Microsoft Word - 476hernandez.docx
CHEMICAL ENGINEERING TRANSACTIONS
VOL. 43, 2015
A publication of
The Italian Association
of Chemical Engineering
Online at www.aidic.it/cet
Chief Editors: Sauro Pierucci, Jiří J. Klemeš
Copyright © 2015, AIDIC Servizi S.r.l.,
ISBN 978-88-95608-34-1; ISSN 2283-9216
Influence of HOT Factors on Risk Assessment Level
Based on Fuzzy Set Theory
Marko Djapan*a,b, Ivan Macuzicb, Danijela Tadicb, Branislav Jeremicb
aPolitecnico di Torino, Department of Applied Science and Technology, Corso Duca degli Abruzzi 24, 10129 Torino, Italy
bUniversity of Kragujevac, Faculty of Engineering, Sestre Janjic 6, 34000 Kragujevac, Serbia
marko.djapan@polito.it
Introducing human, organizational and technical/technological factors in a risk assessment model represents a
significant step forward in the occupational health and safety field. The best known problems of risk
assessment models are their inability to cope with the uncertainties, subjectivity and static nature of the
process. The improvements in the risk assessment model presented relate to these problems. Three groups
of factors are introduced and their hierarchical structures are defined. The relative importance of the factors
and sub-factors and values of the sub-factors are described by predefined linguistic expressions. Fuzzy set
theory is used for the modeling of existing uncertainties. The risk assessment model is based on quantitative
risk assessment and reducing the risk level is possible through changing the initial values of the sub-factors
identified. Varying the values allows the reduction of the risk level between two regular risk assessment
processes or permits planning of proactive actions and activities in order to eliminate hazards or reduce risk
level in the period considered. The proposed model is tested with real data from one production company.
1. Introduction
Nowadays, risk management and risk assessment have become very popular and much used terms.
However, it is worth bearing in mind that the very use of these terms to solve all problems related to risks in
business and production systems can lead to the possibility of improper and wrong focus on the problems.
Additionally, the major misconception in the occupational health and safety (OHS) field is that the
implementation of the risk assessment process is just one great duty, the limiting factor in the functioning of
the system and another reason for spending financial resources. Nevertheless, considering all the objectives
into account it is stated that reducing the number of accidents is one of the most important goals of reducing
the number of unwanted and unplanned events in the workplace (Sawacha et al., 1999). Whether it is about
companies that provide services or which are based on the production of certain types of products, the goals
need to be defined (relating to production and safety) within acceptable risk levels. This should imply an
important balance between production and the concept of safety. The risks themselves are very complex. The
complexity is reflected during the risk assessment process in which risks are identified, analyzed and
evaluated for the whole system, not individually. According to Bischoff (2008) risks are within normal limits or
deemed acceptable if they meet certain conditions, like low uncertainty related to the probability of the
occurrence of consequences; fairly low overall probability of injury; low or intermediate probability of the
hazard; low consistency; impossibility of occurrence of repeated unwanted and unplanned events; small
deviation between the assumed potential harm and the probability of occurrence and low level of risk that is
related to social anxiety and potential dissatisfaction. This paper continues work of Djapan et al. (2013) and
proposes a model as a supplement of the existing risk assessment process or as a model to decrease the
identified risks between two risk assessments. This model is based on the identification of HOT (human,
organizational and technical/technological) factors, its implementation and role in risk assessment based on
fuzzy set theory. A management team describes each uncertainty in the relative importance of the identified
factors and sub-factors, the sub-factor values and the risk level potential. They do this using linguistic
DOI: 10.3303/CET1543206
Please cite this article as: Djapan M., Macuzic I., Tadic D., Jeremic B., 2015, Influence of hot factors on risk assessment level based on fuzzy
set theory, Chemical Engineering Transactions, 43, 1231-1236 DOI: 10.3303/CET1543206
1231
expressions, which reflects the human way of thinking. Fuzzy set theory resembles human reasoning in its
use of approximate information and uncertainty to generate decisions. Hence, applying the fuzzy approach all
uncertainties and imprecisions, which emerge due to lack of good evidence, are managed. These linguistic
expressions are modeled by triangular fuzzy numbers (Zimmermann, 2001). The proposed model, bearing in
mind the whole concept of health and safety at work and the risk assessment process, can be used to solve
the following problems in a production system: 1) prioritization of factors that affect workplace safety, whereby
a factor that is rank first has the most impact on it, 2) determining how changes in one or more sub-factors
affect the risk level and 3) workplace risk assessment.
2. Literature review
Considering the fact that risks appear in all spheres of life, the term risk is widely used and explained in
different ways. Aven (2011) and Leitch (2010) consider that there is a terminological inconsistency within the
standards and concepts from different areas where risk is considered, covering the economy/industry,
management structure and financial resources. They state that there is a very vague and inconclusive
meaning and understanding of the fundamental terms - risk and probability. What significantly complicates risk
assessment is how to choose the right tools and methods for conducting this process. Making this choice is
especially difficult taking into account that the wrong choice leads to bad decisions, and later to reduced
confidence in the whole risk assessment process (Lootsma, 1997). Risk assessment is burdened with
subjectivity and there is a tendency for certain risks not to be taken into consideration. The situation becomes
even worse if some risks are underestimated. This leads to the conclusion that subjective risk assessment is
potentially dangerous and can lead to significant unwanted consequences (Reid, 1992). If the findings of the
probabilities and consequences of accidents are not at a high and acceptable level, the risk assessment can
be classified as irrational, without unscientific character and with the high potential that it will not be conducted
in an appropriate manner (Stirling, 1998). Risk identification, assessment and management should not only be
based on an analysis of the technical risks leading to unidimensionality of the risk assessment, neglecting
other aspects of business and manufacturing systems (Renn, 1999), such as human and organisational
factors (as in Demichela et al., 2014; Monferini et al. 2013) and maintenance management, where the above
cited aspects are mostly important (e.g. Todorović et al., 2014). Analogous to the overview of the tools and
methods for risk assessment in standard ISO/IEC 31010 (ISO, 2009), in the contemporary scientific literature,
there are different opinions on the general division of tools and methods for risk assessment. According to
Arunraj and Maiti (2007) and Tixier et al. (2002) there are two main groups of tools and methods, classified
according to the nature of use, while according to Marhavilas (2011) there are three groups: a) quantitative, b)
qualitative and c) hybrid. The research in the past decade in the occupational health and safety field has
concluded that the percentage of the most common group of tools and methods for risk assessment is a
quantitative method, app. 65,63% Marhavilas (2011). These data informally accept the quantitative method for
the risk assessment process as a the most suitable, as also discussed in Demichela & Camuncoli (2014). The
authors of this paper’s improved model relies on quantitative risk assessment (McDonald, 2004) and makes a
basis for implementing identified HOT factors using fuzzy set theory. Confirmation of the necessity of making
such a structure and this type of classification arises from industry, which was at a particular moment of
history considered as the riskiest field. The history of aviation safety can be divided into three eras that
progressively include three unavoidable factors (ICAO, 2012). In this way, this principle and classification has
been adopted and applied to other industries. The authors are of the opinion that this approach has a real
scientific basis for the improvement of the risk assessment model and its implementation in other industries.
3. Framework of the model
Since, in practice, none of workplaces has the same characteristics and each is specific in its own way, and
risk assessment process should be viewed as separate and unique problem. According to this assumption, in
this paper, we set limits on defined factors, while the total number of sub-factors could vary depending on the
observed workplace. The starting point of the proposed model is workplace risk assessment. This type of risk
assessment (McDonald, 2004) belongs to the group of quantitative methods and tools. In addition, this risk
assessment method is the maximally applicable in practice and the results obtained by using this method
proved to be acceptable and relevant. This type of form covers all essential elements that have an impact on
risk level. It includes four elements: the probability of injury, the severity of injury (because of manifestation of
dangerous situations), the frequency of exposure to hazards and the number of people exposed to the
identified hazards. The focus of the proposed model is on the probability of injury. The expert team is
consisted of licensed OSH persons, professors from universities, OSH engineers, members of OSH council on
1232
national level and engineers who work in observed types of companies. All factors are classified into different
groups by expert team based on data from the literature and the results of good practice. Groups of factors
are formally represent set of indices I={1,...,i,...,I}, where i represents the index for the group of factors, and I is
the total number of defined groups of factors. Sub-factor being represented by a set of indices J={1,...,j,...Ji}.
The total number of sub-factors defined by the factor i, i=1,..,I is referred to as Yi. Index which indicates the
sub-factor, any factor is j, j=1,...,J. It should be noted that, in general, the number of sub-factors might be
different. In our specific case, since the research is about a group of similar production companies, the
number of factors is three and each factor contains five sub-factors (Table 1).
Table 1: HOT factors
Human factors Organizational factors Technical/technological factors
Personal
characteristics (HF1)
Work pace (OF1)
Technical characteristics of
equipment (TTF1)
Experience (HF2)
Organizational and schedule of work tasks
(OF2)
Level of automatization (TTF2)
Training level (HF3)
Informations, procedures and
documentation (OF3)
Characteristics of safety equipment and
devices (TTF3)
Behavior (HF4) Workplace ergonomy (OF4) Maintenance level of equipment (TTF4)
Relations (HF5) OHS system (OF5)
Characteristics of personal protective
equipment (TTF5)
Identifying and assessing the risks of the workplace is one of the most important tasks of management at the
level of each company. This paper considers a group of similar companies. Based on these assumptions, we
can consider that the relative importance of the factors and the relative importance of sub-factors were
determined at the level of a given group of companies. The expert team defines and determines the relative
importance of the factors and sub-factors and it is formally presented a set of indices decision makers
E={1,...,e,...E}, where e represents the index for decision-makers, and E is the total number of members of the
expert team. Each member of the management team estimates the relative importance of the factors. They
expressed their estimates using five predefined linguistic expressions. Modeling of linguistic expressions is
based on the theory of fuzzy sets (Zimmermann, 2001). In other words, determining the relative importance is
set as a problem of group decision making. Aggregating individual assessment in group consensus is
obtained by applying the method of fuzzy mean value. Weight of factors wi, i=1,..,I and weight of sub-factors,
wji, j=1,..,Ji; i=1,..,I are obtained from fuzzy analytic hierarchy process (FAHP) that was developed in (Chang,
1996).
4. Modeling uncertainties
Workplace risk assessment depends on the relative importance of the factors, sub-factors and their current
values. It can be assumed that it is closer to the human way of thinking if the management team’s evaluations,
attitudes, knowledge and experience are expressed by linguistic expressions. In this section, the relative
importance of the factors, sub-factors and their values are described with predefined linguistic expressions.
The number and type of the linguistic expressions defines the management team. In this paper, the modeling
of linguistic expressions is based on the fuzzy sets theory, which is a useful tool to handle imprecision,
vagueness and randomness. It can be considered that the fuzzy set theory supports the human way of
thinking because it uses approximate information and uncertainty to generate decisions (Kahraman, 2009). In
this paper it is assumed that the relative importance of the factors and sub-factors is equal. The values of the
relative importance do not change over time. Each member of the expert team assesses the relative
importance of the defined factors and sub-factors. Assessing the relative importance aims to establish a
practical basis for the introduction of improved risk assessment models.
4.1 Modeling of relative importance of factors and sub-factors
The members of the expert team used five predefined linguistic expressions for factors that are modeled by
triangular fuzzy numbers and seven predefined linguistic expressions for sub-factors that are modeled by
triangular fuzzy numbers. These triangular fuzzy numbers for factors and sub-factors are presented in Table 2.
Relative importance of factors i related to factor i’, i=1,...,I', i≠i’ is modeled using triangular fuzzy number = ( ; , , ). Upper and lower bounds of these fuzzy numbers are represent as and , while
represents modal value. Domain of these triangular fuzzy numbers is in interval [1-5].
1233
Table 2: Relative importance of factors and sub-factors modeled by triangular fuzzy numbers
Factors Sub-factors
very low importance – = ( ;1,1,3.5) very low importance – = ( ;1,1,2.5)
low importance – = ( ;1,2.5,4) low importance – = ( ;1,2,3)
medium importance – = ( ;1,3,5) medium low importance – = ( ;2,3.5,5)
high importance - = ( ;2,3.5,5) medium importance – = ( ;3.5,5,6.5)
extremely importance - = ( ;2.5,5,5) medium high importance - = ( ;5,6.5,8)
high importance - = ( ;7,8,9)
extremely importance - = ( ;7.5,9,9)
Value 1 means that factor i has almost equal importance and value 5 has extreme importance, regarding to
factor i'. Relative importance of sub-factors j, related to sub-factor j', j=1,…,Ji, j≠j’ is modeled using triangular
fuzzy number ′ = ( ; ′ , ′ , ′ ). Upper and lower bounds of these fuzzy numbers are represent as ′
and ′ , while ′ represents modal value. Domain of these triangular fuzzy numbers is in interval [1-7]. Value 1
means that sub-factor j has almost equal importance and value 7 has extreme importance, regarding to sub-factor
j'.
4.2 Modeling of factors’ values
The management team consists of a person responsible for health and safety at work, managers of all the
sectors observed and external experts (if management considers them necessary). The management team
uses the five predefined linguistic expressions for their assessment which are modeled by triangular fuzzy
numbers = ( ; , , ), where and are lower and upper boundaries respectively, and is modal
value. These triangular fuzzy numbers are defined in Table 3. The domain of these triangular fuzzy numbers is
in interval [0-1]. Value 0 means that the value of the factor is negligibly low and value 1 represents the
extreme value of the factor.
Table 3: Relative importance of factor values
Factor values
very low importance – = ( ;0,0,0.2)
low importance – = ( ;0.15,0.3,0.45)
medium importance – = ( ;0.35,0.5,0.65)
high importance - = ( ;0.55,0.7,0.85)
extreme importance - = ( ;0.8,1,1)
5. Developed algorithm
The probability of injury is determined for each workplace respecting the importance of the factors, sub-factors
and the importance of the current value of the sub-factors. The value of the probability of injury is described by
triangular fuzzy number . The determination of this value is presented as equation Eq(1).
= ∙ = 1,…, ; = 1,…, ; = (1)
Prioritization of the factor is determined based on calculation of the probabilities of injury. The factor which is
associated the greatest possibility of injury has the greatest impact on workplace safety, and vice versa.
Figure 1. Developed algorithm
1234
Based on the measure of belief calculated this factor can be ranked in first place and the management team
gets the necessary input data to take appropriate actions. These actions should lead to a reduction in the
value of the probability of injury. Consequently, decreasing this element, the overall risk level is reduced. The
developed algorithm of the proposed model is presented in Figure 1.
6. Case study
For the purposes of this research and paper we used the most general list of possible risks which can be
found in a workplace. The list consists of 25 risks (EASHW, 2007). The case study was carried out for one
SME in the production sector. The characteristics of this company are production processes in operation for
less than a year; a significant number of new employees; the current workload requiring only one shift; the
prevailing of manual operation; internal transportation; the use of improvised hand tools; the use of the
simplest form of gloves as personal protective equipment; the workplace location in the same bigger space
where the rest of the equipment is located with no physical isolation from the influence of microclimate, such
as noise, vibration, dust and temperature. The authors will show what type of input data needs to be available
for this model, the improvement in the percentage of sub-factors according to the financial status of the
company and the determined output. The case study is presented in Table 4.
Table 4: Current value (as a percentage) of sub-factors and sub-factor improvement (as a percentage)
Sub-factors Current Value Planned improvements
HF1 70 % -
HF2 20 % 10 %
HF3 20 % 20 %
HF4 50 % 10 %
HF5 60 % -
OF1 30 % 60 %
OF2 40 % -
OF3 25 % 15 %
OF4 10 % 25 %
OF5 30 % -
TTF1 20% 20 %
TTF2 10 % -
TTF3 20 % -
TTF4 50 % -
TTF5 30 % 20 %
Taking into account the projection of the production increase, that the work is mostly performed manually and
the costs/benefit analysis conducted, three risks R7, R14 and R22 were taken into consideration. The
improvement of these risks needed to be implemented within 12 months. The improvements were possible
because the sub-factors selected did not require capital financial resources. The human sub-factors changed
continuously. The improvements were mostly based on intensive training and individual work with employees.
The improvement of the organizational sub-factors was based on introducing the safest way to work and
creating procedures for a certain number of workplaces. The technical/technological sub-factors depended on
the financial investments and needed to be planned according to the financial state of the company. The
results of the quantitative risk levels and new determined risk levels after implementation of the planned
improvements of the sub-factors using the proposed model are shown in Table 5.
Table 5: Current value (as a percentage) of sub-factors and sub-factor improvement (as a percentage)
Risk Risk level New risk level Percentage
R7 175 153 -13 %
R14 150 136 -9.5 %
R22 200 179 -10.5 %
7. Conclusions
The modification of the value of one or more sub-factors leads to changes in risk level. The risk assessment
model developed is not intended to replace the existing risk assessment tools, methods and techniques. The
main contribution of the paper is to help OHS managers to reduce risks as much as possible. Another
contribution is to proactively plan and prepare activities for OHS improvement. Based on the future financial
1235
and organizational status of their company, OHS managers could make relevant calculations in advance
about which area can be covered and which activities are the most cost-effective.
The general limitations of the model are that a significant increase in some sub-factors does not reduce the risk
to an extent that would be expected and the sub-factors have an influence on the group of similar risks, not on
specific hazards. The first limitation could be solved by introducing an enhanced way to assess relative
importance, because each sub-factor has a certain minimal impact which does not have to be the case in
practice. The second limitation could be solved by making minor model changes, providing a choice for the risk
assessment process for specific hazards. The main advantages of the model presented are an easy to use
methodology and that the ranking of sub-factors and level of risk can be obtained in an exact way. The proposed
method is flexible with regard to the changing in the number of factors and sub-factors and the changes in their
relative importance. Future work on this issue will be to design an appropriate and user-friendly interface for
better and faster calculations. Moreover, an additional improvement for determining the best package of actions
and measures for improving the chosen sub-factors will be considered.
Acknowledgements
This work was supported by the EU FP7 Marie Curie Actions Initial Training Networks [grant number 289837].
References
Arunraj N., Maiti J., 2007, Risk-based maintenance: techniques and applications, J. Hazard. Mater., 142, 653–661.
Aven T., 2011, On the new ISO guide on risk management terminology, Reliab. Eng. Syst. Safe, 96, 719-726.
Bischoff H.J., 2008, Risk Profiles in Modern Work Life, Risks in Modern Society – Topics in Safety, Risk,
Reliability and Quality, Eds. Bischoff H.J., Springer, Dordrecht, the Netherlands.
Chang D.Y., 1996, Applications of the extent analysis method on fuzzy AHP, Eur. J. Oper. Res., 95, 649-655.
Demichela M., Camuncoli G., 2014. Risk based decision making. Discussion on two methodological
milestones. Journal of Loss Prevention in the Process Industries, 28 (1), 101-108.
Demichela M., Pirani R., Leva M.C. 2014. Human factor analysis embedded in risk assessment of industrial
machines: Effects on the safety integrity level International Journal of Performability Engineering, 10 (5),
487-496.
EASHW (European Agency for Safety and Health at Work). 2007, Risk assessment essentials,
https://osha.europa.eu/en/publications/promotional_material/rat2007 (accessed 11 November 2014).
ICAO (International Civil Aviation Organization), 2012, Safety Management Manual (SMM) Doc 9859 AN/474
– 3rd ed. Montréal, Canada.
ISO (International Organization for Standardization), 2009, ISO/IEC 31010 International Standard, Risk
management – Risk assessment techniques.
Djapan M., Tadic D., Macuzic I., Jeremic B., Giagloglou E., 2013, A new model for evaluation of safety grade
of indicators based on a fuzzy logic, Chemical Engineering Transactions, 33, 463-468.
Kahraman C., 2009, Introduction: Fuzzy theory and technology, J. Mult. Valued. Log. S., 15, 103-105.
Leitch M., 2010, ISO 31000:2009 - The New International Standard on Risk Management, Risk Anal, 30, 887-892.
Lootsma F.A., 1997, Fuzzy Logic for Planning and Decision-making, Kluwer Academic, Boston, MA, USA.
Marhavilas P.K., Koulouriotis D., Gemeni V., 2011, Risk analysis and assessment methodologies in the work
sites: On a review, classification and comparative study of the scientific literature of the period 2000 –
2009, Ј. Loss Prevent. Proc., 24, 477–523,
McDonald D., 2004, Practical Machinery Safety, Elsevier, UK.
Monferini A., Konstandinidou M., Nivolianitou Z., Weber S., Kontogiannis T., Kafka P., Kay A.M., Leva M.C.,
Demichela M. 2013. A compound methodology to assess the impact of human and organizational factors
impact on the risk level of hazardous industrial plants. Reliability Engineering and System Safety, 119, 280-289.
Reid S.G., 1992, Acceptable risk, 1992, Engineering Safety, Ed. Blockley D.I., New York: McGraw-Hill.
Renn O., 1999, Three decades of risk research: accomplishments and new challenges, J. Risk Res., 1, 49–71.
Sawacha E., Naoum, S., Fong, D., 1999, Factors affecting safety performance on construction sites, Int. J.
Proj. Manag. 17, 309–315.
Stirling A.,1998, Risk a taturning point, J. Risk Res., 1, 97–109.
Tixier J., Dusserre G., Salvi O., Gaston D., 2002, Review of 62 risk analysis methodologies of industrial plants,
J. Loss Prevent. Proc. 15, 291–303.
Todorović P.M., Gordić D.R., Babić M., Jeremić B.M., Demichela M., Mačužić I.D., 2014, An implementation of
infrared thermography in maintenance plans within a world class manufacturing strategy, Therm Sci. 17,
977-987.
Zimmermann H.J., 2001, Fuzzy Set Theory and its Applications, Kluwer Nijhoff Publishing, Boston.
1236