taken,
trends are identified, and learnings are co mmunicated to appropri ate stakeholders. |
Appendix 217
Process Safety Culture
Compliance with
Standards
Process Safety
Competency
Workforce Involvement
Stakeholder
Outreach
Process Knowledge
Management
Hazard Identification
and Risk Analysis
Operating
Procedures
Safe Work
Practices
Asset Integrity and
Reliability
Contractor
Management
Training and Perform.
Assurance
Management of Change
Operational Readiness
Conduct of Operations
Emergency
Management
Incident
Investigation
Measurement and
Metrics
Auditing
Management Review
and Contin. Improv.
5 3 % 4 7 % 12345 67 89 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0
5% 3% 6% 1% 2% 9% 15% 10% 3% 12% 1% 3% 8% 2% 3% 8% 4% 1% 1% 2%
Year 35 31 16 12 20 3 6 30 52 35 10 40 5 12 27 8 11 29 15 4 3 7
C5.6.1-1Dolphin Energy Ltd.
(Behie 2008) 2009 1 1 1 1
C5.6.2-1Husky Superior Refinery
(CSB 2018a)2018 1 111 1
C5.6.3.-1BP Texas City
(CSB 2007)2005 1 1111 11 11111111 1111
(C5.6.3)
(A.4-1)Bayer CropScience
(CSB 2011a)2008 1 1 1 1 1 1 1 1 1
(C5.6.3)
(A.4-1)Steam Generation
(Sanders 2015; p. 79)1991 1 1 1 1 1 1
(C5.6.3)
(A.4-1)Ethylene Plant Start-up
(Kletz 2009; pp. 408-411)Not
Known 1 1 1 1 1 1 1 1
(C5.6.3)
(A.4-1)Start-up Afterwards
(Kletz 2009; pp. 252)Not
Known 1 1
C6.5-1Monsanto 1969
(Fogler 2011)1969 1 1 1 1 1 1 1Chapter 5 - Table 1.1 Modes 5, 6
Extended Shutdowns
(Start-up) Not discussed in Chapter 5
Chapter 6
Recovery
Pillar IV
Learn from ExperienceIncident
Elements Identi fied as "weak" (See Figure 10.3)
No. of Identified RBPS Causes
Risk Based Process
Safety ElementTransient
Operating
Mode
Pillar I
Commit to Process Safety
Pillar II
Understand Haz. and Risks
Pillar III
Manage Risk
Table A.2-2 Summary of the inci dents during the transient operating mode (Continued) |
Appendix D – Reactive Chemicals Checklist
This checklist is adapted from a CCPS Safety Alert; A Checklist for Inhere ntly Safer Chemical
Reaction Process Design and Operation , March 1, 2004. For additional information on chemical
reactivity tools, see section 5.8.
D.1 Chemical Reaction Hazard Identification
1. Know the heat of reaction for the intended and other potential chemical reactions.
Several techniques are available for measuring or estimating heat of reaction, including
various calorimeters, plant heat and energy balances for processes already in operation,
analogy with similar chemistry (confirmed by a chemist who is familiar with the chemistry),
literature resources, supplier contacts, and thermodynamic estimation techniques. You
should identify all potential reactions that could occur in the reaction mixture and
understand the heat of reaction of these reactions.
2. Calculate the maximum adiabatic te mperature for the reaction mixture.
Use the measured or estimated heat of reaction, assume no heat removal, and that 100%
of the reactants actually react. Compare th is temperature to the boiling point of the
reaction mixture. If the maximum adiabatic reaction temperature exceeds the reaction
mixture boiling point, the reaction is capable of generating pressure in a closed vessel and
you will have to evaluate safeguards to preven t uncontrolled reaction and consider the
need for emergency pressure relief systems.
3. Determine the stability of all individual components of the reaction mixture at the
maximum adiabatic reaction temperature.
This might be done through literature search ing, supplier contacts, or experimentation.
Note that this does not ensure the stability of the reaction mixture because it does not
account for any reaction among components, or decomposition promoted by
combinations of components. It will tell you if any of the individual components of the
reaction mixture can decompose at temperatures which are theoretically attainable. If any
components can decompose at the maximum adiabatic reaction temperature, you will
have to understand the nature of this decomposition and evaluate the need for
safeguards including emergency pressure relief systems.
4. Understand the stability of the reaction mixture at the maximum adiabatic reaction
temperature.
Are there any chemical reactions, other than the intended reaction, which can occur at the
maximum adiabatic reaction temperature? Co nsider possible decomposition reactions,
particularly those which generate gaseous products. These are a particular concern
because a small mass of reacting condensed liquid can generate a very large volume of
gas from the reaction products, resulting in ra pid pressure generation in a closed vessel.
Again, if this is possible, you will have to understand how these reactions will impact the
need for safeguards, including emergency pr essure relief systems. Understanding the
stability of a mixture of components may require laboratory testing. |
FIRE AND EXPLOSION HAZARDS 73
any vents that may interact with them, is key to maintaining the vapor space outside of the
flammable range.
This leaves the heat leg of the triangle which is addressed through ignition source control.
Unfortunately, as Trevor Kletz, a founder of pr ocess safety, said, “Ign ition sources are free”
meaning that ignition sources are prevalent. Potential ignition sources and means to control
them are provided in Table 4.5. These ignition sources and control methods apply to
flammable materials and combustible dusts. Additi onally, for dusts, control of the confinement
and dispersion legs of the dust pentagon through diligent housekeeping is a prevention
measure.
Table 4.5. Ignition sources and control methods
Ignition sources Control Method
Electrical Electrical Area Cla ssification (refer to following
text)
Bonding and grounding
Smoking Prohibition of use or allowance only in specified
areas
Overheated materials Maintain integrity of equipment including that
handling/conveying solids
Hot surfaces Elimination of surfaces above autoignition
temperatures of materials being handled
Burner flames Facility siting such that heaters are located
apart from equipment handling flammable
materials
Sparks Control of hot work through a work permit
system
Spontaneous ignition Control of pyrophoric materials such as ferrous
sulfide scale
Cutting and welding Control of ho t work through a work permit
system
Static electricity Bonding and grounding
Chemical Reaction Understand ing and control of chemical
processes
Lightning Provision of lightning protection
Vehicles Control of vehicular access in areas handling
flammable materials |
28
approaches. First and second order inherently safer approaches are
described in more detail below.
In the strictest sense, or the First Order of Inherent Safety, one could
argue that the definition of inherently safer applies only to the complete
elimination of a hazard. Elimination or complete avoidance of the hazard is a priority of a First Order solution and certainly fits Trevor Kletz’s original dictum that “What you don’t have can’t leak.” Inherently safer
strategies that absolutely eliminat e a hazard are an optimum solution,
while hopefully not introducing another hazard of concern as a result, nor transferring the risk to another part of the value chain for the product or material of concern. Exa mples of first order IS measures
would be shut down or removal of the process presenting the basic
hazard(s) or substituting a hazardous material with one that is totally non-hazardous from a process safety viewpoint.
Alternatively, inherently safer approaches can also address the
hazard by making it less intense or by virtually eliminating it. These
approaches can be labeled as Second Order of Inherent Safety. Such
approaches are clearly in line with th e philosophy of inherent safety but
may not be as powerful as a First Or der change. In the Second Order of
inherent safety, the hazard is only reduced through the application of IS
principles, e.g.,minimization or substitution. It could be that Second Order inherently safer design or operational options result in an acceptable reduction of hazard and, therefore, the risk is adequately
addressed. Examples of second order IS measures would be substituting a hazardous material with one that is still hazardous from a process
safety viewpoint but less hazardous (e.g ., less volatile, or less toxic), or
reducing the inventory of a hazardous material but not completely eliminating its use. This is what is meant by “virtually” eliminated.
In the broadest sense, the overall hazard is not completely
eliminated or reduced by way of Second Order inherently safer
strategies but instead, sublevel hazards are minimized and the likelihood of the event occurring is reduced by adding layers of protection. The
strength and reliability of a layer of protection can vary, with some layers designed to be more “robust” than others. For exampl e, “independent”
protection layers provide more strength and reliability than protection layers that are not independent of each other. This could mean it is more |
8 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION
Figure 1.3. Number of fatal work in juries, by industry sector, 2019
(BLS 2019)
|
186 Human Factors Handbook
Table 16-3: Task planning tact ics for different task errors
Error condition Task planning error management tactic
Unclear, incomplete, or
ambiguous procedures Comprehensive and clear instructions. See Chapters 5
to 7.
Lack of agreement on best way
to perform a task Team review to secure consensus
Inadequate competence Assignment of more experienced people.
Unfamiliar task Comprehensive and clear instructions, task briefing,
and condition verification.
Realistic schedule and task checking.
Task interruptions or
distractions Task design (shielding people from distractions, e.g.,
temporary restriction of acce ss to a work area) and job
aids (see 0).
Determine the level of use of procedures for a task,
such for 1) reference, 2) continuous use in hand or 3)
monitored.
Stipulate “Procedure place keeping” such as checking
off key steps or Hold points to check task completion.
Long or complex task Task and team design.
Checklists and/or other job aids, Hold or Stop Points,
and checkpoints. See Chapters 5 to 7.
Independent task checking. See Chapter 20.
Task focus Dynamic and situation aw areness aids, decision
reviews.
See Chapter 20
Over commitment of team to
complete tasks despite
unforeseen problems Dynamic and situation aw areness aids, decision
reviews.
See Chapter 20
Time pressures Realistic schedule and task checking.
See 17.3 and Chapter 16.
Many team members or
multiple teams – leading to
miscommunication and role
confusion Formal communications an d logs. See Chapter 19
Clear roles and responsibilities.
Unclear, unreliable, or
incomplete process
information List all the information needs and verify their
availability.
Poor physical environment e.g.,
issues with lighting,
temperature, workspace,
humidity, noise Provide conducive environment e.g., ensure
appropriate task lighting, PPE, and sound barriers.
Specify special means of communicating in a noisy
environment, such as hand signals or written
communication.
Allow additional time for task completion.
Equipment not fit for purpose,
such as hand tools Source appropriate equipment.
|
242 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION
The condensate level rose in the absorber to a point where it mixed with the exiting rich
stripping oil stream. Condensate mixed with rich oil flashed over the rich oil level control valve
resulting in a much-reduced temperature in the downstream Rich Oil Flash Tank. This caused
temperatures to drop across the plant as rich oil flowed through the recovery process where
hydrocarbons were stripped from the rich oil befo re returning it to the absorbers as lean oil.
Eventually, the lean oil pumps tripped out, caus ing major thermal excursions on a plant with a
high degree of process and thermal integration. Loss of lean oil was a critical event but was
not communicated to the supervisor until he re turned from the morning production meeting
one hour after the pumps had tripped.
Temperatures in parts of the plant fell to - 48°C (-54°F). At 08:30 AM, a condensate leak
occurred on heat exchanger GP922. The absence of lean oil flow meant that the condensate
flowing through the rich oil system was not warm ed as it entered the recovery section. The
reason for the leak was probably due to a strong thermal gradient created while attempts were
being made to re-establish the process. Other parts of the process showed signs of intense
cold with ice forming on uninsulated pa rts of heat exchangers and pipework.
At 10:50 AM, the leak from GP922 was getting worse, and the Supervisor decided to shut
down Gas Plant No: 1. By 12:15 PM, two mainte nance technicians had completed retightening
of the bolts on GP922 without making any apprec iable difference to the leak. It was decided
that the only way to stop the leak was to slow ly warm GP922 by starting a flow of warm lean
oil through it. However, initial attempts to re start the lean oil pumps were unsuccessful. Ten
minutes later, after operating a hand switch to minimize flow through another heat exchanger,
GP905, that heat exchanger ruptured, releasing a cloud of gas and oil.
It is estimated that the cloud traveled 170 m (558 ft) before reaching fired heaters where
ignition occurred. After flashing back to the po int of release flames impinged on piping, which
started to fail within minutes. A large fireball was created when a major pressure vessel failed
one hour after the fire had started. It took more than two days to isolate all hydrocarbon
streams and finally extinguish the fire (CCPS 2008 a).
The investigation concluded that the immediate cause of the incident was loss of lean oil
flow leading to a major reduction in temperatur e of GP905, resulting in embrittlement of the
steel shell. This was followed by introduction of hot lean oil in an attempt to stop the
hydrocarbon leak in GP922 which led to excess thermal stress in the end plate which failed
catastrophically due to embrittlement. Throug hout the whole sequence of events, operators
and supervisors had not understood the conseque nces of their actions to re-establish the
plant. Esso and the Government were desperate no t to shut down the plant, as it supplied all
the gas to the State of Victoria. They found thei r drawings were out of date and they needed
to walk the lines to discover what to isolate. In the end they had to shut down the plant and
that left the state without gas for between nine and nineteen days, causing major industrial
disruption and job losses.
Lessons
Hazard Identification and Risk Analysis. Gas plant #1 had not been subject to a hazard
identification study as had been done for the other two gas plants at the site. A Hazard and
Operability study, HAZOP, had been planned in 1995, but never carried out. Flow and
temperature deviations, like those that occurred at Longford Plant No. 1, are typically |
98 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS
Human factors are considered a major part of alarm management and
design of the HMI. Example Incident 4.4 comprises an event where the
operators (pilots) were unable to evaluate what has happening under
conditions of high stress (French BEA Final Report).
Example Incident 4.4 – Air Fr ance AF 447 Crash, June 2009
On June 1, 2009, Air France flight AF 447 (Airbus A330-203) flying from Rio
De Janeiro to Paris crashed into the Atlantic Ocean appr oximately 3 hours
and 45 minutes after takeoff. The accident resulted in the fatality of 228
passengers and crew members. Th e French Bureau of Enquiry and
Analysis for Civil Aviation Safety (BEA) investigated the accident and
released the final report three years after the fatal crash.
The report identified the blockage of pitot tubes responsible for speed
measurement as the first of a series of events that led to the accident.
On an aircraft, three sets of pitot tu bes are used to determine key flight
parameters including speed and altitude. Blockage of the pitot tubes
caused inconsistencies in aircraft speed measurement that resulted in
disengagement of the autopilot leading the airplane to a stall position.
The crew failed to recover from the stall position.
According to the investigation report, “The blockage of Pitot probes by ice
crystals in cruise was a phenomenon that was known but misunderstood
by the aviation community at the time of the accident. After initial
reactions that depend upon basic airmanship, it was expected that it
would be rapidly diagnosed by pilots and managed where necessary by
precautionary measures on the pitch attitude and the thrust, as indicated
in the associated procedure. The crew, progressively becoming de-
structured, likely never understood that it was faced with a “simple” loss
of three sources of airspeed information.”
|
Chapter No.: 1 Title Name: Toghraei c13.indd
Comp. by: ISAKIAMMAL Date: 25 Feb 2019 Time: 12:28:12 PM Stage: Proof WorkFlow: <WORKFLOW> Page Number: 241
241
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
13.1 What Is Process Control?
To explain a control system, we can take an example
from everyday life.
A father asks his daughter to avoid the edge of a cliff
while they are hiking. This first step is a “regulatory” measure. A while later, the father sees his daughter playing at the edge of the cliff; this raises an alarm and the father tells her more harshly to be careful and step back. If the daughter continues to play by the edge, then the father may decide to take control of the situation; he may take her by the hand so that they can both leave the park. These three steps relate closely to a typical control system.
Similar to the above analogy, the first step in a process
control system is “regulation. ” Regulatory control is exercised to ensure that the process runs smoothly and according to specifications. The more technical term for this is a “basic process control system” (BPCS).
If the process runs out of control or off‐spec, then an
alarm is raised. This is the second step in the control system, which alerts the operator of the need to imple-ment stronger action to correct the situation.
If the out‐of‐control condition persists, then the control
system moves to the third step, which may involve drastic action. In plant process control, this is called a safety instrumented system (SIS).
So when we talk about a process control system, we
generally refer to three elements: BPCS + alarm + SIS. Collectively, these are called an integrated control and safety system (ICSS). We need to be specific when refer -
ring to any particular system because the general term “process control system” may mean a BPCS to some peo-ple and an ICSS to others.
Manipulating a plant is not limited to ICSS. ICSS is
basically “automatic control, ” but we still need operators’ presence in plants. What operators do is basically “manual control. ”
They should be in the field and in the control room to
monitor parameters to take actions when they see an emergency case.Over the life of a plant, there will be what we can think
of as “sunny days” and “rainy days. ” On sunny days, the process runs smoothly, there is no threat to the system, and everyone is happy. On sunny days a plant is run by a BPCS system. Rainy days are when there is one or more upsets in the plant. On rainy days, a plant is run by SIS actions. SIS actions cannot be manual, but operator intervention is provided for manual interference in automatic SIS actions (Table 13.1).
13.2 Components of Process Control
A
gainst Violating Parameters
As was mentioned in Chapter 5 there are four steering/protecting components for each specific process param-eter in equipment, units, and plant.
These steering/protecting components are:
●BPCS. The main function of the Basic Process Control System is to ensure that the plant runs smoothly and within specifications. This is achieved by using control loops to “measure” certain process parameters, “com-pare” them to specified set points (SPs) and then “adjust” the process accordingly.
●An alarm is incorporated into the system to prompt the operator to take action when the process runs out of control. Some people may ask why you should bother to install an alarm when there is a backup system, called an interlock, to deal with this situation. If they are properly trained, the operator will be able to make the wisest decision to override the control system and take corrective action. This will prevent the loss of production that would occur if drastic action were activated via the interlock system.
●SIS. If a process parameter goes out of control, an alarm is activated, which allows the plant operator to override the system to bring the process back under control. If he is unable to achieve this, the next layer of control comes into play. This is the safety instrumented system, whose main aim is to protect equipment. 13
Fundamentals of Instrumentation and Control |
30 INVESTIGATING PROCESS SAFETY INCIDENTS
The disadvantage of unstructured gr oup brainstormin g is that the
discussion may be dominated by individuals who are not shy about stating
an opinion and who may or may not be experts on the subject. Each person
may also enter the discussion with a bi as that can lead the thinking toward
incorrect conclusions. The results of group brainstorming are very
dependent on the collective experiences of the group, which may be
incomplete if the group is lacking in critical knowledge or a competency skill
set. Two different groups may reach two different conclusion s as to the cause
of an incident. Additionally , unstructured approaches are frequently
inadequate for investigatin g process safety incidents because they produce
incomplete and inconsistent results, and often do not determine all the root
causes.
While brainstorming has weaknesses as an investigation tool by itself, it
has an important role in more structured investigation methodologies.
Brainstorming is useful to encourage all investigation team members to
express their ideas and op inions, particularly follo wing the guideline to
brainstorming that no idea is disallowed. This can be a productive exercise
to develop hypotheses based on evid ence and observations, which is an
inductive reasoning approach. It rema ins to determine whether hypotheses
are true or false through va rious analysis techniques.
3.1.3 W hat If Analysis
A slightly more structured brainstorm ing tool uses What-If Analysis (CCPS,
1992), which involves the team asking “What if?” questions that usually
concern equipment failures, human errors, or external occurrences. Some examples
are: W hat if the procedure was wrong? W hat if the steps were
performed out of order? The questions can be generic in nature or highly
specific to the process or activity where the incident occurred. Sometimes
these questions are prepared in advance by on e or two individuals, which
may also potentially bias the discussion.
3.1.4 5-W hys
The 5-Whys tool is another brainstorm ing tool used to add some structure
to group brainstorming. The tool utilizes a logic tree approach without
actually drawing the logic tree di agram. The group questions why
unplanned, unintended, or adverse occurrenc es occurred or conditions
existed. Typically, the grou p asks “why?” about five times in order to reach
root causes; hence the name. Judgment and experi ence are required to use
the 5-Whys tool effectively to reach management system failures. The level |
US EPA urges chemical industry, universities to embrace
“benign by design” production (August 27, 1993). Chemical Regulation
Reporter , 989-990.
Finzel, W.A. (1991). Use low-VOC coatings. Chemical Engineering
Progress , 87 (11), 50-53.
Flam, F. (9 September, 1994). US EPA campaigns for safer
chemicals. Science, 265, 1519.
Flam, F. (14 October, 1994). Laser chemistry: The light choice.
Science, 266, 215-217.
Forsberg, C.W., Moses, D.L., Lewis, E.B., Gibson, R., Pearson, R.,
Reich, W.J., et al. (1989). Proposed and Existing Passive and Inherent
Safety-Related Structures, Systems, and Components (Building Blocks)
for Advanced Light Water Reactors. Oak Ridge, TN: Oak Ridge National
Laboratory.
Forsberg, C.W. (1990). Passive and inherent safety technologies
for light-water nuclear reactors. Pres ented at the America Institute of
Chemical Engineers 1990 Summer Na tional Meeting, August 19-22,
1990, San Diego, CA, Session 43.
French, R. W., Williams, D.D. and Wixom, E.D. (1995). Inherent
safety, health and environmental (SHE ) reviews. In E. D. Wixom and R.
P. Benedetti (Eds.). Proceedings of the 29th Annual Loss Prevention
Symposium , July 31-August 2, 1995, Boston, MA (Paper 1c). New York:
American Institute of Chemical Engineers.
Friedlander, S.K. (1989). The implications of environmental
issues for engineering R&D and education. Chemical Engineering Progress , 85 (11), 22-28.
Gerritsen, H.G., and Van't Land, C.M. (1988). Intrinsic
continuous process safeguarding. IChemE Symposium Series No. 110,
107-115.
Gibbs, W.W. (November, 1994). Ounce of prevention. Scientific
American, 103-105. 477 |
113
Runaway reactions in batch reactors can sometimes be avoided by
using separate reactor vessels for di fferent stages of the process. In
Figure 6.3, four reactants are added to a reactor to make a product. If
materials C or D are added to the first stage (when A and B are added),
or A or B are added to the second stage (when C and D are added), then a runaway reaction may occur. This situation can be rendered moot by adding materials A and B in one re actor, then piping the resulting
materials to a separate reactor ve ssel where materials C and D are
added. A runaway reaction is not possi ble in the two-reactor design (Ref
6.9 Kletz 1998).
6.5 SIMPLIFYING HEAT TRANSFER
Where possible, cooling systems should be designed so that they can
provide adequate heat removal via natural convective cooling. This
requires a thorough understanding of the hydraulic conditions created
by elevation and temperature differences, and what types of mass and
heat transfer mechanisms are established by these differences. Naval
and commercial reactors utilize elev ation and temperature differences
to drive emergency cooling systems, which will keep a shutdown reactor from overheating due to decay heat. The same principles can also be
applied in chemical/process system designs. For example, cooling fins can be added to vessels, and “fin-fan ” type air coolers can be used which
use natural air currents versus forced-fan draft for airflow.
|
8 PROCESS SAFETY IN UPSTREAM OIL & GAS
Figure 1-2. Scope of Process Safety in Upstream Oil and Gas
formalized for process safety than offshore. Large scale onshore upstream
operations are covered by PSM and RMP, but smaller developments are not covered
by these federal process safety regulatio ns (CSB, 2018) nor are drilling activities of
any size. The regulatory focus for smaller on shore operations is occupational safety
and environment with state and local regula tions predominating. Process safety is
driven by following relevant API standards. Upstream onshore operations have a
larger number of process safety incidents than offshore (as is shown in Section 1.6.3)
but these usually have fewer impacts to people, and this may be a factor in the degree
of regulation.
1.6.1 Analysis of US Offshore Safety Data
Halim et al (2018) analyzed Bureau of Safety and Envir onmental Enforcement
(BSEE) incident reports. Offshore operator s within the US Outer Continental Shelf
are required to report specific incidents to BSEE. Over the period 1995-2017, a total
of 1,617 incidents were investigated. Th e authors further analyzed 137 fire and
explosion incidents over the period 2004-2016 where there was sufficient detail to
establish causation. They identified nine of the most common causes, of which
equipment failure and human error dominate.
COS (Center for Offshore Safety) also pr ovides incident data reported by its
membership. Smolen (2019) summarizing this data shows while process safety
performance has improved over several years in some categories, it may be
plateauing. COS also collect s Tier 1 and Tier 2 process safety incident data.
1.6.2 International Incident Data from IOGP
Incident trend data is available from IOGP (International Association of Oil & Gas
Producers) covering both onshore and offshore incidents. IOGP is a consortium of
companies which operates in 80 countries and collectively produces about 40% of
|
Table 26-3 continued
Investigatory
tools Description of tool
Barrier
Analysis [31] One example of Barrier Analysis mode ls is a Bow Tie Diagram. The diagra m gives an overview of multiple
scenarios in one picture. The Bow Tie te chnique consist of the following steps:
1. Identify the hazard.
2. Define the top event – the exact moment at which control was lost.
3. Define the threats – the factors that caused the top event.
4. Define the consequences – the outcomes of the top event.
5. Identify the barriers.
Bow Tie analysis is also applied to human error.
(reproduced from [31])
|
Containers
165
top portion of the tank volume. The rest of the tank volume
is dedicated to fire water.
The third type of merge is a container with internal
compartments. One good example of this type of merge is in hot lime softener (HLS) in water treatment areas. HLS is an expensive piece of equipment that, in addition to other features, should be installed on legs as it has a sloped floor. This makes the installation of HLSs expensive. This has made companies think of using HLS internal space for other uses. For example a compartment could be fabri-cated inside of HLS and be used as a “backwash water tank” for filters downstream of the HLS (Figure 9.36).
9.20 Secondary Containment
Secondary containment is a physical enclosure around voluminous items to prevent wild liquid escape during an uncontrolled (i.e. accidental) release. The requirements for secondary containment could be technical and/or legal. A regulatory body in a specific jurisdiction may ask for secondary containment to be provided for all tanks and/or vessels and/or pipes in a plant.
Secondary containment is used as a safety measure
just in case there is a sudden rupture in the body of a large tank to prevent the release of a huge amount of liquid to the plant and the neighborhood. Some may think that secondary containment is only necessary when a tank or tanks contain “non‐innocent” (non‐harmful) liquids. It was the case in the old days that secondary containment was only used for non‐innocent liquids. However, these days secondary containment is applica-ble for all type of liquid content, even potable water. However, the local codes define whether you need to put secondary containment on a tank or not.The other issue is if secondary containment is not only
for tanks. Even though it is very common to see second-ary containment for tanks, secondary containment could be done for tanks, vessels, or even pipes. In some process plants where they produce lethal material they may have secondary containment on their tanks and vessels, and even on the pipes.
You may have seen secondary containment in process
plants around tanks or a group of tanks. They are gener-ally in the form of dykes or berms.
There are two main methods to provide secondary
containment for voluminous elements:
1)
Prov
iding a berm or dike around the element
2) Using a double w
all element.
In both the above methods we have primary and
se condary containment. The wall of the container or
pipe is named the primary containment and the second
barrier that we put around the container or pipe for safety purposes is the secondary containment.
These two methods are shown in Table 9.12.There is one critical characteristic for secondary
containment that, if it fails, the physical containment cannot be qualified as secondary containment, which is the space between the primary and secondary con-tainments. The important point is that there should be a space between the primary and secondary con-tainments in a form that the space can be monitored against the leakage. For example, for very common dykes around tanks the space between the dyke and the tank body is visible to the field operator. The field or rounding operator can always check if there is a leakage from the tank to the dyke area and warn the operating personnel before a large rupture and large release of liquid.
DWCD WC
CBC DBC
HLSDWC: Deaerated wa ter compartment
DBC: Dirty back wash wa ter compartment
CBC: Clean back wash wa ter compartmen t
Hot lime softenerFigure 9.36 Mer ging tanks into a
compartmented tank: HLS example. |
KEY RELEVANCE TO PROCESS PLANT OPERATIONS 33
experienced. Some of these situatio ns may have been anticipated, and
procedures or automated systems are in place to deal with them; others
may not be anticipated, and plant pe rsonnel will have to troubleshoot
and resolve the issues using their skills, knowledge, and experience.
Many of these abnormal process si tuations occur during startups,
unanticipated interruptions, or when conducting non-standard
operating tasks.
The design of many process pl ants incorporates complex and
sophisticated process control system s to keep the plant running in
optimum condition and to protect the safety of the plant and personnel.
In addition, the plants should al so follow Recognized And Generally
Accepted Good Engineering Practi ces (RAGAGEP). Nevertheless, the
plants can be subject to periodic fa ilures or upset situations that are
sometimes not easily recognized or controlled. These abnormal
situations may not be immediat ely obvious and are sometimes
undetected or overlooked; but if a situation should develop into an
unwanted event, the consequences co uld be significant. Therefore,
recognizing and managing these situations is vital to maintaining a stable
and safe process.
Situation Awareness, which is some times referred to as Situational
Awar eness (SA), is an ar ea of human factor s that is highly relevant to
abnormal situations. The topic enco mpasses how humans interact with
complex systems, including percepti on of a situation, understanding
what is going on and predicting how the status will change in the near
future. One definition for SA can be found in Stanton (Stanton et al 2001)
and is as follows:
“Situational awareness is the cons cious dynamic reflection on the
situation by an individual. It provides dynamic orientation to the
situation, the opportunity to reflect not only the past, present and
future, but the potential features of the situation. The dynamic
reflection contains logical-conceptual, imaginative, conscious and
unconscious components which enables individuals to develop
mental models of external events.” (Bedny & Meister, 1999)
|
Piping and Instrumentation Diagram Development
256
Measuring process parameters is uniquely done by
“process analyzers, ” or “ AEs. ” They are too complicated
to be built in the form of AG or gauge.
However there are some gauges than can be connected
to indicators too.
Table 13.14 shows a list of sensors.Some companies use acronym for gauges instead of
elements interchangeably.
The portable gauges need specific locations on the
plant to be able to “cling” on to them to measure the parameters. These “sensor locations” should be tagged on the P&IDs. A list of them is given in Table 13.15.
It can be seen from the table that gauge points are
generally limited to pressure points and temperature points, or thermowells.
Some companies, instead of the acronym “PP” , use
“PT” , which means “pressure tap. ”
Now we are going to give a brief review of primary
elements or sensors. It is important to stress that this is only a “summary” for the purpose of P&ID development. There is a huge amount of knowledge regarding sensors and the selection of them for different applications is beyond the scope of this book.
In this section we talk about each of five sensors, tem-
perature sensors, pressure sensors, level sensors, flow meters, and process analyzers. For each of them we briefly talk about each group and the common types of sensors. Then the different methods of connecting sen-sors to process items are shown. These methods are known as “hook up” arrangements. The hook up arrange-ments are not always shown on P&IDs, but it doesn’t mean they don’t exist. Some companies prefer to not showing them just to avoid crowdedness on P&IDs.
The P&ID symbol for each of the sensors is introduced
here too. However, sometimes, especially in the early stages of P&ID development, the exact type of a sensor is not yet decided. In such cases just the tag of the sensor may be used as its symbol. Obviously this should be replaced later with the exact symbol of the selected sensor.
13.11.1.1 Temper ature Measurement
Temperature measurement can be done everywhere on flow of gas, liquid, or flowable solids.
It also can be done for parts of equipment. One exam-
ple is “skin temperature, ” which is the temperature of the coil wall inside fired heaters. The other example is measuring the temperature of winding in high power electric motors.
A unique feature of temperature measurement is that
they are types of sensors that work remotely. For exam-ple the temperature a flue gas stream can be measured remotely.
Temperature sensors generally don’t have any specific
symbol on P&IDs.
Table 13.16 is a non‐exhaustive list of common tem-
perature sensors.
Table 13.14 Sensors .
Parameter T P L F A
Tag TE PE LE FE AE
Table 13.15 Gauge poin
ts.
Parameter T P L F A
Tag TW PP Not generally needed Not generally needed Not available
Table 13.16 Temper
ature sensors.
Type Unique advantage Unique disadvantage Application
Thermocouple Default choice for wide
temperature rangeSensitive to noise Specially in burner management system (refer to Chapter 16)
Resistance temperature detectors(RTD)Very accurate Needs power source Process pipeline temperature measurement
Thermistor High accuracy
●The narrow range
●The most inexpensive temperature sensorMore in laboratory but not in industrial process control.
Infrared Non‐contact type Needs line of sight Common for remote temperature sensing in combustion systems |
FIRE AND EXPLOSION HAZARDS 79
BP Isomerization Unit Explosion, Texas City, Texas, U.S., 2005
Buncefield Storage Tank Overfl ow and Explosion, U.K., 2005
Olive Oil Storage Tank Explosion, Italy, 2006
BLSR Deflagration and Fire, Texas, U.S., 2007
Valero-McKee LPG Refinery Fire, Texas, U.S., 2007
Imperial Sugar Dust Explosion, Georgia, U.S., 2008
University of California at Los Angeles La boratory Explosion, California, U.S., 2008
Varanus Island Pipeline, Australia, 2008
CITGO Refinery Fire, Texas, U.S., 2009
ConAgra Foods, North Carolina, U.S., 2009
NDK Crystal Vessel Rupture, Illinois, U.S., 2009
Petroleum Oil Lubricants Explosion, Jaipur India, 2009
Big Branch Mine Explosion, West Virginia, U.S., 2010
Kleen Energy Explosion, Connecticut, U.S., 2010
Pike River Coal Mine Explosion, South Island, New Zealand, 2010
Texas Tech University Laboratory Explosion, Texas, U.S., 2010
Hoeganaes Dust Fires, Tennessee, U.S., 2011
Shell Refinery Fire, Singapore, 2011
Chevron Richmond Refinery Fire, California, U.S., 2012
West Fertilizer Company Explosion, Texas, U.S., 2013
Williams Olefins Heat Exchanger Rupture, Louisiana, U.S., 2013
University of Hawaii Laboratory Explosion, Hawaii, U.S., 2016
Port of Beirut Ammonium Nitr ate Explosion, Lebanon, 2020
Exercises
List 3 RBPS elements evident in the Imperial Sugar Dust explosion incident summarized
at the beginning of this chapter. Describe th eir shortcomings as related to this accident.
Considering the Imperial Sugar Dust explosio n incident, what actions could have been
taken to reduce the risk of this incident?
Search for a Safety Data Sheet from a supplie r of t-Butyl Amine. What is the Flash Point
(FP), Upper and Lower Flammable Limits (UFL and LFL), and the Autoignition
Temperature (AIT)? What is your reference?
Find the same flammability properties from a different source. What is the Flash Point,
Upper and Lower Flammable Limits, and the Au toignition Temperature? Does this agree
with the SDS you referenced?
What are the upper and lower flammability limit s for diesel fuel, gasoline, and propane?
What are their flash points?
What type of fire is possible from an atmospheric pressure storage tank storing a
flammable liquid? And from an industrial process unit handing flammable materials
under high temperature and pressure?
What is the difference between a deflagration and a detonation? |
26. Learning from error and human performance 339
of the accident, and recommendations had been made on the application of better
cementing practices [108]. For example, just one year earlier the Montara Oil Spill
(2009) accident had occurred. The inquiry re port for this accident [109] noted that
a direct cause of the accident was the defective installation of a cemented shoe casing, intended to operate as a primary barrier against blowout.
The root causes of the accident noted in the investigation report [110] were
cited as “organizational and safety management failures”, including:
• Lack of an adequate risk assessmen t/hazard procedure, and inadequate
details within the procedure.
• Inefficient recognition or timely responses to early warning signals.
• Poor communication.
• Lack of leadership, and an absence of a culture of leadership responsibility.
• Lack of learning from the lessons of previous incidents and recent near misses.
• Lack of appropriate emergency training to personnel.
Among other recommendations, the report focused on learning, and
highlighted the following “learning” recommendation:
26.3.3 A Human Factors perspective
From a Human Factors perspective (and learning focus) it was evident that
reporting systems were weak, which impa ired lesson learning. Lessons learned
from a similar near miss (caused by a negative pressure test failure) which
occurred on December 23rd, 2009, in the North Sea [111], were not shared across
the wider organization soon enough.
The United Kingdom Health and Safety Executive was satisfied with the
corrective actions implemented by Shell and Transocean following the North Sea incident. The Executive also noted that the shortcomings that had led to the accidents had been addressed [112]. This suggests that the 2009 near miss lessons
may have been shared and applied in the North Sea. The fact that the 2010
Deepwater accident occurred suggests that this learning had not yet been shared
with the Gulf of Mexico site.
Lessons learned in the aftermath of the 2009 near miss and the subsequent
2010 accident, suggest that it is vital that systems to investigate accidents are
appropriately designed. Such investigation systems must be able to identify
The need for increased transparency, reporting of incidents and near
misses for the purposes of learning lessons. |
HUM AN FACTORS 269
important that the investigators rely on facts based on evidence in
developing the incident scenario.
Example:
“On arriving at the site of a major incident, an investigator was informed by a local manager that data from the control room were
useless as the instrument air to the pneumatic instruments had failed
during the ensuing fire. Ignoring this advice, the investigator studied the data and was able to exactly determine all process parameters at
the time of the incident, which ultimately confirmed a different
scenario from that being supported by local management.”
(Broadribb, 2012)
11.2.2 H uman Factors during the Causal Analysis
“Failure
to follow established procedure” is a common premature stopping
point for incident investigation related to human factors. In many cases, the investigation team identifies the fact that a person failed to follow
established procedures, then does no t attempt to investigate further and
determine the underlying reason for th e behavior. In most cases there is
an underlying correctable root cause th at should be iden tified and fixed.
The failure to follow established procedure behavior
on the part of the
employee is not a root cause, but instead is a symptom of an underlying
root cause and warrants fu rther root cause analysis. For example, if an
employee failed to fo llow an established correct procedure, the root cause
may involve training. However, if the employee failed to follow an
established incorrect procedure, this would be a symptom of a root cause
involving the development of procedures.
Chapter 10 addresses root cause analysis in detail.
The investigation team has an obligation to try to find the underlying
cause for the failure to follow established procedure behavior. Typical
symptoms and corresponding underlying system defects that can result in
an employee failing to follow procedure include:
• Out-of-date written procedure that no longer reflects current
practices or current configurat ion of the physical system, due to
defects in the process safety information, or operating procedures
management systems
• Employee perceives that his or he r way is better (safer or more
effective), due to deficiencies in the system for establishing and |
278 | Appendix D High Reliability Organizations
time” and where substandard perform ance is not
tolerated.
Compressed time factors whereby major activities may need
to take place in seconds.
While the nature of these operations differs from chemicals,
oil, and gas, several applicable lessons-learned about culture can
be gleaned. In the above-referenced literature study, the UK HSE
organized the characteristics and the lessons-learned in a figure.
Figure D.1 has been modified from the original to fit in this book.
Figure D.1 High Reliability Organization Map (After Ref D.2)
HROs exhibit common characteristics that enhance their
ability to deal with errors, including:
Containm ent of Unexpected Events
Deference to expertise
Redundancy
Oscillation between hierarchical
and flat/decentralized structures
Training and com petence
Procedures for unexpected events
Problem anticipation
Preoccupation with
failure
Reluctance to simplify
Sensitivity to operations
H RO s
Definition
Tight coupling
Catastrophic
consequencesInteractive
complexity
Learning Orientation
Continuous technical
training
Open com munication
Root Cause Analysis of
accidents/incidents
Procedures reviewed in
line with knowledge base
M indful Leadership
B ottom-up
comm unication of bad
news
Proactive audits
M anagement by exception
Safety-production balance
Engagement with front-
line staff
Just culture
Encouragement to report
without fear of blam e
Individual accountability
Ability to abandon work on
safety grounds
Open discussion of errors•
•
•
•
• • |
124 PROCESS SAFETY IN UPSTREAM OIL & GAS
Figure 6-4. Example fault tree logic
(from NASA, 2002)
Another form of HIRA analysis is Event Tree Analysis, also described in
Guidelines for Chemical Process Quantitative Risk Analysis (CCPS, 1999). Fault
trees build to the top event. The event tree takes this event through the many possible
outcomes depending on whether safety barriers are effective. An example of the
many possible outcomes is shown in Figure 5-2.
6.3.4 Asset Integrity and Reliability
Topics related to safety systems are addr essed in a number of RBPS elements. It is
convenient to address them under Asset Integrity and Reliability given the
importance of safety system reliability.
Safety Critical Systems and Equipment
The term ‘Safety Critical Elements’ is used in the UK to describe those controls put
in place to prevent or mitigat e significant process safety events. These may be full
barriers or individual barrier elements (see Section 2.7). These must be identified
|
326 Human Factors Handbook
Figure 25-2: Gathering and reviewing feedback
25.4.3 Operational debriefs
Operational debriefs provide rich inform ation on what was done well versus what
could have been done better. Operational de briefs can look at the execution of the
tasks and at the non-technical skills exhibited during tasks.
Operational debriefs can take place after doing a process start-up, and after
process upset or abnormal events, for example. They should reflect on:
• Individual and shared situation awareness.
• Effectiveness teamwork and task sharing.
• Effectiveness and efficiency of decision-making under pressure,
including decision-making under time pressure and under stress
conditions.
Collect feedback
Action
Review feedback
Improvement
Positive
feedback
Negative
feedback
Monitor effectiveness
of action |
Overview of the PHA Revalidation Process 17
focused on the scope of the individual changes and do not revalidate the PHA
as a whole. Thus, the PHA revalidation cycle is unchanged. The advantages of
this practice are (1) preparation for th e formal PHA revalidation can be vastly
simplified and (2) it is more likely that teams reviewing changes might spot a
concern related to another change because the cumulative history is
documented in the PHA. Continuously revi sing a working copy of the most recent
PHA can also shorten the PHA revalid ation team meetings because the
evergreen PHA should accurately document the risks associated with the current
process.
A common PHA revalidation cycle is five ye ars. This interval is specified in
the United States standards/regulations of the Occupational Safety and Health
Administration (OSHA) and the Environmen tal Protection Agency (EPA), and is
the typical frequency recommended by in dustry associations and a frequency
that has been used historically by ma ny companies. Facilities not covered by
government regulations may establish their own, appropriate revalidation
frequencies. For example, some companies choose to perform PHAs on all
processes, but they extend the revalidatio n cycle of voluntary PHAs to seven or
ten years for lower hazard or non- regulatory covered processes.
Additional guidance on establishing th e PHA revalidation schedule to meet
the revalidation cycle requirements is provided in Section 6.1.3.
Any facility may schedule its PHA re validations more frequently than
applicable regulations demand, either ro utinely or under special circumstances.
Companies may choose to shorten revalidat ion cycles for reasons that include: When Should the PHA Revalidation Meetings Start?
What determines the required date for beginning the PHA revalidation
meetings if the revalidation cycle is set at five years? Is it five years from (1)
the prior PHA first meeting date, (2) th e prior PHA last meeting date, (3) the
date the prior PHA report was issued , (4) the date management approved
the final PHA documentation, or (5) so me other date? PHA meetings can span
weeks or months, and the final PHA repo rt can be issued several weeks (if
not months) later.
The meetings must be started far enough ahead of the required completion
date to allow a high-quality product to be produced in compliance with all
requirements. |
Costa, R., Recasens, F. and Velo , E. (1995). Inherent thermal
safety of stirred-tank batch reactors: A prognosis tool based on pattern
recognition of hazardous states. In G. A. Melhem and H.G. Fisher (Eds.).
International Symposium on Runaway Reactions and Pressure Relief Design , August 2-4, 1995, Boston, MA (pp. 690-709). New York: American
Institute of Chemical Engineers.
Cottam, A. N. (1991). Risk assessment and control in
biotechnology. In IChemE Symposium Series, No. 124, 341-w348.
Crabtree, E.W., and El-Halwagi, M.M. (1994). Synthesis of
environmentally acceptable reaction s. In M. El-Halwagi, and D.P.
Petrides (Eds.).. Pollution Prevention Via Process and Product
Modifications (pp. 117-127). AIChE Symposium Series, 303. New York:
American Institute of Chemical Engineers.
Cusumano, J A. (August, 1992). New technology and the
environment. Chemtech, 482-89.
Dartt, C.B., and Davis, M.E. (19 94). Catalysis for environmentally
benign processing. Ind. Eng.Chem. Res. 33, 2887-299.
Davis, G.A., Kincaid, L., Menke, D., Griffith, B., Jones, S., Brown,
K., and Goergen, M. (1994). The Product Side of Pollution Prevention:
Evaluating the Potential for Safe Substitutes . Cincinnati, Ohio: Risk
Reduction Engineering Laborato ry, Office of Research and
Development, U. S. Environmental Protection Agency.
The design of inherently safer plants (1988). Chemical
Engineering Progress , 84 (9), 21.
DeSimone, J.M., Maury, E.E., Guan , Z., Combes, J.R., Menceloglu,
Y.Z., Clark, M.R., et al. (1994). Homogeneous and heterogeneous polymerizations in environmentally-r esponsible carbon dioxide. In
Preprints of Papers Presented at the 208th ACS National Meeting , August
21-25, 1994, Washington, DC (pp. 212-214). Center for Great Lakes
Studies, University of Wisconsin-Milwaukee, Milwaukee, WI: Division of Environmental Chemistry, Am erican Chemical Society.
DeVito, S. C. (November, 1996). Designing safer chemicals:
Toxicological considerations. Chemtech, 34-47. 474 |
356
An ISS analysis that is incorporat ed into the existing PHA review
process. This would require that the PHA for each process be re-
done in its entirety to include an initial ISS analysis. Revalidated
PHAs that examine only portio ns of the process may not be
sufficient to satisfy the initial IS S review if the whole process is
not evaluated. Checklists or guid eword analysis contained in the
guidance document that incorporates ISS can be used to accomplish this analysis.
Whichever type of ISS analysis is implemented by the facility the
CCHS guidance specifies that th e analysis be conducted and
documented in the following manner:
The facility will document the qualifications of the team facilitator/leader and team makeup , including positions, names,
and any relevant experience or training.
The facility will document the ISSs considered as well as those
implemented.
If the facility chooses to do an independent ISS analysis, the
facility should document the meth od used for the analysis, what
ISS were considered, and the result s of each consideration. If the
checklist for ISS was used, for it ems that were not considered,
document why those items were not considered, i.e., not applicable or were already consider ed in previous consideration.
The facility will document for the ISS considered and not
implemented, the grounds that were used to make the feasibility
determination. See CCHS’s defi nition of feasibility below.
The documentation for incorporatin g the guidewords for ISS into
a HAZOP should be consistent with the documentation used
during any HAZOP Study.
For any other ISS analysis, the fa cility should document the ISS
considered, the ISS implemented, and the ISS not implemented.
The ISS analyses should be revalid ated at least once every five
years. The revalidation should incorporate improvements made
in method since the last review was conducted or select a new
method to perform the ISS analys es; ISS review(s) for all changes
that have been made since the last ISS analysis; review of all major chemical accidents or releas es or potential major chemical |
Table C-1 continued
HF Competency Performance/ Knowledge
Criteria Level 1 - Operator Level 2 - Supervisor* Level 3 - Manager**
Non-technical skills
Decision Making
(continued) Understands the factors
which impair effective
decision-making Can identify factors which
affect effective decision-
making (such as tunnel
vision, confirmatory bias,
group think etc.) Can recognize when
decision-making
(cognitive processes) are
impaired and decision-
making bias are present Is able to assess the
effectiveness of a decision-
making process
Is able to make effective
decisions and avoid decision-
making bias Can describe
techniques/strategies to
avoid decision-making bias Can apply techniques to
prevent decision-making
bias Is able to apply techniques
to prevent/mitigate decision-
making bias
|
298 INVESTIGATING PROCESS SAFETY INCIDENTS
and regulatory agencies. Although it may be unreasonable to expect that
all the needs will be met completely, considering them during the writing
phase will help approach that goal. The large variation in the readers’
technical backgrounds, the need to in clude technical information and the
need to be reasonably concise may limit the usefulness of a single report,
although this challenge may be ad dressed by including an executive
summary or similar section in the re port for those with a less technical
background or less need/desire to know the details. Every re port represents
a balanced trade-off of content, details, quantity of in formation, to meet
the expected needs of the readers and user s. It is reasonable to expect that
the report user has some general knowledge of chemical process
technology and hazards. It is also reas onable to expect that the readers have
some genuine interest and a desire to gain from understanding and
applying the available lessons. The report should not only document and
communicate the findings and recommendat ions, it should also be a tool to
motivate or inspire action.
Carper, in his book Forensic Engineering (Carper, 1989), recognizes
multiple audiences. Carper ac knowledges the re ality that the report should
not be expected to reach all audiences equally and satisfy all questions.
Professional Accident Investigation by Kuhlman devel ops the concept that
different levels of management have di fferent needs and priorities (Kuhlman,
1977).
Although it is the most important single document, the investigation
report is only a portion of the overall record of the investigation. Other parts of the investigation record include photographs, measurements, process
data, witness accounts, laboratory anal yses, engineering an alyses, and other
facts and analyses that support determination of causal factors and root
causes. Consideration should be given to compiling and maintaining a
full
and complete set of documents for fu ture reference. This systematic
documentation package is sometimes referred to as the audit trail . It
provides subsequent reviewers and investigators with the opportunity to
understand the team’s dec isions and analysis more completely. The
document set should contain lists of relevant files. All documents associated with the investigation should be pr eserved according to company records
retention policy.
An investigation report:
• Describes the incident in full detail (with timelines if possible),
• Explains the sequence of events and failures that led to the |
70
Use of a reaction solvent with a high enough boiling point to
prevent vaporization in case of excess reaction
Innovative chemical synthesis proc edures have been proposed as
offering economical and enviro nmentally-friendly routes to
manufacturing a variety of chemical s. These novel chemical reactions
may also potentially offer increa sed inherent process safety by
eliminating hazardous materials or chemical intermediates, or by
allowing less severe operating conditions. Some examples of interesting
and potentially inherently safer chemistries include:
Electrochemical techniques, pr oposed for the synthesis of
naphthaquinone, anisaldehyde, and benzaldehyde (Ref 4.41
Walsh).
Extremozymes, or enzymes which can tolerate relatively harsh
conditions, suggested as catalyst s for complex organic synthesis
of fine chemicals and pharmaceuticals (Ref 4.8 Govardhan).
Domino reactions, in which a series of carefully planned
reactions occur in a single ve ssel, used to prepare complex
biologically active organic compounds (Ref 4.9 Hall; Ref 4.34 Tietze).
Laser light “micromanaged” reacti ons directed at the production
of desired products (Ref 4.7 Flam).
Supercritical processing, which allows the use in chemical
reactions of less hazardous solven ts like liquid carbon dioxide or
water. This benefit must be balanced against the high
temperatures and pressures required for handling supercritical fluids. Johnston (Ref 4.13 Johnston), DeSimone, et al. (Ref 4.5 DeSimone), and Savage (Ref 4.28 Savage) review some potential
applications of supe rcritical processing.
The use of glucose in lieu of benzene (a toxic and flammable hydrocarbon) for the production of adipic acid. It may be possible to produce glucose from biological residue materials,
such as plant husks and straw (Ref 4.16 Kletz 1998).
The substitution of toxic/flamma ble gases, such as phosphine,
diborane, and silane, in the manufacture of semiconductors with less hazardous liquids, such as trimethyl phosphite, trimethyl borate, and tetraethyl-o -silicate (Ref 4.16 Kletz 1998). |
RISK ASSESSMENT 327
Figure 14.11. Types of ALARP demonstration
(HSE a)
The Hong Kong criteria shows that risks above a certain level (the gray area) are
unacceptable. Below a certain level (bottom left ), the risks are acceptable. The risks in the
middle are in the ALARP region.
“ALARP” stands for “as low as reasonably prac ticable”. The ALARP concept is illustrated in
Figure 14.11. The intent is that risks in the ALARP region warrant further attention. They should
be mitigated to a level, beyond which, it is no t practicable to reduce the risk any further. The
“practicable” aspect includes consideration of time, effort, and money. This requires a
company to exercise judgment when making ALAR P decisions. No simple method is available
for determining if a risk is ALARP. Often cost-b enefit analyses are used to aid in decision
making. The Health and Safety Executive gives an example of ALARP as the following. (HSE b)
To spend £1m to prevent five staff suffering bruised knees is grossly
disproportionate; but
To spend £1m to prevent a major explos ion capable of causing 150 fatalities is
proportionate.
Layer of Protection Analysis (LOPA)
LOPA is a simplified form of risk assessment. The purpose of LOPA is to determine if the
scenario has sufficient layers of protection to ma ke the scenario risk tolerable. The concept of
layers of protection is illustrated in Figure 14.12.
LOPA evaluates single cause-consequence pairs, as compared to a QRA that calculates the
cumulative risk. LOPA typically follows a haza rd identification study which develops cause-
Risk Reduction
Regardless of Cost
Relevant Good
Practice
plus
Risk Reduction
Measures
plus
Gross Disproportion
Relevant
Good
PracticeIntolerable
Tolerable if ALARP
Broadly Acceptable |
161
Before studying alternative ty pes of equipment, the process
requirements must be understood. For example:
Is a solvent necessary?
Must the products or by-products be removed to complete the
reaction?
What mixing and/or time re quirements are necessary?
What sequencing is necessary for material additions?
Is the reaction exothermic, endothermic, or adiabatic?
These and other relevant questions must be answered before
alternate reaction schemes can be ev aluated. Similarly, different unit
operations are available to accomplis h the same processing objective.
For example,
Should a filter, a centrifuge, or a decanter be used to separate a
solid from a liquid?
Should crystallization or distillati on be used for a purification
step?
It is inherently safer to develop processes with wide safe operating
limits that are less sensitive to variat ions in the operating parameters, as
shown in Figure 8.3. Sometimes this type of process is referred to as a
“forgiving” or “robust” process. If a process must be controlled within a
very small temperature band in or der to avoid hazardous conditions,
that process would have narrow safe operating limits. For some reactions, using an excess of one reactant can enlarge the safe operating limits.
8.4.2 Unit Operations - Specific Some examples and considerations fo r specific common unit operations
are described as follows. Reaction . Reactor design is particularly critical because reactors involve
chemical transformations, and ofte n potentially significant energy
releases. Evaluation of the safety characteristics for a given reactor
design requires an understanding of what physical or chemical
processes control the rate of reacti on (catalysis, mass transfer, heat
transfer, etc.), as well as the total potential energy consumption or generation involved in the reaction. Energy generating pressure and/or
undesired side reactions should also be evaluated. This information is |
162 | 12 REAL Model Scenario: Overfilling
Barrier at the Hoek van Holland to effectively deal with flooding. However, sea
level has risen 20 cm over the past century. Alexandre said, “Climate change
appears to be a big concern for us. While we aren’t going to be hit by a
hurricane like Harvey, our rainstorms seem to be intensifying, which could
logically lead to more severe flooding. I recommend that we review our
inspection policy for tank foundations. We should also review our process
hazard analyses to determine what our procedures should be in the event of
an unprecedented flood. My question to you is, how do you define
unprecedented?”
Reed spoke up and said, “You don’t know what you don’t know, until it
happens. That’s what I learned from the incident that occurred in Crosby, TX.
Who would have ever predicted that much rain in such a short period of time?”
Pamela and Frederik smiled, perhaps a little bit nervously, at the comment.
Pamela said, “Good point. We have to start somewhere, and ultimately, it will
be Jan’s decision on how far we go to protect the site against flood.” Frederik
offered, “Let’s review the PHAs before we decide to do anything. We should
also consider the frequency of these reviews.” Alexandre quickly tapped the
keys of his computer, making note of these action items.
It was Reed’s turn to talk about the current issue at hand, the tank
overflow issues. “After much investigating,” he said, “the minor tank overflows
were caused by the severe weather impacting the float-and-tape gauges, just
as I suspected. We’re fortunate that we have a good crew with many years of
experience who know what to look for after a storm, but those guys are
eventually going to retire. You can try to capture their knowledge, but
sometimes, what one of us thinks is common sense isn’t the same for others.”
Frederik responded, “We do have a great team, but you’re right, the future
will be tough if we don’t find and train the right replacements. Do you have
any suggestions on how to handle this?” “I’m glad you asked,” Reed said. “My
good buddy Alexandre and I have been doing some research on alternatives
to the float-and-tape gauges. We figured, we can’t solve the workforce issue,
but we can help you with gauges that aren’t as susceptible to breaking down
in extreme weather.”
Pamela said “You’ve hit the nail on the head. Our plant relies on accurate
level measurement as a key risk-reduction measure. But as you’ve mentioned,
with the increased storm intensity, the current level indicators are increasingly
losing their reliability. We need to find a way to maintain the same level of risk
reduction.” |
16 | 1 Introduction
culture will then lead to a robust PSMS that in turn drives
improved and sustained process safety perform ance.
Since process safety follows the PDCA approach used in other
operational and business system s, improving process safety
culture will also likely lead to improvements in other cultures,
such as EHS, Quality, and technology, and therefore lead to
stronger business perform ance (see section 1.5 and Appendix A).
Likewise, process safety culture of an organization does not
exist in a vacuum. Instead, it inextricably links to the organization’s
overall culture, including other subcultures such as business
practices, overall EHS, quality, and even the culture of
stakeholders that interface with the organization (e.g., neighbors,
customers, etc.), and others.
In the ideal case, a strong positive process safety culture m ates
with other strong positive cultures to build an overall strong
positive corporate culture, as discussed by Musante (Ref 1.19).
This kind of strong and integrated culture is sometimes referred
to as Operational Excellence. The Musante reference, titled Doing
Well by Doing Good: Sustainable Financial Performance Through
Global Culture Leadership and Operational Excellence, is reproduced
with permission as Appendix A.
A weak or negative process safety culture may be coupled
with, for example, a strong business culture. This may provide
financial success and avoid process incidents for some tim e, but
ultimately a major process safety incident can cause it to fail
catastrophically. However, strong business culture can be
leveraged to build a strong process safety culture. Likewise, when
both process safety and business cultures are relatively weak, first
strengthening the process safety culture can be a stepping stone
to building an overall positive business culture. |
General Rules in Drawing of P&IDs
43
This helps keep all involved parties understand the
P&IDs without looking elsewhere.
3) The P&I
D set by manufacturing companies may have
the brand name of items used on each P&ID sheet.
4) The P&I
D set by manufacturing companies tends to
have less design and operational notes. They may have less notes because of a closer relation with their design groups within their company. Also because the design groups work in one specific area, they are experts and already consider the details of design requirements, which are in the Notes block of other P&IDs.
5)
P&IDs
created by manufacturing companies may
have more detail regarding instrument air.
4.7 Dealing with V endor or
Licensor P&IDs
The engineering company responsible for designing a process plant and developing the P&ID most likely does not have any item it built. The engineering company buys pumps from one vendor, vessels from another ven-dor, and tanks from a third vendor. Therefore, all the items on the P&IDs are supplied by the vendor. There is generally no “footprint” on P&IDs pinpointing the exist -
ence of vendors except in one important case. If there are several pieces of equipment by one vendor that are already assembled on skids or should be assembled in the field by the vendor, it should be shown on the P&ID (Figure 4.33).
Figure 4.33 Vendor bor derline.
Figure 4.34 Vendor‐supplied loose it ems. |
14. Inherent Safety Regulatory Initiatives
14.1 INHERENT SAFETY REGULA TORY DEVELOPMENTS AND ISSUES
In the past decade regulators and legislators in juri sdictions across the
globe have recognized the risk redu ction potential in IS and have
debated as to whether encouragin g or mandating these approaches
through regulation could improve ov erall process safety or security
results. Debated options have rang ed from requiring that facilities
“ con si d er ” IS as on e of sev er al ch oi c es, to a ctu al m an d ates th at IS b e
implemented, the latter option back ed up by giving agencies the
authority to override facility determ inations and require installation of
specific IS elements.
Several key obstacles or misperceptions have resulted in IS being
underutilized by industry, including the following:
A perception that IS is technically and economically practical for
only new processes, though it has been demonstrated to be
potentially useful for existing facilities.
The lack of an inherent safety infrastructure , or a framework for
evaluating IS systems, includ ing technologies supporting IS
approaches, and methodologies that permit IS to be integrated
into technical, economic, safety and security design
considerations. This includes th e lack of consensus metrics for
IS.
The lack of specific guidance on how to conduct an inherent safety
study, particularly for existing facilities and processes.
A lack of understanding of IS principles and how to practically
apply them to both new and existing facilities.
Therein lies the dilemma. Policy makers may view implementing IS
as a relatively simple and effective wa y of minimizing or eliminating the
hazards or consequences from proc ess incidents, but in practice
companies encounter obstacles sinc e these requirements must be
integrated with other design and operational considerations. Despite
350 (VJEFMJOFTGPS*OIFSFOUMZ4BGFS$IFNJDBM1SPDFTTFT"-JGF$ZDMF"QQSPBDI
#Z$$14
¥5IF"NFSJDBO*OTUJUVUFPG$IFNJDBM&OHJOFFST |
Figure 3-2: Human performance modes, errors and mistakes
Type of human
performance
Type of error or mistake
Example
Knowledge based
Mistake e.g., lack of
knowledge of process
hazards
Unable to understand a
rare process upset
Rule based
Misinterpret an event and
apply wrong emergency
response
Mistake e.g., wrong
procedure selected
Forgetting a step in a long
procedure
Error – slip
Error – lapse
Skill based
Accidentally pressing the
wrong button |
PREFACE
Many adverse events have occurred in industry and elsewhere, due to
abnormal situations that took place, or developed, but were not
recognized in time or managed in a way that could have prevented the
incident. Various tools and techni ques, including complex automated
control systems, are available to he lp manage such situations. However,
such systems are not always effective if operators are not trained on
them properly. By carefully consider ing how these abnormal situations
might occur and by developing me thods to identify, respond and
manage them, the consequences that arise from them can be prevented
or at least mitigated. This book examines such methodologies and
management systems and provides a resource for operations and
maintenance staff to be able to effe ctively handle abnormal events, as
well as reduce the frequency and ma gnitude of process safety events.
The American Institute of Chemic al Engineers (AIChE) has been
closely involved with process safe ty and loss control issues in the
chemical, petrochemical, and allied industries for more than four
decades. AIChE publications and symposia have become information
resources for those devoted to process safety and environmental
protection.
AIChE created the Center for Chemic al Process Safety (CCPS) in 1985
after significant chemical disasters in Mexico City, Me xico, and Bhopal,
India. The CCPS is chartered to develop and disseminate technical
information for use in the prevention of major chemical accidents. The
center is supported by more than 200 chemical process industry sponsors
that provide the necessary funding and professional guidance to its
technical committees.
The major product of CCPS activities has been a series of guidelines
to assist those implementing variou s elements of a process safety and
risk management system. This b ook is part of that series.
CCPS strongly encourages companies around the globe to adopt and
implement the recommendations contained within this book. |
414 INVESTIGATING PROCESS SAFETY INCIDENTS
NOTE 2: It is the intent that the “Potential Chemical Impact” definitions shown in Table 2 to
provide sufficient definition such that plant owners or users of this metric can select from the
appropriate qualitative severity descriptors without a need for dispersion modeling or calculations. The user should use the same type of observation and judgment typically used to
determine the appropriate emergency response acti ons to take when a chemical release occurs.
However, CCPS does not want to preclude the use of a “sharper pencil” (e.g. dispersion
modeling) if a company so chooses. In those cases, the following notes are being provided, as
examples, to clarify the type of hazard inte nded with the four qualitative categories:
A: AEGL-2/ERPG-2 concentrations (as available) or 50% of Lower Flammability Limits (LFL) does
not extend beyond process boundary (operating unit) at grade or platform levels, or small
flammable release not entering a potential explos ion site (congested/confined area) due to the
limited amount of material released or location of release (e.g., flare stack discharge where pilot
failed to ignite discharged vapors).
B: AEGL-2/ERPG-2 concentrations (as available) extend beyond unit boundary but do not extend
beyond property boundary. Flammable vapors gr eater than 50% of LFL at grade may extend
beyond unit boundaries but did not entering a potential explosion site (congested/confined area); therefore, very little chance of resulting in a VCE.
C: AEGL-2/ERPG-2 concentrations (as available) exceeded off-site OR flammable release
resulting in a vapor cloud entering a building or potential explosion site (congested/confined
area) with potential for VCE resulting in fewer than 5 casualties (i.e., people or occupied
buildings within the immediate vicinity) if ignited. D: AEGL-3/ERPG-3 concentrations (as available) exceeded off-site over the defined 10/30/60
minute time frame OR flammable release resultin g in a vapor cloud entering a building or
potential explosion site (congested/confined area ) with potential for VCE resulting in greater
than 5 casualties (i.e., people or occupied buildings within the immediate vicinity) if ignited.
NOTE 3: The Potential Chemical Impact table reflects the recommended criteria. However, some
companies may object to making a relative rank ing estimate on the potential impact using the
terms described. In those situations, it would be acceptable for those companies to substitute
the following criteria corporate wide: Severity Leve l 4: 1X to 3X the TQ for that chemical, Level
3: 3X to 9X, Level 2: 9X to 20X, and Level 1: 20X or greater the TQ for that chemical. However, if
a company elects to use this al ternative approach they should be consistent and use this
approach for all releases. They should not select between the two methods on a case-by-case basis simply to get the lowest severity score.
NOTE 4: The category labels can be modified by individual companies or industry associations
to align with the severity order of other metrics. It is important is to use the same severity point
assignments shown.
NOTE 5: The severity index calculations include a category for “Community/Environmental”
impact and a first aid (i.e., OSHA “recordable inju ry”) level of Safety/Human Health impact which
are not included in the PSI threshold criteria. However, the purpose of including both of these values is to achieve greater differentiation of severity points for incidents that result in any form
or injury, community, or environmental impacts.
|
Fundamentals of Instrumentation and Control
247
number is 1001. Sometimes the elements of a single
control loop do not all appear on the same P&ID sheet, so if these elements have all been assigned the same sequence number, we can still find them easily in different P&ID sheets.
Now, we will continue to look at the second example of
a simple level control loop.
Figure 13.6 illustrates fluid level control in a tank,
rather than fluid temperature control.
The devices in this loop are defined as follows:
●LE: level element, or sensor
●LT: level transmitter
●LC: level controller
●LV: level control valve.
There is no “LE” in Figure 13.7 and nothing is missed
there! There are some cases that there is no specific sen-sor for a process parameter and the signal is initially developed in the transmitter, which is “LT” here.
For level measurement, there are two types of sensor:
contact sensors and roof sensors.
Contact level sensors are the most famous sensors in
that their signal is developed in their transmitter and no “LE” exists for them.Using the above definitions, we can outline the
sequence depicted in Figure 13.6:
1)
The L
T transmits the level value in the tank, as
measured by the contact sensor, to the LC. The
transmission is represented by the dashed line.
2) The LC c
ompares the received value against its regis -
tered SP .
3) The LC s
ends a level adjustment signal to the LV.
4) The L
V adjusts the control valve to modify the
outflow from the tank in accordance with the instructions received from the LC.
For example, let us assume the SP to be 2
m. Thi
s
means the level needs to be constantly adjusted to 2 m
fr
om the bottom of the tank. If the level is reported at
2.5 m, the LC c
ompares this value against the 2 m SP and
r
ecognizes that the tank contains an average of 0.5 m
more t
han the SP . Therefore, the LC sends a signal to the
LV to adjust the level downward by opening the valve to increase the outflow from the tank.
This example illustrates that each control loop has
three functions: measuring, comparing, and adjusting.
13.8 Instruments on P&IDs
A piping and instrumentation diagram (P&ID) is a sche-matic drawing used to illustrate all of the elements used in the control of a process. It is a diagram that shows how all the pieces of process equipment are interconnected, together with the instrumentation used to control the process. Symbols used for both equipment and instru-mentation conform to the global guidelines given by the ISA, the International Society of Automation.
13.8.1
Fundamen
tal Terminology
I would like to explain some fundamental terms that
are used in process control, specifically relating to
con
trol loops (Figure 13.7). First, you have a primary
element (sensor), which is usually an instrument that measures a process variable. A signal is sent via a transmitter to a controller. The controller then sends an adjustment signal to the final element.
The final element is some mechanical means to
con
trol
the process. This is often a control valve on a pipe, or a variable speed drive for a pump or compressor.
13.8.2
Identifiers f
or Equipment
and Instrumentation
Here we want to learn the identifiers of instruments.
We will start our discussion with a table that compares equipment versus instrument identifiers.1Measure
3AdjustLC
1001Set point
LT
1001
LV
10012Compare /uni290D to set pont
Figure 13.6 Lev el control loop.
Primary element
Final elementInstruments
SignalsFT
215
FE
215FC
215
FV
215
Figure 13.7 Fundamen tal terminology. |
104
design and create an over-co mplicated process (or one which
relies on control of hazards).
Flexibility and redundancy . While some level of redundancy may
be necessary and desirable wi th basic process equipment,
particularly where the failure of the component will have serious
effects, this should be limited to what carefully performed PHAs
and other studies reveal as the correct level. For every extra pump, heat exchanger, or othe r basic component, additional
controls, utility requirements, piping / valves and other mechanical equipment will follow, thereby greatly expanding the complexity of the process. Additionally, not every risk can or should be solved by specifying some piece of equipment to deal with it.
Kletz (Ref 6.9 Kletz 1998; Ref 6. 10 Kletz 2010) also offers some
suggestions on use of simple technologi es in lieu of high or more recent
technologies to solve certain types of problems. One such suggestion is that flare systems should be kept as simple as possible, and not be
equipped with other appurtenances, such as flame arrestors, water seals, filters, etc. These components are prone to breakdowns as the result of e.g.,plugging and reducing flare capacity.
These suggestions describe a de sign philosophy where simple—and
sometimes old technologies work just as well as newer, more sophisticated ones. Such a philosoph y should be employed wherever
possible before resorting to complex solutions.
Examples of simplification are disc ussed in the following sections.
Additional examples can be found in Kletz (Ref 6.8 Kletz 1991; Ref 6.9 Kletz 1998; Ref 6.10 Kletz 2010), and in Chapter 8 of this book.
6.1 LEAVING THINGS OUT
In the spirit of Trevor Kletz’s quote “What you don’t have can’t leak”, an
effective simplification technique is to combine the functionality of two
or more vessels or pieces of eq uipment into one and leave out the
redundant equipment. For example, rather than a separate knockback
condenser installed on a reactor vapor line, the vapor line can in some cases be left uninsulated, and the condenser eliminated. In another example, a refrigeration compressor can have a suction-side catch pot |
336
Historically, an overemphasis on minimizing initial capital investment,
and on time constraints, which often favor active or procedural systems, has resulted in underutilization of i nherently safer solutions. Instead,
there is an increased dependency on alarms and SISs to reach
acceptable risk levels. Economic anal yses in the initial design stages
often fail to take into consideratio n the cost of maintaining and proof-
testing these systems, which can be significant for large process facilities. When comparing inherently safer desi gn solutions to other solutions,
designers should include the total lif e cycle cost of each alternative
before reaching a decision. For example, Noronha, et al. (Re. 13.25 Noronha) describe the use of deflagra tion pressure containment design
in preference to using deflagrati on suppression or other means of
explosion prevention based on life cycle cost and reliability considerations (Ref 13.6 CCPS 1998).
13.5.2 Often more econom ical, but not necessarily
Figure 13.4 presents a comparison of the four categories of design
solutions with respect to several cost and functional parameters, for a
heat exchanger failure scenario. Inhe rently safer/passive solutions (such
as exotic metallurgy) tend to have higher associated initial capital
outlays; however, operating costs ar e usually lower than those for the
other design solutions. For active solutions (such as on-line monitoring and instrumentation), as compare d to inherently safer/passive
solutions, reliability is typically lower, and complexity is greater.
Operating costs are also likely to be the greatest for active solutions . While procedural solutions are tempt ing due to their initial very low
capital cost and typically lower complex ity, they are often also the least
reliable, and should be considered on ly after other solutions have been
explored. (Ref 13.6 CCPS 1998).
|
3 ABNORMAL SITUATIONS AND KEY
RELEVANCE TO PROCESS PLANT
OPERATIONS
This chapter discusses focus areas for the management of abnormal
situations and explains their releva nce to process operating personnel.
It includes plant design aspects, new technologies, operating modes
during which abnormal situations ca n occur. The chapter also provides
the links between abnormal situat ion management and CCPS’ Risk
Based Process Safety elements. The chapter discusses procedures for
managing abnormal situations. Ex amples of abnormal situations
encountered in a variety of example incidents are also included, and the
associated lessons learned can be es pecially valuable for sharing with
frontline supervisor s and operators.
3.1 FOCUS AREAS FOR ABNORMAL SITUATION MANAGEMENT
Several areas relevant to the manage ment of abnormal situations have
been identified by the Abnormal Situation Management® Consortium
(ASMC) as research areas, as summarized.
3.1.1 ASM Research Areas
The ASMC refers to seven areas of research concerning abnormal
situations, recognizing that the connection between the system and the
human, as well as human strengths and limitations, must be understood. A
detailed understanding of these focus areas, and others as outlined in
Section 3.1.2, is required for management of abnormal situations at the
plant level. Most of these areas have a direct link to key elements of Risk
Based Process Safety (CCPS 2007a), as introduced in Section 2.1, the
relevance of which is detailed in Section 3.3.
|
102 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS
4.2.8 Other Parties
Other individuals who could potentially be involved in resolving abnormal
situations are as follows:
Laboratory technicians, who are able to identify or characterize
issues with intermediate products as part of routine or special
troubleshooting analysis.
Incident command center, where key decisions are typically made
including how to deal with losse s of containment, involvement of
external resources, mutual aid, communication with higher
management and local external parties.
Corporate Headquarters, where major strategic decisions are made
on handling significant events and communication with the media,
shareholders, and other interested parties.
4.3 GUIDANCE FOR ORGANIZING AND STRUCTURING TRAINING
4.3.1 Organization of Training
Training workers and assuring their reliable performance of critical tasks is
one of the nine elements in the RBPS pillar of managing risk. Establishment
of a training management system is the initial step, key elements of which
will normally include objectives, measurements, training materials, and
effective trainers. This approach is generally accepted as the fundamental
basis for most programs, although the objectives may vary greatly across
industries and occupations. For exam ple, the objectives of a training
program for astronauts would be differ ent from training for front-line plant
operators. However, both need to understand the functioning details of the
equipment and control systems; and both need to know how to recognize
and respond to abnormal situations.
Training is often conducted within specific workgroups, such as the
operating team, the maintenance team, or the engineering department. For
abnormal situation training, however, it may be more appropriate to adopt
a holistic or “systems-thinking” approa ch when conducting at least some of
the training, and to involve various workgroups and experts.
|
Fundamentals of Instrumentation and Control
265
In the following subsection, we will go through of the
concepts of each of these control loops.
13.12.1 Le
vel Control Loops
A level control loop can be set up for non‐flooded liquid
containers. This means that they are applicable for tanks or non‐flooded liquid vessels.
Figure 13.40 shows a schematic for a typical level loop.As was mentioned, if the level element (sensor) is
connected to the side body of the vessel, we don’t show “LE” on the schematic. Instead, the schematic shows a signal going to the LT and then to the LC. Here I have indicated the SP input, but this not always shown on the P&ID. After that, the signal goes to the converter and then finally to the control valve.
Generally speaking, we always control the liquid level
in non‐flooded containers for the purpose of inventory control. If a container is flooded with a liquid, inventory control can be achieved by a pressure loop.
13.12.2
Pr
essure Control Loops
Pressure control loops can be used on pipes or on con-
tainers. Figure 13.41 shows a schematic for a typical pressure loop.
Process control loops can be used in containers and
pipes and for liquids and gases. Table 13.21 shows these applications.
Pressure control loops are applicable for gases in pipes
and in containers. You can think of gas pressure as simi-lar to liquid level in tanks. Pressure control loops on gas pipes somehow shows “flow” of the pipe!
Pressure control loops are also used for liquid‐flooded
containers.
The use of pressure control loops on pipes is not very
common; however, there are cases where we can obtain benefit from them on liquid‐containing pipes.Below are a few examples of using pressure control
loops for liquid‐containing pipes:
●To protect the downstream equipment, e.g. by open-ing a relieving line.
●To ensure the liquid remains in a liquid state in upstream equipment. This is important when the liquid is at a high temperature, is volatile or entering the upstream equipment at high velocity. For example, you may want to pump a liquid at high temperature using a centrifugal pump. In order to limit the damage to the pump due to gas in the line, you need to use a pressure loop upstream of the centrifugal pump to ensure the liquid doesn’t vaporize.
●On utility headers. For example, on a utility water header, you may need to install a control loop to ensure that the pressure is high enough to feed the plant.
13.12.3
Temper
ature Control Loops
There are instances when temperature control is vital to the operation of a particular piece of equipment. Examples are furnaces, boilers, heat exchangers and temperature‐fixed reactors.LT
100SP
LV
100LC
100
I/P
LT
100
Figure 13.40 Lev el control loop schematic.SP
PC
100
I/P
PY
100
PV
100PT
100
Figure 13.41 Pr essure loop schematic.
Table 13.21 Applica
tion of pressure control loops.
Liquid Gas/vapor
Container Only if the container
is floodedP‐loop
Pipe Not common P‐loop
(or F‐loop if it is around a gas mover) |
Selecting an Appropriate PHA Revalidation Approach 103
5.3 PRINCIPLES FOR SUCCESSFUL REVALIDATION
APPROACH SELECTION
Actual experience in conducting revalid ations, by PHA practitioners across a
number of companies, has highlighted so me keys for success, as well as some
things that can impede success. While none of the items listed are absolute rules,
they do provide valuable guidance.
Successful Practices:
• Beginning the revalidation approach decision process well in
advance of the revalidation due da te so time constraints do not
unduly affect the choice of revalidation approaches
• Using the Redo approach for any new and separate requirements
(e.g., a quantitative analysis of facility siting issues) that do not
affect the core analysis scenarios, and Updating the core analysis
• Reviewing process safety manageme nt system audit (internal or
external) recommendations and other process safety performance
indicators prior to deciding if a Redo or Update is appropriate, and
ensuring any deficiencies are addressed during the revalidation
• Performing a Redo periodically (e.g., every 2nd or 3rd cycle)
• If performing a Redo based on schedule (e.g., a Redo is being
performed on the third cycle or 15 years after the initial PHA),
including a post- Redo gap analysis to help ensure no scenarios in
the prior PHA were missed or interpreted incorrectly
Obstacles to Success:
• Completely ignoring the prior PHA when using the Redo approach
and losing process history and design knowledge contained therein
• Failing to methodically evaluate the prior PHA and operating history
when selecting a revalidation approach |
8 • Emergency Shutdowns 154
engineering or administrative controls usually activated at the time of
the incident.
8.6.2 Incidents occurring during the emergency shutdown time
C8.6.2 -1 – DPC Enterprises, L.P. [83]
Incident Year :2002
Cause of incident occurring during the emergency shut-down : Upon
activation, the emergency shutdown system (ESS) ball valve did not
work and did not stop the chlorine release.
Incident impact : Failure of a chlorine railcar unloading hose resulted
in release of 21,800 kg (24 tons) before emergency responders could
stop release. 66 people sought medical evaluations; three were
hospitalized. Trees and other vege tation surrounding the unloading
station were damaged.
Risk management system weaknesses:
LL1) At the time of the incident , DPC did not have an adequate
Inspection, Testing, and Preventive Maintenance (ITPM) program to
ensure asset integrity and reliabili ty. In particular: 1) the transfer
hoses did not meet design specifications (there was no “positive
materials identification” progr am); 2) the Emergency Shutdown
System (ESS) ball valve did not work when needed due to severe
build up (it had not been tested).
Relevant RBPS Element :
Asset Integrity and Reliability
LL2) At the time of the incident, DPC did not have a clear emergency
response plan, did not provide adequate accessibility to its
emergency response equipment, did not perform emergency
response drills, and had not invol ved the local emergency response
planning committee.
Relevant RBPS Ele ment:
Emergency Management |
22. Human Factors in emergencies 287
Training individuals in stress management techniques for coping with stress
reactions can take the form of:
• General exercises, such as realistic situations or case studies.
• Specific techniques, such as simulated emergency drills.
Building experience of stressful situations increases individuals’ coping ability,
builds confidence, and reduces the likelihood of stress and consequent cognitive
deterioration or paralysis.
Training content on coping with stress in emergency situations includes
techniques that directly target stress resp onses. It also includes techniques to
increase technical skill proficiency e.g., automated task execution.
Examples of these techniques are shown in Figure 22-4.
“The OIM had gone a matter of seconds when he came running back in what
appeared…to be state a panic…The OIM ma de no specific attempt to call in
helicopters from the Tharos (a rescue ve ssel) or elsewhere, or to communicate
with the vessels around the installation, or with the shore or other installations;
or with personnel on Piper…” (para 8.9 [93, pp. 152-153])
“The OIM did not give any other instructions or guidance. One survivor said
that at one stage people were shouting at the OIM and asking what was going
on and what procedure to follow. He did not know whether the OIM was in
shock or not, but he did not seem to be able to come up with an answer.” (para
8.18 [93, pp. 156-157]) |
176 | 5 Aligning Culture with PSMS Elements
Emergency Management (Element 16)
When process safety incidents occur, facility personnel should
take actions that help reduce the consequences of the incidents.
These actions include evacuation to a safe location, use of
emergency m asks, sheltering in place, first response, offensive
response (e.g. to close an isolation valve), and firefighting, among
others.
Since each emergency is different, it is impossible to develop
specific procedures to address every scenario. Instead, specific
emergency m anagement personnel need to be expert at putting
together the skills and resources at the disposal to effectively
address the emergency. Everyone else at the site needs to be
trained on a range of specific emergency management skills.
Training should be done regularly, so everyone at the facility can
carry out their role correctly and without delay.
In many plants, emergency m anagement personnel m ay come
from outside the plant. This can include industrial neighbors who
partner with the facility in a mutual aid agreement as well as
emergency responders from the local community. The cultural
implications of these external stakeholders were discussed in
section 5.1.
Emergency management can readily becom e subject to
norm alization of deviance. Since process safety incidents are
infrequent, it can be easy to forget to plan, evaluate emergency
procedures, and conduct drills. Ironically, the temptation to
deviate from emergency preparedness could increase as culture
and PSMS performance improves and incidents become even less
frequent. However, emergency management is an integral part of
risk m anagement, and must be m aintained, just as process
equipment must be m aintained. Culturally, emergency
m anagement should be treated as part of the im perative for
process safety and m onitored through the management review
elem ent (see section 5.1). |
226 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION
Different from corrosion, erosion is also a po tential threat to piping integrity. Erosion
occurs as a material mechanically damages or th ins the pipe wall. This can be caused by flow
of the catalyst, such as in catalytic cracking, fl ow through the piping. In the upstream industry,
it can be caused by sand that is entrained in th e crude oil. Although it is a different mechanism
than corrosion, the threat of lo ss of containment is the same.
Erosion and corrosion can team up to accelerate piping thickness loss. Some materials
corrode, and the corrosion product forms a passi ve protection layer which limits corrosive
damage. This protective layer can be strippe d away by erosive forces, which exposes the
underlying material to renewed corrosive attack.
An additional threat, primarily to small bore pi ping, is vibration. Small bore piping failures
occur more frequently than for larger piping. Ma ny releases have occurred due to vibration of
small-bore piping such as instrument piping, which led to piping failure. Vibration can be
caused by induced vibration, pulsating equipm ent such as reciprocating pumps, equipment
subject to ocean waves, and fluid shock or ‘ha mmer’ caused by rapidly stopping or starting
flow. (CCPS 2020)
Design considerations for process safety. Recognizing the potential for piping corrosion
is the first step. The piping material may be selected that will not corrode in the service
conditions, the piping may be designed to with stand the corrosion for many years by providing
appropriate wall thickness, or the process may include a chemical injection to prevent or
minimize the corrosive impact. Considering corrosi on of materials in the design stage is only
the first step; monitoring it throughout the lif e cycle is required. This is addressed in the
following section.
With respect to vibration, a challenge is that frequently only larger piping is shown on
piping isometrics with a note that small-bore pi ping is field installed. The result is that the
installation is dependent on the skill of the inst aller and may not be subject to the engineering
review that other piping and equipment receiv es. The length of unsupported or unrestrained
piping should be reviewed. To isolate vibration, flexible connections may be used, but they are
also weaker components that can fail. Both the amplitude (amount of movement) and the
frequency (rate of movement) can affect how qu ickly vibration can cause equipment to fail.
Technology exists to test and analyze vibrat ion to determine the exact source. (CCPS 2020)
In addition to piping, flexible hose asse mblies may be used, typically in loading and
unloading operations, to transfer materials. “G uidelines for the management of flexible hose
assemblies” provides information on maintain ing the integrity of these systems. (EI)
Asset Integrity and Reliability
Asset integrity and reliability is the RBPS element that helps ensure that equipment is properly
designed, installed in accordance with specificat ions, and remains fit for use until it is retired.
The previous sections addressed how equipment can fail and provided design considerations
for process safety Even with the best design, integrity issues can occur during operations.
Putting in place a system to manage asset inte grity and reliability is important to production
and process safety. |
383
unlikely that requirements for cond ucting IS reviews and implementing
such technology “where practicable” will necessarily result in large-scale
risk reduction against security-related risks.
14.3.1 Consistent Understanding of Inherent Safety Misunderstandings or misperceptions about IS tend to localize around
four concepts – goals , applicability , scope and economic feasibility .
The goal of both a safety and security program should be to reduce risk.
Inherent safety is an approach to re ducing and managing risks; it is not
an end in itself. IS policies and regu lations will be most successful when
they clearly state a risk reduction and management goal with a recognition that some risks are inhe rent to the production of some
critical goods and services, and that such risks can be managed within
acceptable ranges.
IS may be applicable to existing as well as new facilities and processes .
There may be a perception that IS is relevant only for new facilities and
that there are no feasible opportunities once the process is operational.
While it is true that the potential for major improvements may be
greatest during process development, this book has demonstrated that
facilities have reduced or even e liminated hazards or have managed
change to avoid new hazards by applying IS methods throughout the facility life cycle. The majority of th e applications for IS are with the
installed industrial base, whereas th e feasibility of applying IS to the
fullest diminishes as the facility is built. This leaves many companies
where new processes (and particul arly new technologies) are rarely
implemented with fewer occasion s to practice the methods.
The scope of IS is not limited to hazard reduction ., IS concepts are
applicable to the layers of pr otection surrounding the remaining
hazards. A narrow view argues that IS only applies to major changes in the degree of hazard, while a broade r viewpoint finds any changes that
increase safety through the application of IS principles to be an advantage. This includes the prog rammatic aspects of PSM programs
(see Chapter 11).
Changes to facilities and processes must be economically feasible . Costs
are a primary concern when consid ering modifications to existing
facilities. Inherent and passive approa ches are strategic, usually must be |
168 PROCESS SAFETY FOR ENGINEERS: AN INTRODUCTION
Figure 10.7. Asset life cycle stages including project phases
(CCPS 2019 a)
Table 10.1. Asset life cycle stages including project phases
(modified from CCPS 2020)
Asset Life Cycle
Stages including
Project Phases Selected Activities and Process Safety Studies
Appraise phase Develop and evaluate a broad range of project options, assess commercial
viability, and rank feasible options to take forward. This stage is also referred to as
Front End Loading (FEL) 1.
Studies: Preliminary Hazard Analysis
Select phase Evaluate concept options, maximizing opportunities and minimizing threats or
uncertainties. A single concept to progress is normally chosen at this stage. This
stage is also referred to as FEL 2. Evaluate inherently safer design options.
Studies: Preliminary Hazard Analysis, What-If Analysis
Selected PSI: chemical properties and composition
Define phase Develop a basic design including plot plan, process flow diagrams, material and
energy balances, and equipment data sheets. Schedule and cost are updated, and
financial investment decisions may be made. This stage is also referred to as FEL
3.
Studies: What-If Analysis
Selected PSI: operating limits
|
332
13.3 INHERENT SAFETY – ENVIRONMENTAL HAZARDS
13.3.1 PCBs Polychlorinated biphenyls (PCBs) were originally introduced in the 1930s
as non-flammable cooling and insulating oils for electrical transformers.
PCB manufacture was banned in May 1979 due to environmental concerns (Ref 13.1 Boykin). This is an example of how new data and
information led to a change in the us e of a material due to an improved
understanding of its hazards and a reevaluation of the relative importance of different types of hazard (Ref 13.16 Hendershot 1995).
13.3.2 CFCs With current concerns about the ad verse environmental effects of
chlorofluorocarbons (CFCs), it is easy to forget that these materials were
originally introduced as inherently safer replacements for more
hazardous refrigerants then in us e. These included ammonia, light
hydrocarbons, such as isobutene, ethyl chloride, methyl chloride,
methylene chloride and sulfur di oxide (Ref 13.18 Jarabek). These
materials are flammable, acutely toxic, or both. A release of one of these
substances in the home potentially causes an immediate fire or toxic exposure hazard.
Thomas Midgley, Jr. dramatically introduced CFCs in a lecture to the
American Chemical Society in 1937. Midgley filled his lungs with CFC
vapors, and then exhaled, extingui shing a candle. This graphically
illustrated that CFCs are not flammable or acutely toxic (Ref 13.20 Kauffman). Now, after many decades of use, we have discovered that
CFCs cause environmental damage by depleting the stratospheric ozone
layer, and their use is being phased out. Since it is unlikely that our
society will give up refrigeratio n or air conditioning, substitute
refrigerants are needed. Perhaps hydrofluorocarbons (HFCs) and hydrochlorofluorocarbons (HCFCs) emerged as safe and environmentally acceptable replac ements (Ref 13.30 Wallington).
H o w e v e r , i n s o m e c a s e s , w e a r e g o i n g b a c k t o t h e r e f r i g e r a n t s w h i c h CFCs originally replaced. For ex ample, home refrigerators using
isobutene refrigerant are now available in Europe. The manufacturers are not producing “frost-free” versions of these refrigerators because of concerns associated with ignition hazards because of the small heater |
Plant Interlocks and Alarms
345
Alarm management specifies all the activities related to
adding an efficient alarm system in a plant. Alarm man-
agement requires a specific skill and some control engi-neers are specialist in this area. Here, again, we do not want to go to deep knowledge of alarm management.
16.10.1
Ana
tomy of Alarm Systems
Figure 16.23 shows the anatomy of an alarm system.
In an alarm system there is a sensor that sends a signal
to a logic processor, and then to an alarm.
The primary element of an alarm system could be a
sensor or switch. The sensor/switch could be linked to
any of the process or non‐process parameters such as level, flow, temperature, pressure or composition.
The logic could be a simple logic handled by a DCS or
PLC. There are even cases that for a single parameter one alarm, e.g. at low level, is handled through a DCS but the low–low level of the same parameter is handled by a PLC system.
There are also cases that the primary element of an
alarm system is connected to the final element through hard wires. These alarms are named hard alarms to be recognized from “soft alarms” , which are through a DCS or PLC.
The final element of an alarm system or alarm could
be classified in different ways. Alarms can be audible or visual and can be in on monitor screen or a stand‐alone type.
Table 16.4 shows different types of alarms.Alarms can also be classified by their location, i.e. in
the field or in the control room. When they are in the field, they are definitely stand‐alone alarms. The field alarms are mainly audible type.When they are in the control room they are, theses
days, on monitor screens. There could still be, in some cases, alarms in control rooms that are installed on a panel, in the form of stand‐alone alarms.
16.10.2
Alarm R
equirements
An alarm is an alert system to notify the operator of an
“upset” event. The logical reply for an alarm is an action or a set of actions. The action could be done by SIS (automatically actions) or by plant operators.
There are two questions regarding alarms on P&IDs
that needs to be answered: which parameter needs an alarm, and which level of parameter needs to be used for the alarm triggering point.
Deciding about parameters needing an alarm is not
always easy.
The designer has to optimize the number of alarms in
the system. Very few or too many alarms are dangerous for a plant. If the intent is very few alarms, some impor -
tant alarms may be ignored, which is dangerous. Too many alarms may lead to a situation where the operator becomes insensitive to them and doesn’t deal with them with the required degree of urgency. This is again dan-gerous for the plant.
In some instances, if there are too many alarms in the
control room, the operator becomes overloaded and some of alarms may be overlooked.
One rule of thumb is that if a SIS exists for a parameter,
it definitely needs an alarm too. If a safety interlock Table 16.3 Fea tures of two types of alarm systems.
Group 1 Group 2
Name Alarm system Fire and gas detection system (FGS)
Schedule of action Before loss of containment After loss of containment
Point sensitive Space sensitive
Inside of units Outside of units
Target parameters Temperature, pressure, flow rate, level,
composition and non‐process parametersOnly fire and gas
Footprint on P&ID Can be seen on every sheet of P&ID Could be only several sheets
dedicated to it as “FGS P&IDs”
Sensor Logic Alarm
Figure 16.23 Ana tomy of an alarm system.Table 16.4 Differ
ent types of alarms.
Audible alarms Visible alarm
Monitor screen alarmBeeper Flashing icons
Stand‐alone alarmBuzzer, horn, siren, bellFlashing lamps, rotating lights, strobe lights, beacon |
W ITNESS M ANAGEM ENT 119
Promptness in gathering information is critical. Information from people
is among the most fragile form of evi dence, (i.e., it is easily forgotten,
distorted, or otherwise influenced by personal conflicts.) For most people,
short-term memory for retaining an d recollecting details degrades rapidly.
The second reason for promptness is rooted in the fact that contact and
communication with others can significantly affect our “independent”
recollection of occurrences. It is best to prevent any exchange of information
among witnesses, if possible, immediat ely after an event. In most cases,
complete isolation is not practical, so as a minimum, the witnesses should be
asked to refrain from discussing the in cident with anyone until their initial
interview. The use of social media makes this a challenge.
The interaction among witnesses ca uses modulation of details and
changes emphasis both consciously and subconsciously. Recollection is
affected by our emotions, by perceived unf airness, by fear of embarrassment,
by fear of becoming a scapegoat, and by preexisting motives, such as grudges and attitudes. Many
people are so reluctant to be identified as
betraying their peer group that they may withhold information if they
perceive the peer group would desire them to do so. There is often value in
repeating portions of the interview; a witness might be stimulated by reviewing his or her own initial testimony.
Investigations involving complex human performance problems can
benefit from simulations. Process simu lators are often used for operator
training. In some cases, these process simulators can be excellent tools for
learning more about human error causation. The incident investigation team can expose operators to simulated proce ss upsets and gain valuable insights
into the operator’s response to rapidl y and accurately diagnose the problem
and execute the proper action.
The talk-through exercise is a technique sometimes used by
investigators to gain insight and to verify conclusions drawn from verbal
testimony. This technique, often used by human reliability analysts, has
particular application for learning more about specific tasks or occurrences. It is a method in which an operator describes the actions required in a task,
explains why he or she is doing each action, an d explains the associated
mental processes. To be effective, su ch exercises must be planned by the
investigator. The actual talk-through itself is seldom very time-consuming,
but the burden is on the investigator to take good notes and observe any
potential problem areas. When the procedures call for the manipulation of
a specific control or for the monitoring of a specific set of displays, the
operator and the investigator approach them at the cont rol panels and the |
3.2 Characteristics of Leadership and Management in Process Safety Culture |89
Knowledge of management systems, particularly the
com pany’s PSM S,
Ensuring that employees (and managers themselves)
operate within the constraints defined by the PSM S and
the operating and m aintenance procedures, Attention to detail, particularly of m aintaining safeguards
in full working order and approve all safeguard bypasses
according to the corporate policy; and Verification of the PSMS perform ance within their scope of
control
Candidate m anagers’ com petence related to the PSMS should
be screened before their appointment. Any additional training or
coaching needed should be identified and provided. While useful
for all competencies, a thoughtful and up-to-date succession plan
(See section 3.5) and organizational m anagement of change
procedure (OM OC) is especially helpful for process safety. Poor
m anagement skills are a key cultural warning sign (Ref 3.16) of
potential catastrophic incidents.
B e Visible Leaders should be visible in the field to evaluate conditions,
understand site specific process safety issues, and be available to
answer questions. Leaders should com m unicate process safety
issues and requirements to site personnel in person, and seek
productive feedback. They should engage the organization and
assess if the line organization understands their responsibilities
and perform ance expectations. Leaders should m ake process
safety expectations and evaluations visible and explicit in their
team m em bers’ individual goals and performance reviews.
Drive Good Morale, Especially During Change Morale influences culture. Many of the things that drive good
m orale, such as trust, open comm unication, and a com mon •
•
•
• |
78 PROCESS SAFETY IN UPSTREAM OIL & GAS
A weakness of the FMEA method is that it does not consider human factors
well. Brainstorming techniques such as What-if and HAZOP are superior in that
regard. These methods address both the ergonomics of systems (e.g., display layout,
physical effort required) and factors that can degrade human responses (e.g., stress,
fatigue, information overload). These can be qualitative or quantitative in approach.
Fault Trees
Fault trees help to understand how complex systems can fail and identify less
obvious problems such as common mode failures. The main application of fault trees
for well construction relates to BOP operation and reliability prediction. The fault
tree method is described in detail in Section 6.3.3.
Risk Ranking Assessment
Risk ranking is an optional additional step in hazard evaluation that can be applied
to most methods (e.g., PHA, What-If, HAZOP, FMECA). It is now very common
to extend hazard identification to include risk ranking. But for this to be effective,
the company should select a single risk ma trix for its decision making and not have
each project choose its own. This helps w ith consistency in pr ioritizing decision
making.
Teams examine each scenario and assign a consequence and likelihood level.
Many companies have their own risk matrix for this purpose, or they may use the
version in ISO 17776 (2016) with six levels of consequences and four levels of
likelihood. The number of levels must match the risk matrix being used as the results
are plotted onto the matrix. Figure 4-4 shows three decision bands.
●lower risk – manage for continual improvement
●medium risk – incorporate risk reducing measures
●higher risk – fail to meet scr eening criteria, change required
Other risk matrices employ more or fewer levels (e.g., a 5 x 5 matrix is common)
and some matrices include four bands for decision making. Likelihood is often easier
for teams to assess when expressed qualitatively as in ISO 17776, rather than
quantitatively as in some matrices which specify a frequency band (e.g., 10-4 to 10-
3 per year). The bands provide a consistent basis for decision making for many
scenario decisions. Generally, more senior levels of management are involved in
making decisions regarding higher levels of risk or where mitigations may be
difficult to implement.
The risk matrix shown provides for four types of consequences – to people,
assets, the environment, and reputation. Consequences are usually judged on
reasonable worst case, but individual companies have their own approaches.
Although the risk ranking approach is simple and easy for teams to understand,
there are some disadvantages. Teams may have difficulty selecting the likelihood
category if they are no t aware of the wider industry hi storical record. Also, the risk
matrix approach applies decision making to one risk at a time. It does not accumulate
risk. Thus, many risks all assessed at the lower risk category, when |
4.3 Maintenance of Barriers/Barrier Integrity | 45
Companies throughout the industry continue to forget that if barriers are
not adequately maintained, the incidents they are designed to prevent can
happen. Bhopal, MP, India, and Buncefield, Hertfordshire, UK, two landmark
incidents where barrier integrity was compromised, will be discussed in detail
in Chapter 8.
Storage tank overflow incidents are a recurring example of failure to learn
about barrier maintenance. Table 4.2 provides a sampling of these incidents.
Table 4.2 Storage Tank Overflow Incidents
Year Location Material
1983 Newark, NJ, USA (CSB 2015) Gasoline
1988 Yamakita, Kanagawa, Japan (ASF) Hydrogen peroxide
2005 Buncefield, Hertfordshire, UK (HSE,2011) Gasoline
2008 Petrolia, PA, USA (CSB 2009a) Oleum
2009 Bayamón, PR, USA (CSB,2015) Gasoline
2011 Reichstatt, Bas-Rhin, France (ARIA 2013) Gasoline
2014 Fukushima, Japan (BBC 2013) Radioactive water
Many of these incidents could be described with the same report,
changing only the place, date, and chemical name.
In their investigation of the
Bayamón incident (Figure 4.2), the CSB
found poor maintenance of level
indicators and alarms, inadequate
redundancy, and a poor safety
management system, citing similarities
to the Buncefield incident 4 years
earlier. The facility failed to maintain
their level indicators, relying instead on
manual calculations to estimate level.
The facility also failed to maintain the secondary containment barriers, leaving
the dike drain valve open, allowing spilled gasoline to enter the waste
treatment plant, where the vapors ignited. Additionally, the CSB noted that a
safeguard protecting against hurricane-force winds may have exacerbated the
consequences of the release (CSB 2015):
Similar to the Buncefield incident, during the overflow, gasoline sprayed
from the tank vents, hitting the tank wind girder and aerosolized, forming
a vapor cloud, which eventually ignited. Figure 4.2 Bayamón Fire
(Source: CSB 2015)
|
DETERM INING ROOT CAUSES 205
This approach would consider evidence gathered related to the
following issues:
• How did the oil come to be on the floor in the first place?
• What is the source of the oil?
• What tasks were underway when the oil was spilled?
• Why did the oil rema in on the floor?
• Why was it not cleaned up?
• How long had it been there?
• Was the spill reported?
• What is the usual condition of walking surfaces in that unit?
• What influenced the employee to step into the oil?
• What type of shoes wa s the employee wearing?
• Why didn’t the employee go around the puddle of oil?
• Was the area barricaded to prevent entry?
• Are there training or consistency of enforcement issues involved?
As these questions are answered, the continuing prompt for a better
understanding of why the incident occurred should be, “Why? Why did this
particular event occur?” These answers take the investigators deeper into the
origin of the incident. Once this evi dence has been analyzed and the causal
factors identified, the root cause analysis can commence to identify
weaknesses in the management systems involved. For instance, if the oil was
determined to have leaked from a defective container, one might ask:
• Why was a defective container used?
• What are the procedures for inspecting, repairing, or replacing the
containers?
• Are the procedures clearly understood and enforced?
• Is the system to manage the cont ainers properly designed or are
there gaps?
If a failure occurs and no changes are made to the management system,
then the failure will likely occur again. Often corrective action is taken — yet
the failure still recurs. Frequently this is because the corrective actions
address symptoms rather than root causes. |
21. Fostering situation awareness and agile thinking 259
Figure 21-1: Behavioral Markers for “A ctively seeks relevant information”
21.2.3 Training Techniques and Assessment
Situation awareness training methods include:
• Information-based methods in a classroom
setting. The training could include an
interactive deck of slides, case studies, and
group activities. The case studies would aim to
engage trainees’ cognitive processes and
deepen their understanding of situation awareness.
• Practice-based methods using simulations, where specific cues and
events can be manipulated, along with workload and distracting
conditions. The simulation scenario ca n be stopped at any time to assess
trainees’ situation awareness, followed by review and coaching.
See Chapters 13 and
14 for more
information on
training and
assessment.
Regularly checks key sources of information including
alarms and other prompts
Makes use of all available information sources – e.g.,
instruments and colleagues – to check status of the
operation or assumptions about the operation
Shows concern and takes action if important information
is not available when it is needed
Asks for regular updates fr om colleagues who may have
relevant information
Is proactive in addressing missing relevant information |
176 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS
THE CAUSE OF THE AUTOPILOT DISCONNECT– ICE BLOCKAGE
The freezing and blockage of one of more pitot tubes led to a
discrepancy in the airspeed from the three pitot tubes and the
disconnect of the automated flight control systems. This was due to a
combination of the atmospheric condit ions and the design of the pitot
tubes that had an evolving history of blockage due to ice. The report
stated that this was a known, but misunderstood phenomenon
at the time.
The expectation was that pilots woul d be able to recognize what had
happened and take appropriate acti on using a standard procedure.
However, this did not happen for th is event, nor for several of the
previous occurrences of pitot tube blockage.
The recovered flight data record er (FDR) shows the CAS that is
available to the PNF, irrespective of the source of the data that can be
switched between ADIRU 1 and 2, an d the CAS from th e ISIS standby
system. The data displayed on th e PF’s display is not recorded.
I t c a n b e s e e n f r o m F i g u r e 7 . 3 (BEA 2012) that the CAS diverged
significantly at about 02:10:06, briefly came together again at 02:10:17
and then diverged from 02:10:34, befo re coming together again at about
02:11:07. This was a result of the pitot tubes freezing and unfreezing at
different times, due to the extreme ic ing conditions outside the aircraft,
associated with the meteorological conditions.
This trend data, obtained from the FD R, is not available to the pilots;
they are provided with instantane ous readouts via their respective
screens, which may have been displa ying different data derived from
different pitot tubes (Tube 1 or 2).
|
124 | 4 Applying the Core Pr inciples of Process Safety Culture
and some are not. Com mon ethical dilemmas that could occur
include:
Hazard Identification and Risk Analysis: Intentionally
refraining from needed or relevant recommendations
because of the fear of m anagement response to them.
Also, reducing the frequency or consequences of a
scenario to avoid a costly or difficult action item. Auditing: Deleting or altering findings because of business
concerns or embarrassment to those responsible for the
PSMS. Also bowing to pressure from superiors to
downplay the risk of audit gaps.
MOC: Creating MOC records after a change has been
implemented, or marking MOC action items as com pleted
when they are not. MI: Definition of what is overdue for ITPM tasks and
deficiencies so that the num ber of overdue ITPM tasks and
open deficiencies is artificially low. For exam ple, defining
overdue as any tasks that were due by a certain date but
ignoring those still not performed from before that date. Metrics: Definition of PSMS Key Performance Indicators
(KPIs) to make the program appear to be in better shape
than it really is. For exam ple, liberal extension policies for
overdue PHAs, incident investigations, audits, and other
PSM-related action items so that they can be easily
deferred so they are not captured in metrics.
4.4 EXTERN AL IN FLUEN CES ON CULTURE
Nearly all facilities interact with m any external parties. Each m ay
have a different culture than the facility, and exert an influence on
the culture of the facility. Understanding these external cultures
is im portant to encourage supportive external cultures and
defend against cultures that could have a negative influence. •
•
•
•
• |
OVERVIEW OF RISK BASED PROCESS SAFETY 45
Example Incident: P-36 off Brazil
The P-36 FPU explosion and sinking event offshore Brazil (Barusco, 2002) is
an example where hazard identification did not address a possible problem. The
event was initiated by a pressure burst of a drain tank in one of the leg columns.
The cause was an incorrect isolation of the vessel during long-term maintenance
as one connection was isolated using only a valve, not a blind. That permitted
drain water containing hydrocarbons to seep past the valve into the vessel. The
relief had also been isolated, and this meant that the internal pressure increased
as the liquid volume built-up compressi ng the trapped vapor space above. The
vessel ultimately burst and released flammable hydrocarbons into the column
space.
The emergency response team was not aware that any hydrocarbons would be
present, and they accidentally ignited th e flammable mixture. This killed several
members and ruptured the main cooling water supply line that ultimately sank
the vessel. The team and operations personnel did not recognize the potential
presence of the hydrocarbon hazard.
RBPS Application
Safe Work Practices directly addresses the need fo r safe work practices to be
correctly followed. In this case it would refer to Isolation Procedures.
HIRA (Hazard Identification and Risk Analysis) should have identified the
potential for hydrocarbons to be present in the column area due to the direct
connection of the process into the drain tank. This should have been
communicated as part of the emergency response procedures.
Equipment and control systems can be affected by harsh onshore and offshore
environments. Some equipment can be hard to inspect, particularly on offshore
installations. Offshore and remote onsho re installations may have accommodation
limits that reduce the availability of visiting personnel to perform integrity tasks. In
the EU offshore, and for many companies onshore and offshore, there is a focus on
safety critical elements and achi eving performance standards.
Upstream reservoirs decline with time and new wells may be drilled or
stimulation activities with potentially corrosive chemicals employed. This may
bring asset integrity issues. Similarly, many upstream facilities are operating beyond
their intended design life and are managing aging issues.
RBPS Element 11: Contractor Management
Contractor management is a system of controls to ensure that contracted products
and services support (1) safe operations and (2) the company's process safety and
occupational safety performance goals. It includes the selection, acquisition, use,
and monitoring of contracted products an d services. These controls ensure that
contract workers perform their jobs safely, and that contracted products and services
do not add to or increase safety risks. |
164 Human Factors Handbook
15.2.3 Contributing Human Factors
Fatigue can reduce a person’s ability to process information, reduce levels of
attention and alertness, impair memory, reduce reaction time both physically and
cognitively, impair physical coordination , and potentially cause errors. Examples
include:
• Forgetting which steps have been completed in a procedure, due to
memory lapse or incorrectly thinking something has been done.
• Being unable to understand information such as from instrumentation.
• Being unable to understand what is happening or to make a decision.
• Inflexible thinking and poor planning.
High levels of fatigue can cause people to uncontrollably fall asleep or have
“micro naps”. People will not be aware that their performance is affected by fatigue
and may think incorrectly that they can “power through” or use stimulants such as
caffeine, to combat fatigue. This is not true.
Sleep allows the brain to recharge and remove toxic waste by-products which
accumulate when awake. Sleeping helps to “clear” and reset the brain. A reduction
or disruption of the sleep cycle prevents the brain from maintaining their normal
function. Sleep is important for optimal cognition and judgement.
Common causes of fatigue include:
• Working for long periods without a rest break.
• Working many hours in one day.
• Working many days without a rest day.
• Working a shift system that disrupts the sleep/wake cycle.
• Night working and early starts.
Lack of sleep and inadequate rest breaks will affect more complex tasks and
tasks that require judgment and decision-making more so than simpler tasks.
However, tasks that place very low levels of demand on people, such as monitoring
a process, are also vulnerable to fatigue.
Fatigue is a decline in physical and/or mental performance caused by factors
such as prolonged exertion, insufficient sleep, and/or disruption of the
sleep/wake cycle. Fatigue manifests as a sense of tiredness, weakness, or lack
of ener gy. |
FIRE AND EXPLOSION HAZARDS 59
Training and Performance Assurance. Initial and annual safety training was done, but it
seems to have focused on occupational safety. Safety training had not covered the hazard of
dust accumulations since 2005.
Management of Change . The belt conveyor was enclosed without conducting a
Management of Change (MOC) review. The lack of hazard awareness, ignoring of near misses,
and lack of an MOC review led to the creation of an unprotected enclosure containing
combustible dust clouds. An MOC review, pe rformed by competent people knowledgeable
about dust explosion hazards, would have eval uated the need for explosion protection such
as venting, suppression, or inerting within such an enclosure.
Conduct of Operations. Written housekeeping programs were not effectively
implemented. What cleaning was done did not always include elevated surfaces. Dust
collection system design and maintenance may al so have been contributing factors to the
fugitive emissions, but no action was taken to reduce leaks or fix the fugitive dust collection
system. Also, there had been many small fires in this and other Imperial Sugar locations, which
did not lead to larger fires or explosions. These may have caused the staff to become
complacent regarding the hazards of combus tible dust. This phenomenon is known as
Normalization of Deviance, in this case, thinking that having many small fires was normal and
tolerable.
Incident Investigation . It has already been mentioned that this facility, and other
Imperial Sugar refineries had many small fires and near misses. For example, in this facility
operators noted that buckets in the bucket elevators sometimes broke loose and fell to the
bottom of the elevator. In one case this started a fire. An explosion in a dust collector occurred
10 days before this incident. These near misse s and the explosion were warning signs that
were not heeded.
Introduction to Fires
Fire is a chemical reaction; it is an oxidation reaction. Fire, however, is a rapid, exothermic
oxidation reaction. It generates heat and light (a n exception is a hydrogen fire as well as low
light emittance in methanol and carbon disulfid e fires) and produces smoke as a product of
incomplete combustion. Fire requ ires three things to occur:
Fuel,
Oxygen, and
Ignition source
|
RISK MITIGATION 343
Figure 15.5. Terminology describing layers of protection
Swiss Cheese Model
James Reason (1990 and 1997) developed the Swi ss cheese model which uses layers of Swiss
cheese to represent layers of protection. The laye rs of protection can protect the hazard being
realized and the consequence from occurring. The holes in the Swiss cheese indicate that these
layers may have weaknesses and degrade over time and thus are not 100% effective. When
the holes in the Swiss cheese align, represen ting each layer being compromised, then the
consequence can occur. A Swiss cheese model is shown in Figure 15.6.
Figure 15.6. Swiss cheese model
|
Appendices 175
APPENDIX B PHA QUALITY AN D COMPLETENESS CHECKLIST*
OBJECTIVE: To evaluate the prior PHA against quality and completeness criteria
established by company and regulatory requirements. A “No” response to any
item requires that the issue be ad equately addressed during the PHA
revalidation.
The columns in this example checklist are generic Q, T, and E.
• Q - Question column. Typically, questions are organized by topic.
The writers of this book expect that when used in industry, these
questions can be modified and upda ted by facilities to meet their
needs and unique situations. For exam ple, the topics listed in these
checklists could be used (with or without the full listing of
questions) to help a team consid er the quality of a prior PHA.
• T - Team evaluation column. The team should enter their response
to the question (e.g., “Yes,” “No,” or “Not Applicable”), followed by
brief documentation of the discussions and justification of the
response. While it is generally acceptable for occasional responses
to contain a simple “Yes” or “No,” more responses should contain
some detail of the team discussions, justifications, and concerns (if
any). If safeguards or controls protect against the topic of the
question, those should also be liste d in this, or a separate, column.
• E - Evidence of compliance/Team comments column. Brief
documentation of the discussions and evidence supporting the
team evaluation. Even if the evidence seems obvious, it is generally
better to document some deta il of the team discussions,
justifications, and concerns (if any).
* This checklist is provided for illustrati ve purposes only and is specific to
United States regulations. Readers may wish to develop such a checklist
specific to their own situation, requirements, and needs. |
128 Human Factors Handbook
11.3 Step 1: Identify safety critical tasks
Safety critical tasks should be identified, followed by
other tasks required to complete activities. The final
output of this phase consists of a list of safety critical
activities, and their competency standards, as shown in Appendix C. As noted in Chapter 6, the following
activities can help to identify safety critical tasks:
• Safety Critical Task Analysis (SCTA).
• Difficulty, Importance and Frequency Analysis.
• Identification of safety critical tasks noted in:
o Operation and Maintenance procedures.
o Risk assessments.
o Job/Task analysis.
o Existing lists of safety critical roles and tasks.
The assessment should include normal
process operations, process upsets, planned
and unplanned maintenance, and infrequent activities such as start-up and shut down.
Figure 11-1 links the level of safety cr iticality to the level of training and
competency assurance. Each task can be rated against:
• Task criticality.
• Task complexity.
• Task frequency.
• Time available to complete the task.
Learning needs and their requirements range from “Very High” to “Very Low”.
For example:
• “Very High” learning requirements correspond to:
o Experiential learning – on-the-job learning, instructions and assessment
o Extensive, multiple method of learning (e.g., on-the-job learning combined with mentoring and coaching)
o Assessments (e.g., in situ assessment, knowledge questions and observation of performance).
• “Very Low” requirements correspond to:
o Classroom learning/training – ba sed, largely on theoretical
knowledge with little or no assessment. See Chapter 1 for more
information on Safety
Critical Task Analysis
and Hazard
Identification and Risk
Analysis
Operators and supervisors are
often involved in this phase, as
they are knowledgeable of the tasks and associated risks. |
16.2 ENCOURAGING INVENTION WITHIN THE CHEMICAL AND
CHEMICAL ENGINEERING COMMUNITY
Publicizing the virtues of i nherent safety beyond the process safety
community and into the broader chemistry and chemical engineering
community is essential to spur inno vation. The authors of this book
encourage readers to look for oppo rtunities to “beat the drum” for
inherent safety at every opportunity as they interact with this broader
audience. Awareness can be raised in a variety of ways. Books by Trevor
Kletz (Ref 16.12 Kletz 1984; Ref 16.13 Kletz 1991), CCPS (Ref 16.1 CCPS
1993), and Englund (Ref 16.7 Englund) have proven to be successful vehicles. The IChemE and IPSG Inherently Safer Processing training program (Ref 16.10 IChemE) is anothe r successful format for promoting
this topic. Other means need to be explored and exploited.
16.3 INCLUDING INHERENT SAFE TY INTO THE EDUCATION OF
CHEMISTS AND CHEMICAL ENGINEERS
Teaching inherently safe r design concepts in undergraduate chemistry,
chemical engineering, and related disciplines will provide great benefit
as students move into industry after graduation.
16.4 DEVELOPING INHERENTLY SAFER DESIGN DATABASES AND
LIBRARIES
The industry, and more importantly, individual companies, need to
develop inherently safer design data bases that are readily available,
cataloged, cross-referenced, indexed, and shared across the broader
company. These might take the form of libraries of information. Several
examples of needed databases are:
•A continually updated database that describes and catalogs
inherently safer design successes and failures.
•A collection of databases of chemic als, and of functional groups,
ranked relative to their reac tivity, stability, toxicity, and
flammability categories. This would assist in the evaluation of the
potential benefits of substituting one, somewhat safer, chemical
for another. 434 |
EDUCATION FOR MANAGING ABNORMAL SITUATIONS 93
indicators, and marking of the instrument range on field gauges will help
the field operator in their troubl eshooting and mitigating actions.
Additionally, the instrument design engineers and process engineers must
also ensure that the ranges of the in struments cover all potential operating
modes and scenarios.
Further tools and techniques on HMI design are discussed in
Chapter 5, Section 5.7.1.
If a process can be designed to be inherently safe, then this is always
the preferred option, as discu ssed in Example Incident 4.2.
Example Incident 4.2 – Cher nobyl Disaster, April 1986
One feature of the design of the RBMK nuclear reactor at Chernobyl is
that it was primarily graphite-moderat ed and cooled by water. However,
water is also a neutron moderator and therefore when allowed to boil
within the reactor, the void created by the steam results in a reduction in
moderation. This creates a thermal feedback loop where more power
creates more boiling and less moderation, which in turn leads to more
power. The condition is called “positiv e void coefficient” and the reactor
design had the highest positive vo id coefficient of any commercial
nuclear reactor.
Once the water in the core started to boil beyond a certain rate, the
operators could do very little to control the reaction.
Many other factors were associated with this incident, although a
major design feature of the nuclear reactors in other industrialized
countries is that they are designed with negative void coefficients, which
makes them inherently safe from this type of runaway situation.
|
Toxic Hazards
Learning Objectives
The learning objectives of this chapter:
Explain chemical toxicity hazards,
Identify the pathways for toxi cs to enter the human body,
Explain exposure limits, and
Identify where to find resource s for chemical toxicity data.
Incident: Methyl Isocyanate Release Bhopal, India, 1984
Incident Summary
Just after midnight on December 3, 1984, a pesticide plant in Bhopal, India released
approximately 40 metric tons of methyl isocyanate (MIC) into the atmosphere. The incident
was a catastrophe; the exact numbers are in di spute; however, lower range estimates suggest
at least 3,000 fatalities, and injuries estimates ranging from tens to hundreds of thousands.
The impacted area is shown in Figure 6.1. The event occurred when water contaminated a
storage tank of MIC which resulted in the release of a large toxic cloud.
Key Points:
Process Safety Culture – Culture is about what you do when no one is
watching. When the culture degrades to the point that mechanical integrity
and resources are in disrepair, it may be time to stop the operation.
Hazard Identification and Risk Analysis – As Trevor Kletz said, “what you
don’t have, can’t leak”. (Kletz) Do you really need that chemical in that
quantity?
Management of Change – Multiple layers of protection only work if they
are functional. Removal of any layer sh ould be subject to management of
change oversight. |
ACKNOWLEDGMENTS
The American Institute of Chemical Engineers (AIChE) and the Center for Chemical Process
Safety (CCPS) express thei r appreciation and gratitude to all members of the Process Safety for
Engineers: An Introd uction, Second Edition and their CCPS member companies for their generous
support and technical contributions in the preparation of this book.
The collective industrial experience and know-how of the subcommittee members makes
this book especially valuable to all who strive to learn from incidents, take action to prevent
their recurrence and improve process safety performance.
Project Writer:
This manuscript was written by Cheryl Gro unds who thanks the Subcommittee Members and
Peer Reviewers for their content contribution to this book and their dedication to teaching
process safety. Final technical editing was completed by CCPS staff – Jennifer Bitz leading with
support from Bruce Vaughen and Anil Gokhale.
Subcommittee Members:
Jerry Forest Celanese, CCPS Project Chair
Cheryl Grounds CCPS Staff Consultant and Writer
Dan Crowl Michigan Technological University
Kobus Diedericks Nova Chemicals
Ken First CCPS Staff Consultant
Warren Greenfield WG Associates LLC
Barry Guillory Louisiana State University
Jack McCavit JL McCavit Consulting, LLC
Robin Pitblado DNV
Before publication, all CCPS books are subjecte d to a thorough peer review process. CCPS
gratefully acknowledges the thoughtful comments and suggestions of the peer reviewers.
Their work enhanced the accuracy and clarity of these guidelines. Although the peer reviewers
have provided many constructive comments and su ggestions, they were not asked to endorse
this book and were not shown the fi nal manuscript before its release.
Peer Reviewers:
Brian Farrell CCPS Staff Consultant
Jeff Fox CCPS Emeritus
Jerry Fung Canadian Natural Resources Limited
Jim Klein ABS Consulting
Ray Mentzer Purdue University
Hocine Ait Mohamed Rio Tinto
Greg Nesmith Dow Chemical Company
Bala Raman Ecolab
Mark Setterfield Tronox
Jonathan Slater 3M
Rajagopalan Srinivasa Indian In stitute of Technology Madras
Ron Unnerstall University of Virginia
Bruce Vaughen CCPS Staff Consultant
Ronald J. Willey Northeastern University |
1 Introduction
1.1 Introduction
This chapter discusses the scope of, th e audience for, and the benefits
for the readers of this guideline. Fo r readers unfamili ar with the CCPS
Risk Based Process Safety (RBPS) approach, this chapter also includes
a brief overview of its framework, including how lessons learned from
experience (one of the RBPS pill ars) are incorporated into each
chapter. The last section in this ch apter provides the reader with the
guideline’s framework: how the chap ters are organized based on the
types of operations at a facility (normal, abnormal, and emergency)
and how the risks associated with each transient operating mode—the
subject of this book—depends on which mode of operation the
process is undergoing at that time.
1.2 Scope
The scope of this guideline addresses process safety activities that are
essential for effectively managing the risks associated with the
different transient operating modes , recognizing that not all activities
will apply to every mode. Since the risk of incidents can be high during
the start-ups and shut-downs fo r normal operations in most
manufacturing facilities, this book presents incidents that occurred
during start-ups and shut-downs, providing insights as to why they
happened and guidance on how to minimize the risk in the future. The
important distinction between “transient operations” and the
“transient operating mode” should be understood. This guideline
defines the transient operating mo de in the context of normal,
abnormal, and emergency operations, providing a clear and 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 |
251
Separation technology
Requirements for additional information
During design scoping, the team will concentrate on minimizing
equipment, reducing inventories, si mplifying the process, reducing
wastes, and optimizing process cond itions. Inherent safety concepts
s h o u l d a l s o b e c o n s i d e r e d d u r i n g process hazards reviews, such as
HAZOP, for both new and existing pr ocesses. The initial design should
be “mistake-proofed,” and each safe ty device and procedure examined
to see if there is a way to eliminate the need for it.
When the inherent safety proces s has been expanded to review
regular or routine operation, the te am should look at all aspects of
inherent safety to provide suggested improvements for both the existing
facility and for the next plant. Even if the process was originally designed
with inherent safety in mind, th ese improvements may arise from
advances in technology, changes in prod uct specifications or application,
or lessons learned from incidents an d near-misses, both in the facility
being studied or in similar facilities elsewhere.
Table 10.2: Focus of Different Inherent Safety Reviews
Note: The number of check marks indicates the relative importance of the
strategy
Chemistry
and Process Selection Design Scoping Regular Operation
Minimize
• Reduce quantities √ √√√ √√
Substitute • Use safer materials √√√ √ √√
Moderate
• Use less hazardous
conditions √√ √√ √√ |
PEOPLE MANAGEMENT ASPECTS OF PROCESS SAFETY MANAGEMENT 437
Table 21.1 is an example of a listing of proc ess safety training course for new employees.
This is an abbreviated example; a full training matrix will likely include information such as
prerequisite course and whether the course is computer based or classroom training.
Table 21.1. Example process safety training course list
Course Target Audience Triggers
Understanding and
Managing
Flammable
Atmospheres Required for all Engineers,
Chemists, involved in design,
maintenance and operations First Two Years
PHA Methodology &
Team Leader
Training Recommended for technical people
involved in design, operations, and
safety reviews, including MOCs and
PHAs
Required for PHA Team Leaders First Two Years as well
as PHA Team Leader
Requirement
MOC Safety Review
Team Leader
Training Recommended for MOC Core Team
Members. Required for MOC Safety
Review Team Leaders that have not
taken the PHA Team Leader
Training Class First Two Years
Consequence
Assessment Recommended for people involved
in modeling releases of chemicals
and energy Prior to use of
consequence
modeling tools
Pressure Relief
Device (PRD)
Application Required for engineers and
recommended for designers
involved in PRD design, application,
sizing and selection Prior to involvement
in design, application,
sizing and selection of
PRDs.
Design and
Application of SCAI
and Safety
Instrumented
Systems Required for I&E, Control, and
Process Engineers and
recommended for designers
involved in shutdown system
design, review, and specification Required prior to
involvement in
Shutdown System
review, design or
operation OR
recommended within
the first two years
Fire Protection and
Fire Suppression Required for engineers and
recommended for designers
involved in fire protection systems Prior to involvement
in design of fire
suppression systems
Incident Investigation Recommended for incident
investigators and participants Prior to leading or
participating in
incident investigations
All employees, contractors and visitors are ty pically required to attend training on the
occupational safety and process safety basics at a facility. This is intended to prevent harm |
107
Figure 6.2 – The Eastman Chemical reactive distillation process for
methyl acetate US Patent No. US4435595 A (Ref 6.1 Agreda)
6.3 INHERENTLY ROBUST PROCESS EQUIPMENT
In many cases, it is possible to desi gn process equipment that is strong
enough to contain the maximum positi ve or negative pressure (i.e.,
maximum overpressure or maximum vacuum) resulting from the worst-
case process incident(s) (Ref 6.2 CCPS 1993). If such a design, under all
feasible circumstances, eliminat es the possibility of a loss of
containment due to overpressure /underpressure, then it can be
considered an inherently safer design . If not, it then only reduces the
likelihood of a release, and then it becomes a form of passive safeguard
design (although this may still be desirable). Containment of potential overpressure within the process vessel, or elimination of the possibility of vacuum collapse, simplifies the desi gn by eliminating elaborate active
|
A.3 Index of Publicly Evaluated Incidents | 197
are those that contributed most to the incident; the secondary findings
contributed less but still may have learning potential.
3. Find the titles of the reports in the listings in Section A.4. We do not
provide a web address foreach report because web addresses change
from time to time. However, searching the organization’s website should
take the reader quickly to the relevant report.
4. Read the reports and follow the remainder of the REAL Model as
described in Sections 6.2–6.8.
A.3 Index of Publicly Evaluated Incidents
Each of the 441 incidents indexed by the CCPS Learning from Investigated
Incidents Subcommittee has been assigned a code, consisting of a letter
followed by a one- to three-digit number. The letter refers to the collection of
incident reports:
A Agência Nacional do Petróleo, Gás Natural e Biocombustiveís of Brazil.
C The US Chemical Safety and Hazard Investigation Board (CSB).
D The Dutch Safety Board (DSB).
HA Alerts published by the Health Safety Executive (HSE) of the UK.
HB Bulletins published by the Health Safety Executive (HSE) of the UK.
J NPO Association for the Study of Failure (ASF) of Japan.
S Selected stand-alone incident reports.
The number refers to a unique report found in that collection.
This index is organized in four sections:
• Section 1. Codes for reports with potential findings related to most RBPS
elements.
• Section 2. Codes for reports related to most CCPS Culture Core Principles.
• Section 3. Codes for reports related to many causal factors.
• Section 4. A cross-reference to contents of the above sections from many
elements, core principles, and causal factors that were not directly
indexed.
Once codes that may be relevant to your effort have been identified, go to
Section A.4 to find the report title and how to obtain it.
|
Piping and Instrumentation Diagram Development
352
The type of orders by SIS could be start‐up and
sh
utdown of the electric motor.
●A manual command. This could be a command that
comes in from the operator in the field or control room.
“Command” signals are the orders that are sent to a
motor, or more correctly, to the MCC of a motor. C could be used as the representing letter for these types of
func
tions in P&ID symbols.
“Command” signals are always available around a
motor because they are the arrangement to make the plant and/or operator to control a motor.
Category 2 is the reports provided by the motor.
“Response signals” are the reports that are generated by the motor, or more correctly, by the MCC of a motor.
There are mainly two types of signals: the signals that
report if the motor satisfied a “command” and “responses” that report parameters on the “health” of the motor, which are running reports and trouble reports.
An example of command report signals is the signal is
sent from the motor if it turned off after receiving a com-mand for turning off or not.
An example of a running report signal is when a motor
reports the total hours that it is working. This is espe-cially important for motors connected to parallel pumps; they need to work roughly the same number of hours each to ensure their optimum health.
The other example of a trouble report signal is a “com-
mon trouble alarm. ” This signal is very common to see and it is an alarm by the motor that warns the operator of some type of problem inside the motor.
“Response signals” are the “motor’s talking” that are
sent by a motor, or more correctly, by the MCC of a motor. S could be used as the representing letter for these types of functions in P&ID symbols. A signal S could acti-vate an indicator, a lamp or an alarm on the control panel.
“Response signals” are not always available. A designer
may put them around a motor if it is critical to know the condition of the motor.
The main element of category 3 is HOA switch. The
principal arrangement for inspection and repair is started with an HOA switch. However, there are some other switches around an HOA switch that need to work together to be able to perform a complete inspection and/or maintenance.
In Sections 16.12.4–16.12.6 the P&ID representation of
three categories of electric motor functions are discussed.
16.12.4
P&ID Repr
esentation of Commands
and Responses
As it was stated there are two types of signals. The sig-
nals that are generated in “reaction” to a command and the signals to report the health of a motor. These two types of signals are shown in Table 16.7.There are at least two issuing regarding showing motor
control in P&IDs.
The first one is that there are plenty of parameters
involved that are not defined by the ISA. The solution is using non‐specific letters from ISA, like M, N, or Y and then explaining them right beside the balloon. Usage of Y, however, is very common because it refers to any event or state.
Several examples are shown in Table 16.8.The second issue is that each motor so many control
items around it on P&IDs that sometimes it is not easy Table 16.7 Symbols of c ommands and responses.
Meaning Representation
Commands
MDo it! Automatic:
Regulatory:
Interlock:
And in a combination form could be:
SD
Manual:
In field: HS
In control room: HS
or: XCR
115SS
command
Responses
MI did it sir!
Here is the proof!They could be in any of three
types of indicators, alarms, or lamps.
Indicator example:
YKQI
00
(I at the end of tag)
Alarm example: YA
00102FLT
(A at the end of tag)
Lamp example: YL
2500
STATUS
(L at the end of tag) |
38 GUIDELINES FOR MANAGING ABNORMAL SITUATIONS
Example Incident 3.3 – BP Texas City 2005 – ( cont.)
During the startup,the level indica ted by the gauge in the Raffinate
Splitter appeared to drop from 100% to 80% despite the feed rate
remaining at 20,000 bpd (132 m3/h) and a block valve remaining shut
on the bottoms line. However, the actual level in the tower was far
above the full-scale reading of th e level gauge and increasing.
The level device and its associated transmitter were designed to
measure the liquid level in a 5-foot (1.5 m) span such that 100% of its
calibration corresponds to approximat ely 10 feet 3 inches (3.1 m) in
a tower 164 feet (50 m) tall. Contra ry to several published reports
and technical papers, the displacer-ty pe level device on the Raffinate
Splitter worked as designed befo re, during, and even after the
incident, when it was tested by an independent third-party expert.
The apparent reduction in leve l was a result of the higher
temperature and therefore lower dens ity of the hydrocarbons within
the displacer level device, wh ich did not have temperature
compensation. As the bottoms temperature in the Raffinate Splitter
increased, the density fell, whic h was reflected by the apparent
reduction in indicated level from 100% to 80%. The displacer level
device no longer measured the leve l in the column. It was responding
instead to changes in the density of the fluid.
This confused the operator into believing that the level was not
abnormal (in range) when, in fact , the column was full. One of two
independent high-level alarms (the high-high alarm) was not
functional, the high alarm was not acted upon, per the norm during
startup, and a local sight glass wa s unreadable due to a buildup of
residue.
The subsequent explosion and fires led to 15 fatalities, 180 injuries
and financial losses exceeding $1.5 billion.
|
15. Fatigue and staffing levels 177
Whenever analyzing safety critical tasks, it is important to be realistic about the
time and effort needed to perform an activi ty. It is also important to recognize that
new problems may occur, and that more time and effort may be needed to
perform an activity than previously.
If the activity is complex or different, the activity can be analyzed. An example
is shown in Figure 15-9. The activity can be subdivided into sub-activities. These
can be plotted over time. Tasks that coin cide can be spotted. The time taken to
perform each task can be estimated, such as by observation of tasks or consulting
people who perform the tasks. In this example, five sub-activities coincide, and the
task time splits over two shifts. At least five people are required, and rest breaks
will be necessary.
Further guidance on workload and staffi ng needs analysis methods is available
from the Energy Institute [64].
Figure 15-9: A simple task timeline
|
Emergency Management
Learning Objectives
The learning objective of this chapter is:
Understand the importance of planning for and managing emergencies.
Incident: West Fertilizer Explosion , West, Texas, 2013
Incident Summary
On April 17, 2013, a fire occurred at the West Fertilizer Company (WFC) in West, Texas that
triggered an explosion of about 27 metric tonne (30 ton) fertilizer grade ammonium nitrate
(FGAN) at 7:51 PM. The explosion registered as a 2.1 on the Richter scale. (See Figure 20.1.)
Fifteen people were fatally injured, 12 of them were emergency responders, 3 members of the
public. One of the public fatalities was in a nur sing home (from a stress induced heart attack)
and the other two were in an apartment complex. An additional 260 people were injured. The
overpressure from the blast damaged 150 buildin gs offsite, including 4 schools, a nursing
home (later demolished), an apartment comp lex, and 350 private residences (142 beyond
repair) (CSB 2013).
This was a significant incident in the U.S., due to the extensive public impact, and the
prevalence of FGAN storage and handling facilitie s in the U.S. The CSB identified over 1,300
facilities handling ammonium nitrate (AN) within close proximity to a community, so the U.S.
President issued Executive Order EO-13650. This established a working group consisting of the
U.S. Department of Homeland Security, the U.S. Environmental Protection Agency, and the U.S.
Department of Labor (under which OSHA is lo cated), Justice, Agriculture and Transportation.
The purpose of the working group was to improve the identification and response to the risks
of chemical facilities (EO 2013).
Figure 20.1. Video stills of WFC fire and explosion
(CSB 2013)
|
E.37 Playing J eng® with Process Safety Culture |327
functions are carried out. What other culture factors could the
com mission have considered?
Did the fact that the operation involved a transfer from one
com pany to another create a “not my problem” attitude? The
com mission noted a lack of training in the procedure. What was
the general status of training in the facility? Were workers trained
to recognize and control hazards and risks? Did they take part in
“man-down” drills? What was the current focus of corporate
process safety efforts? Were employees empowered to fulfill their
safety responsibilities ?
Provide Strong Leadership, Maintain a Sense of Vulnerability,
Understand and Act Upon Hazards/Risks, Empower Individuals to
Successfully Fulfill their Safety Responsibilities, Combat the
Normalization of Deviance.
E.37 Playing Jenga® with Process Safety
Culture
Jenga® is a Parker Brothers strategy and skill
gam e. Players construct a tower of blocks, and
then take turns removing a block from the m iddle of the tower
and adding it to the top. The last to successfully remove a block
without toppling the tower is the winner.
A Vice President of Operations of a com pany, a long-time
employee well-steeped in the com pany safety culture, noticed
that process safety leading indicators and near-m iss metrics were
beginning to trend negatively across the com pany. While the
trend was not strong, the Vice President called a global meeting
of safety and operations leaders that all were required to attend.
The purpose of the meeting was to develop an action plan to
ensure the unfavorable trend did not continue and the company
could get back to its previous performance.
Not long afterward, the company began shifting the focus of
its business. Coincidentally, the Vice President of Operations B ased on
Actual
Situations |
58 PROCESS SAFETY IN UPSTREAM OIL & GAS
Figure 4-2. Two-barrier diagram for drilling, coring and tripping with a
shearable string
4.1.3 Drilling the Well: Fluid Column
The fluid column with sufficient hydrostatic pressure is one complete barrier on its
own. It is a mixture of fluids (water-based, non-water based, or gaseous) and solids
engineered to specific densities, collectively called “mud”, to match the
requirements of the pore pressure and fracture gradie nt curves (Figure 4-1). Lower
depths in the wellbore require higher density drilling mud and casing sections isolate
higher portions of the well where the mud pressure exceeds the fr acture gradient. In
offshore US federal waters, BSEE generally requires a drilling margin of 0.5 pound
per gallon (i.e., 0.5 ppg) below the lowest estimated fracture gradient to provide a
safety margin. As previously mentioned, drilling should not be thought of as a static
situation; conditions change requiring response to maintain the correct mud weight.
Mud returns from the well carrying drill cuttings have the soil/rock cuttings
separated using shakers to allow cleaning, reconditioning, and reuse of the mud.
Careful monitoring of mud flowrate is nece ssary to determine if there is mud loss
into the formation or an influx of reservoir fluids into the wellbore.
|