Microsoft Word - 1.docx
CHEMICAL ENGINEERING TRANSACTIONS
VOL. 77, 2019
A publication of
The Italian Association
of Chemical Engineering
Online at www.cetjournal.it
Guest Editors: Genserik Reniers, Bruno Fabiano
Copyright © 2019, AIDIC Servizi S.r.l.
ISBN 978-88-95608-74-7; ISSN 2283-9216
Complexity and Uncertainty Management in Process Safety
Education
Jaime E. Cadenaa,*, Felipe Muñozb
aSchool of Civil Engineering, The University of Queensland, St. Lucia, Queensland, Australia
bChemical Engineering Department, Engineering Faculty, Los Andes University, Colombia
je.cadena@uq.edu.au
At the end of the past century, the field of process safety was initiated thanks to the skills and industrial
experience of pioneers such as Trevor Kletz, Hans Pasman, Joaquim Casal and André Laurent. Pioneers
relied on their skills and desire to solve problems since formal safety education was not developed until the
late 1990s with examples such as the Mary Kay O'Connor Process Safety Center at the Texas A&M
University in the US. After 30 years of Kletz's first safety-related book, the panorama has changed with
engineering education integrating safety, thanks to initiatives such as ABET accreditation and others detailed
in recently published reviews. However, safety is still a concept being debated and constructed in both
education and practice, with open questions such as "How is safety linked to risk?" The authors view process
safety as a set of evolving tools and growing knowledge supporting risk assessment and aiding decision-
making. This process is carried out under uncertainty related to the complexity of the systems, the availability
of data and the competence of the analysts involved. Uncertainty and its management constitute critical
challenges for process safety educators and practitioners. In this context, the authors want to answer: how
does process safety education integrate complexity and uncertainty management? To answer it, the authors
conduct a review of formal educational programs and specific courses with a focus on uncertainty, as well as
their teaching and consulting experience. The results help to formulate a set of recommendations to improve
the handling of complexity and uncertainty management in different levels of safety education.
1. Introduction
Growing complexity and economical optimization constraints make the development of new hazardous
facilities a challenge for all engineering fields involved. At the same time, the deployment of new chemical and
petrochemical facilities continues to increase (Reniers, Amyotte 2012). The complexity and economic
constraints of these facilities can lead to failures and possibly to major accidents with unacceptable
consequences such as off-site casualties and environmental damage. Process safety engineers are in charge
of the risk management processes that prevents, controls and mitigates such events and their consequences.
Such responsibility requires competence from all engineers involved and particularly from the process safety
engineers, which is built through learning at all educational levels: undergraduate and postgraduate education
and continuous professional development. Mkpat et al. (Mkpat, Reniers, Cozzani, 2018) present a detailed
review of process safety education, with a particular emphasis on the way chemical engineering curricula
integrate process safety and the interactions between key stakeholders, i.e., academia, industry, government.
At the University education level, both Mkpat et al. (Mkpat, Reniers, Cozzani, 2018) and Dee et al. (Dee, Cox,
Ogle, 2015) identify those main topics being taught include mechanical integrity, hazards identification,
consequence, and emergency planning. These topics can be associated with the steps of the risk
management process as described by ISO 31000:2018, with an essential focus on the components of risk
assessment. Figure 1 presents the elements of risk management, and identifies those predominantly
addressed by practitioners in the industry and therefore with less learning components in academia.
DOI: 10.3303/CET1977073
Paper Received: 16 November 2018; Revised: 19 April 2019; Accepted: 20 June 2019
Please cite this article as: Cadena J., Munoz F., 2019, Complexity and Uncertainty Management in Process Safety education, Chemical
Engineering Transactions, 77, 433-438 DOI:10.3303/CET1977073
433
Figure 1: ISO 31000:2018 risk management process and its relation to education and practice
To ensure that University education provides students with the competency as process safety engineers,
accreditation institutions such as Accreditation Board for Engineering and Technology (ABET) establish a set
of outcomes in which safety is highlighted, with particular focus in hazards identification. ABET sets out 11
outcomes expected from a chemical engineering graduate, from which the (c) outcome relates to the design of
a process meeting realistic constraints including health, safety and environmental. Although safety can be
associated with all outcomes, this outcome is of particular relevance for educational purposes given that risk
management begins with process design, which is usually overseen by chemical engineers. At the same time,
risk assessments are initiated at early design stages and following through all the life-cycle of the operation,
requiring the use of different input information and models that support calculations and estimations, such as
gas dispersion modeling. Risk assessments results support the decision-making processes of the
stakeholders concerning safety measures in operations, including land-use planning. Bringing all these
elements together, risk assessment and the technical elements it involves constitute a central element of the
responsibilities of a process safety engineers and therefore, of their education. Recent research (Goertlandt et
al 2016; Rae 2012) has identified considerable uncertainty in the results of quantitative risk assessments
(QRA) which is explained when analyzing each of its steps. The first step is a process hazard analysis (PHA),
which allow identifying hazards and include a wide range of techniques, e.g., Hazards and Operability
Analysis (HAZOP). In these analyses, different scenarios are analyzed to identify potential consequences and
determined measures required to prevent, control and mitigate them. Researchers have presented that PHAs
can miss up to 47% accidents occurring in facilities (Goertlandt, Khazad, Reniers, 2016), given the qualitative
nature of these analyses and that, analysts do not know what they do not know. The second step is comprised
of consequence and probability (or frequency) analyses, using inputs related to the operating conditions and
characteristics of the facility. For consequence analysis of each accidental scenario, a wide range of models
are used which differ in complexity, precision, and accuracy, all requiring competent users and adequate
inputs. For probability analysis, different options exist including the use of qualitative descriptions, generic
failure frequencies (e.g. OREDA) for design stages, probability distribution functions and the use of random
sampling techniques such as Monte Carlo or (ideally) failure data from the facility.
The issues across each step of risk assessments are essential to explain the reported uncertainty and to
throw light on the role of process safety education in addressing these issues. As Mannan (Mannan, 1999)
stated about two decades ago, regulations and recommended practices are only an element in the solution to
the safety issues an operation can face. The other element is the education received by engineers and in
particular that these professionals understand the fundamental issues of implementing safety from the design
and throughout the life-cycle of the operation. In light of the technical and human advancements achieved in
the field of safety in these two decades and the improvement of chemical engineering curricula thanks to
accreditations such as ABET, this paper presents a detailed picture of how complexity and uncertainty in risk
assessments are addressed. The picture is presented in Section 2 and is based on a search for curricula and
specific case-studies, as well as in the experience of the authors. Finally, a set of conclusions along with
recommendations in Section 3, aiming at answering the following questions: 1) How can we better address
uncertainty and complexity in risk assessments?; 2) How can we better convey the importance of risk
assessment results in decision-making processes?; 3) What are the essential skills to be developed and what
decisions do they support?
Risk Assessment
Scope, context, criteria
Hazards identification
Risk Analysis
Risk Evaluation
Risk treatmentCo
m
m
un
ic
at
io
n
&
c
on
su
lta
ti
on
M
o
nitoring &
R
evision
Rec ording & reporting
Step predominantly addressed in industry
Step addressed by industry and academia
434
2. Complexity and uncertainty in process safety education
The starting point of this discussion is the fact that process safety is not part of all chemical engineering
curricula across the globe. In the United States of America alone, only 23% of chemical engineering
departments require Process Safety although this percentage is expected to grow as institutions adapt to
requirements of accreditation entities such as ABET (Voronov et al., 2017). Furthermore, in the institutions in
the US the credit average is 2.4 in institutions that directly address process safety, which is the lowest among
all categories of courses; in Europe a parallel can be seen as the European Federation of Chemical
Engineering (EFCE) recommends a 10% to 20% content in curricula, the lowest among all other categories.
This means that safety does not constitute an integral element of the knowledge transferred to chemical
engineering students in all education institutions, which in turn result in professionals with a gap in safety
knowledge. This is a critical element of this discussion, since knowledge gaps are directly associated with the
occurrence of incidents as pointed out by Krause (2016), and it is a reflection of epistemic uncertainty.
Examining undergraduate and Masters Curricula in Europe (Degreve, 2012; Brenig et al., 2013), it is
noticeable that complexity is introduced to students by gradually transitioning from the use of process safety
tools in simple systems to the considerations of its use in real systems. This is usually done through safety
experts' lectures and the visit to real plants. However, the application of process safety tools to a real system –
as most practitioners would recognize- requires considerable time and human resources that can often be
unavailable at an educational level. This implies that complexity is introduced through examples and there is
seldom explicit guidance to deal with a more complex (real) system. The challenge of recognizing complexity
begins at educational environments, having as an example the lack of awareness of the hazards and
associated risks with chemical laboratories operations. As presented by Olewski (2017), there is a
"misperception that university laboratories are ‘low risks' and ‘inherently safer," which can be extended to
small operating companies. This means that chemical engineers, process engineers and even educators in
these fields fail to recognize and adequately deal with complexity. In summary, most curricula do not explicitly
address a lack of safety knowledge, while the consequences incidents in university laboratories and small
operations reflect the consequences of this epistemic uncertainty. It is noted that the responsibility for this lack
of knowledge lays in both the educators and the students. For the latter, this lack of knowledge is an
"unknown," and therefore they cannot manage it unless made aware of it. Eliminating these unknown
unknowns is a task for more experienced and knowledgeable individuals, i.e., process safety educators, with
the help of an explicit recognition of complexity and its impact on risk management. The role of education at
University level and the need for its link with Industry is analyzed by Benitendi (2016), finding a need to
construct a collaboration scheme that allows integrating process safety into chemical engineering courses.
Constructing a compendium of courses such as those proposed by Benitendi might be unfeasible, given the
challenges it presents to the educational staff and the commitment required by the industry. However, this
proposal allows visualizing that the academia-industry link is essential to deal with the uncertainty inherent to
process safety considerations, e.g., variable and unpredictable hazardous scenarios. This uncertainty or
unknown unknowns appear again as a critical element to address by using industry's experience and the
complexity a real setting has when compared to a typical educational textbook problem. Such element is also
addressed by other educational initiatives for professionals, such as the one presented by Kennedy (Kennedy
et al. 2015) in which a course is constructed by academia to respond to the process safety knowledge gaps of
a particular operation. In this course, one of the specific objectives is to "enable staff to recognize and
challenge uncertainties". Although the course itself addresses the knowledge gap, the contents of the course
do not explicitly address the uncertainty involved in process safety studies and their impact on decision-
making.
Given that effectively supporting decision-making is the end goal of risk assessment, the role of complexity
and uncertainty is embedded in a broader context. This context is presented in the form of a flowchart in
Figure 1, where the key stakeholders involved in risk management converge. In particular, it can be seen the
crucial role of academia as it educates and trains professionals. Process safety education uses the previously
mentioned curricula and courses, but can easily omit the underlying uncertainty and complexity of operations,
which result in failures and unacceptable loss. Explicitly recognizing the role and importance of these two
elements is a challenge given the limited resources in academia; however, the authors believe this can be
gradually overcome.
435
Figure 1. Flowchart of Safety Science construction and its relation with key stakeholders
3. Conclusions and recommendations
This review briefly presents a picture of the current role of complexity and uncertainty in process safety
education, through selected examples of curricula and particular courses. This shows that complexity is a
fundamental element of process safety education, as of chemical engineering, depicted by the use of
simplified examples and then by more complex case studies.
Furthermore, it was found that examples exist in which the link between industry and academia allow to take
this case studies even further and provide students with real-life conditions and limitations, which has a direct
effect on their formation as competent safety engineers. The review also allowed constructing a flowchart for
knowledge in the form of Safety Science, as per the definition provided by Aven (2014). This picture shows the
complexity behind the proper education and training of safety engineers, requiring all stakeholders to be
directly or indirectly connected (e.g., society through the acceptability criteria defined by regulating bodies)
and the need for continuous feedback from failures. Based on this flowchart and the typical contents of the
reviewed curricula, the authors notice an absence of explicit reference to the uncertainty sources that can
considerably affect the results of a risk assessment and therefore the following decision-making process. The
omission of uncertainty in process safety education can be as potentially dangerous as the omission of
process safety contents from a chemical engineering education. Based on the own experience of the authors
in risk assessment and management projects with the government and the private industry, as well as on the
typical contents of process safety programs, a list of uncertainty sources was formulated. This list is used to
formulate a series of suggestions that should help educators to consider them. Here, the goal is to provide
students (regardless of the level of formation) with additional tools to understand process safety not as a
checklist of studies, but as an activity in which uncertainty and complexity are always present. By explicitly
recognizing uncertainty and providing the students with the currently available tools to deal with it, decision-
making in the chemical process industry can be enhanced. Taking the previous into consideration, to answer
the questions: How can we better address uncertainty and complexity in risk assessments? and How can we
better convey the importance of risk assessment results in decision-making processes?, the authors make use
of the typical contents in process safety education. These topics range from hazards characterization to risk
communication. Table 1 presents the sources of uncertainties and the associated suggestions for each one of
these topics, as well as examples of successful applications. Education institutions should consider this during
the construction of process safety programs and by educators for the design of syllabus.
436
Table 1: Uncertainty associated to process safety education
Topic Sources of uncertainty Suggestions
Hazards
characterization
Availability of thermodynamic and hazardous
properties
Availability of data for mixtures
Complexity of process conditions e.g.
multiphase systems
Emphasis during fundamental courses of
chemical engineering on the weaknesses
present in thermodynamic data generation, both
in equipment and interpretation
Hazards
identification
Completeness of operation information
Availability of accepted guidelines for
executing techniques
Multi-disciplinary nature of workshops
Participants expectations and interests
Participants’ personalities
Presence of unknown unknowns (e.g., failure
modes)
Impossibility of managing unknown unknowns
Run of workshops in which hazards
identification techniques are exemplified using
real-life case studies, such as those
investigated by the Chemical Safety Board of
the U.S. Noakes (Noakes et al. 2011) provides
an example of an innovative module for HAZOP
education, consistent with these suggestions.
Risk analysis Definition of risk The definition by Kaplan (1981) set the
foundation for QRA. This is not the only
definition available (Aven, 2009), and students
should be made aware of this.
Consequence
analysis
Completeness of thermodynamic data for
calculations for mixtures
Availability of modeling options and the
assumptions that support them
Degree of model validation
Technical competencies for model selection,
use and results interpretation
Sensitivity of results to model parameters and
assumptions
Run of simple scenarios using a wide range of
tools, e.g., use of the TNT equivalence model,
TNO multi-energy and Computational Fluid
Dynamics codes for comparing overpressure of
a vapor cloud explosion.
Probability of
failure
Available knowledge regarding failure modes
and dependency between them
Availability of data to select parameters for
the probabilistic distribution
Availability of data/frequencies for specific
system’s conditions
Sensitivity of results to model parameters and
assumptions
Given the importance that scientists,
practitioners and regulating bodies give to
probabilistic analysis, the limitations on defining
limit state functions and the use of probability
distribution functions should be provided.
Risk evaluation Nature of criteria: prescriptive, consequence-
based, individual risk, societal risk
Availability of guidance to use and interpret
acceptability criteria
Sensitivity of evaluation to risk analysis
parameters and assumptions
Use of publicly available quantitative risk
assessments to show the consequences in the
decision making of using internationally differing
criteria for individual and societal risk (Pitblado
et al. 2012)
Risk treatment Completeness of current safety barriers, e.g.,
BowTie
Cost-benefit limitations
Organization’s internal guidance and
accepted safety measures
Provide examples of successful implementation
of safety barriers, using case studies of
industry. In case this link is not available, use
examples of daily life to exemplify the selection
of safety measures under cost and time
constraints.
Risk
communication
Completeness of stakeholders’ identification
and characterization, including their
expectations and risk aversion
Availability of communication guidance that
supports the use of risk assessment
information
Invite regulating bodies and industry's
representatives that can present the importance
of the risk management decisions that are
taken, and how the role and competencies of
the safety engineer directly affect these.
Regarding the skills of a process safety engineer, Beard (2005) stated that a knowledgable user is one of the
three elements required for an acceptable use of a model in fire safety engineering; the other two being a
model and the methodology to use it. Beard defined a knowladgable user as that “Who is capable of
employing the methodology to a model which has the potential to be valuable in a particular case in a
comprehensive and explicit manner, and interpreting results justifiably”. This is a challenge, as Beard (1996)
had previously stated using an example for a temperature calculation, in which the associated errors with it
are the assumptions made, the numerical errors, software and hardware malfunctions, and the application
437
errors. The reality is not different even for a simple overpressure calculation using the TNT equivalence
model. Furthermore, the Center for Chemical Process Safety establishes competency as one of the 20
elements of the Risk Based Process Safety model. This is presented in CCPS (2010) as one of the key
elements for process safety commitment pillar, where a general framework for maintaining competency in the
industry is presented. However, this framework does no state the relevance of explicitly addressing
uncertainty in the course of process safety tasks.
To the question of what are the essential skills to be developed and what decisions do they support? The
authors believe that the capacity of safety engineers to recognize their limitations in knowledge and tools,
especially for complex systems, is key to achieve a continuously better risk management. Recognizing there
are unknown unknowns over which no feasible control exists besides permanent monitoring and an explicit
consideration of assumptions, is essential. This brief review and the considerations presented regarding
treatment of uncertainty in process safety education, an sets the foundations for further work.
References
Amyotte P., Reniers, G., 2012, Prevention in the chemical and process industries: Future directions, Journal of
Loss Prevention in the Process Industries, 25, 227-231
Aven T., Renn O., 2009, On risk defined as an even where the outcome is uncertain, Journal of Risk
Research, 12, 1, 1-11
Aven T., 2014, What is safety science?, Safety Science, 67, 15-20
Beard A., 2005, Requirements for acceptable model use, Fire Safety Journal, 40, 477–484
Beard A., 1996, Fire Models and Design, Fire Safety Journal, 28, 117-138
Benintendi R., 2016, The bridge link between university and industry: A key factor for achieving high
performance in process safety, Education for Chemical Engineers, 15, 23-32
Brenig, H., Haputmanns U., Klais O., Schmidt J., Moritz H., Schönbucher A., 2013, Recommendation for
education in a Bologna Three-Cycle Degree System, Model Curriculum: Process and Plant Safety,
DECHEMA, ISBN 978-3-89746-141-3
Center for Chemical Process Safety (CCPS), 2010, Guidelines for Risk Based Process Safety, John Wiley &
Sons, Chapter 5, 89-122
Dee S., Cox B., Ogle R., 2015, Process safety in the classroom: The current state of chemical engineering
programs at US universities, Process Safety Progress, 34, 316-319
Degreve J., Berghmans J., 2012, Master of science in safety engineering at KU Leuven, Belgium, Procedia
Engineering, 45, 276 – 280
Goerlandt, F., N. Khakzad and G. Reniers, 2016, Validity and validation of safety-related quantitative risk
analysis: A review, Safety Science, 127-139
Kaplan, S. and B. J. Garrick, 1981, On The Quantitative Definition of Risk, Risk Analysis, 1, 11-27
Kennedy P., Duffy N., Horgan S., Lynch M., 2015, Chemical Process Safety Education: Joint Industry-
Academic venture, HAZARDS 25, 1-8
Mannan S., Akgerman A., Anthony R., Darby R,, Eubank P., Hall K., 1999, Integrating process safety into ChE
education and research, Chemical Engineering Education, 33, 198-209
Mkpat E., Reniers G., Cozzani V., 2018, Process safety education: A literature review, Journal of Loss
Prevention in the Process Industries, 54, 18-27
Noakes N., Chow C., Ko E., McKay G., 2011, Safety education for chemical engineering students in Hong
Kong: Development of HAZOP Study teaching module, Education for Chemical Engineers, 6, e31-e55
Olewski T., Snakard M., 2017, Challenges in applying process safety management at university laboratories,
Journal of Loss Prevention in the Process Industries, 49, 209-214
Pitblado R., M. Bardy, P. Nalpanis, P. Crossthwaite, K. Molazemi, M. Bekaert, V. Raghunathan, 2012,
International comparison on the application of societal risk criteria, Process Safety Progress 31(4): 363-
368
Rae A., Alexander R., 2012, The Science and Superstition of Quantitative Risk Assessment, Proceedings of
PSAM 11 & ESREL 2012, International Association of Probabilistic Safety Assessment and Management,
IAPSAM. 3: 2292-2301.
Voronov R.S., Basuray S., Obuskovic G., Simon L., Barat R.B., Bilgili E., 2017, Statistical analysis of
undergraduate chemical engineering curricula of United States of America universities: Trends and
observations, Education for Chemical Engineers, 20, 1-10.
438