Engineering Failure Analysis 126 (2021) 105456

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Human factor risk management procedures applied in the case of

open pit mine
Snezana Kirin a, Aleksandar Sedmak b, Wei Li c, Miodrag Brzaković d,
Igor Miljanović e, Ana Petrović b, Simon Sedmak a, *
a
Innovation Centar of Faculty of Mechanical Engineering in Belgrade, Serbia
b
Faculty of Mechanical Engineering in Belgrade, University of Belgrade, Serbia
c
Faculty of Mathematics and Statistics, Xidian University, Xi’an, China
d
Faculty of Applied Management, Economics and Finance, University of Business Academy in Novi Sad, Serbia
e
Faculty of Mining and Geology, University of Belgrade, Serbia

A R T I C L E I N F O A B S T R A C T

Keywords: Risk management aims to provide a controlled work environment to ensure the safe operation of
Risk management the high-risk systems. It is a dynamic process which work in a continual state of change. The issue
Human factor of human factor risk and rules and regulations in open pit mine is a main focus of this paper in
Bucket wheel excavator
order to develop predictive models of behavior of workers in relation to compliance with the
Finite element method
procedures and rules. Presented survey was conducted in open pit coal mine, as high-risk system,
involving 476 mineworkers. The survey was in the form of a questionnaire, consisting of
approximately 45 questions. It aimed to find out the opinions of the mining workforce about risk
attitude generally as well as about safety rules and regulations. The first goal was to determine
factor with the biggest influence on risk of human factor. The second goal was creation model for
prediction behavior of mining workers. The survey also aimed to examine: (a) major human risk
factors at a specific open-pit mine site (b) mine workers’ opinions about policies and procedures;
(c) the manner in which mine safety rules and regulations are perceived and understood; (d) the
frequency of deviation from rules and regulations; (e) attitudes related to risk taking and their
interaction with rules and regulations; and (f) anticipate the behavior of mining workers with
respect to compliance with policies and procedures. In addition, a case study is presented, related
to failure of mining equipment, the SchRs 630 bucket wheel excavator, which had occurred due to
excessive loads caused by a human factor, such as inadequate operation of the equipment. The
goal here was to further relate the effects of the human factor to the structural integrity, by
providing a reliable numerical model which could be used to determine the critical locations
where failures could occur, and by combining this approach with the previously described risk
analysis methodology, structure safety can be improved.

1. Introduction

The mining industry is a strategic economic sector for many countries which has been viewed as a high-risk activity, and therefore,

* Corresponding author.
E-mail addresses: skirin@mas.bg.ac.rs (S. Kirin), asedmak@mas.bg.ac.rs (A. Sedmak), liweilw@mail.xidian.edu.cn (W. Li), igor.miljanovic@rgf.
bg.ac.rs (I. Miljanović).

https://doi.org/10.1016/j.engfailanal.2021.105456
Received 15 February 2021; Received in revised form 21 April 2021; Accepted 22 April 2021
Available online 7 May 2021
1350-6307/© 2021 Published by Elsevier Ltd.
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

risk management in this area is an important research topic. Global socio-political and technological changes are causing new risks.
Mining Risk Review from 2019 addressed the uncertainties of mining risk and mining risk transfer pointed out new risks related to
digitization; bottlenecks; geopolitical risk and social economic development, [1]. In the same paper, review of the literature is pre­
sented, dealing with risks in the mining industry, [1], with the main research topics: human factor, machines and environment. Risk
management is a complex process enabling safe working process if performed well and it is influenced by organisational factors,
environment condition or task condition, human actions (team or individual) or failed defence or even absence of defence in case of a
hazard situation. Risks analysis today had shown that standardisation and introduction of mechanisms for risk management moves the
focus of the safety management problems from purely technological area to an area which involves human factor as well. Risk
management should not be based on response to past accidents, but contrary to that, increasingly proactive. Thereby, an adaptable
strategy based on a closed feedback loop and measuring or monitoring of current safety levels needs to be applied, taking into account
the explicitly formulated level of desired safety. One of the major problems in risk management is difficulty in predicting the human
factor for risk: people make mistakes, may not succeed in performing an operation, or may experience health conditions during work,
[2]. In empirical study about 91 mining accidents in the platinum mine in South Africa, [3], it was concluded that “Unsafe acts were
identified in 98.9% of the accident reports analysed”.
Risk management deals with event, which carries a risk of a specific unwanted outcome (accidents causing certain amount of
damage), probability of this event occurring and its significance (value, quantity and amount). Methods for risk assessment may
include [4,5]: Preliminary potential hazards analysis (PHA); Third party risk analysis; Hazard and Operability studies - HAZOP;
Analysis of records; Failure Mode Effects Analysis - FMEA; “What if” method; “lean” approach to risk management; Failure Tree
Analysis - (FTA). One can also classify methods for risk assessment as qualitative and quantitative. Qualitative risk analysis is the
measure of risk based on descriptive categories such as low, medium, high, irrelevant, relevant, very important, etc. Qualitative
analysis includes method for determining of priorities for further actions or planned reactions to risks. Qualitative risk analysis
evaluates the priority of identified risks using the probability of their occurrence and effects on the system. It is usually a fast and
efficient way to establish priorities for deeper analysis and, if necessary, should be used throughout the entire lifespan of the systems in
order to notice possible problems on time. Qualitative risk analysis uses results of risk management planning and risk identification
process as input data, and can lead to direct planning of response to risk or more detailed risk analyses using quantitative methods.
Quantification of risk emphasizes quantitative risks and provides methods for its assessments. Thereby, a limited number of
consequences is taken into account, since only those consequences that can be quantified are considered. Statistical analysis and
system analysis techniques are used. The drawback of this analysis is that it does not take into account the risk of the human factor.
Often neglected, the human factor involves human capabilities and limitations in the scope of their work, as well as their mutual
interaction and interaction with machines, systems and the environment. People activate machines, make and adjust the organization
of work processes and apply rules and procedures, and typically cannot change as fast as technology and work processes do nowadays,
[5]. According to [6], “unsafe behavior of frontline workers is considered as a direct, critical factor contributing to workplace injuries
and accidents across diverse high-risk industries”. According to [7] “most analytical effort for including human errors has been spent
on verification of the safety of existing, operating process plants of traditional design with respect to the man–machine interface.”
Detailed risk analysis of human factor in combination with standard engineering risk analysis, based on risk matrix, is described in [8].
According to [9] top 10 causes of accidents on opencast mines are: behaviour problems; poor communication; vehicle accidents;
explosives, electricity, working at heights; confined spaces; fire; lifting and falling objects; fall of ground. The importance of respecting
procedures and rules throughout the work in the mine was strongly emphasized, [9].
Rules and procedures are key features for a modern organization to function, [10]. Policies and procedures are an important
segment of risk management to ensure worker safety, process safety and environmental security of modern industrial systems. Modern
management system promotes a wide scope of norms, rules and procedures in all activities, [11,12]. Post-incident reports are most

Fig. 1. Collapse of bucket wheel excavator SRs 1760.

2
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

often associated with procedures or rules: either to report about bending of rules, or to initiate improvements. For example, according
to [13] “to avoid unexpected failures of bucket wheel excavators (BWE) and ensured its structural integrity during operation, it is
necessary care during operational life, monitoring and diagnostics of all vital elements of the supporting structure, and sometimes
repair and redesign are also required”. All that is related to human factor competences, establishing adequate procedures and rules and
their behaviour in accordance with the them. According to [14] “the majority of accidents occurred in the working face, especially in
the heading face; the direct causes of accidents were still the behaviour that ‘‘neglecting or breaking the system regulations, procedures
and norms’’”. He pointed out mostly accepted quantification of human error rates in terms related to the structure of external task, but
not in terms of human capabilities and their limitations.
In order to avoid disasters like the one when “BWE SchRs 1760, unexpectedly and catastrophically failed in 2004, after 17 years of
regular service on an open surface mine in Serbia, Fig. 1”, [13], risk management system should be continually improved, [15].
Since man is the key element of the working process, every serious risk analysis includes the influence of the human factor. Analysis
of open pit failures, Table 1, shows that the human factor is responsible for the largest percentage of them, because of both design flaws
and machine operating and their maintenance. Two more examples of bucket wheel excavator failure due to human errors are shown
in Fig. 2 and Fig. 3.
During the analysis of the human factor it is very important to determine the conditions in which the incident occurs such as night
shifts, or working on several things simultaneously, poor evaluations during design, human tendency to overestimating themselves,
routine, poor communication, as well as the fatal combination of different reasons. Fig. 4 shows the difference between the risk of
occupational accidents and major disasters, indicating that the occupational accidents occur frequently, but their consequences are less
severe. The most serious effects occur in the area of small probability.
Even though focus of this research is on human aspects of risk analysis, first we will introduce the case study, as analysed by
engineering methods, including finite element and fracture mechanics approach. This analysis explained well what was the problem
and direct cause of the accident, but also point out that human error is in the background. Results of research, [16], “show that the
proportion of total mine fatalities attributable to the equipment ranged from 37% to 88% per year.“ They also presented “that workers
with less than five years of appropriate mining experience constitute 44% of all fatalities that occurred during the period of
1995–2005.”

2. Case study – Numerical modelling of the schrs 630 bucket wheel excavator

This section of the paper will consider the numerical analysis of the SchRs 630 bucket wheel excavator, commonly used for mining
applications. This analysis focuses on the stress distribution in the bucket wheel excavator structure, for the purpose of determining the
locations of stress concentration, as potential critical locations for crack initiation and/or failure. Up to this point, the failure of
aforementioned equipment was considered from experimetal and probabilistic point of view, [17-25], including suggested recon­
straction of the excavator, [26-32]. The analysed structure was made using S355JR2 steel, with the following mechanical properties
which were used as the input data for the calculation – elasticity modulus of 210 GPa, yield stress of 355 MPa and density of 7800 kg/
m3 [24].
Numerical simulation was performed in home-made KOMIPS software [33], involved the development of a sub-scaled model of a
large structure, which was subjected to various load cases, including dead weight, work loads (static and dynamic) and inertial forces
which occur during bucket wheel excavator operation. The main goal of these simulations was to create a representative model which
could be used instead of testing of real structures, since this could be a very complicated, or even impossible to achieve. However, the
focus of this case study will be on the inertial forces and their effects on the structural integrity of the bucket wheel excavator, since this
load case was the cause of failure. Another important factor here was the fact that these inertial forces are directly affected by the
bucker wheel excavator operater, hence there is a very close correlation between the failure and the human factor. Fig. 5 shows the
exact location of the crack which was initiated at one of the boom supports.
Numerical model used for finite element method calculation is shown in Fig. 6. Linear elements were used in orded to simplify the
complex geometry, with puprle arrows indicating loads acting on the structure, and orange triangles representing boundary condi­
tions, Fig. 6. A number of different load cases were assumed during this analysis, including dead weight, a combination of dead weight
and work loads and, finally, the load case which included the inertial forces generated by the movement and subsequent braking of the
boom. More details about each individual load case and the stress distribution related to them can be found in [24]. The most
interesting case for this analysis was the combination of dead weight and working load. While a possibility of failure due to high
inertial forces caused inadequate operation of the bucket wheel excavator boom does exist, it is extremely unlikely to happen in real

Table 1
Structure of failure causes.
Failure cause Failure share [%]

Difficult exploitation conditions 27

Design and assembly flaws 22
Operator errors 18
Maintainer errors 13
Material fatigue, equipment, wear and corrosion 8
Inadequate design solution 7
Others 5

3
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

Fig. 2. Failure of SRs 1200 bucket wheel excavator failure due to human error, [15]

Fig. 3. Failure of SchRs 630 bucket wheel excavator failure due to human error, [15]

Disastrous
accidents
Accident severity

Occupational
accidents
Probability/frequency

Fig. 4. Professional and disastrous accidents.

4
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

Fig. 5. Crack location in the SchRs 630 bucket wheel excavator [24]

Fig. 6. Finite element representation of the bucket wheel excavator [24]

working conditions, due to safety systems in place, and for this reason, that particular scenario was not considered in this analysis.
Stress distribution is shown in Fig. 7.
The locations of the highest stress can be clearly seen in the model. Their magnitudes are significantly lower than the allowed
stresses for this load case, and yet failure of the whole structure occurred. It should also be noted that the entire structure got separated
from its foundation (Fig. 3), instead of failing at stress concentration locations, shown in Fig. 7, which would be the expected. This
suggests that the design and regular working conditions were not the cause of failure, as the stresses are on the safe side. Hence, this
calculation confirms that it was indeed human error that resulted in the catastrophic failure. This is further reinforced by the fact that
this load case, in theory is not the most dangerous one – the case with inertial forces produces higher loads, as shown by the the
following table. The individual values in each row are related to four selected finite elements from the model. Stress magnitudes of
136.9 and 139.1 MPa are particularly interesting, as these values could result in crack initiation, Table 2.
It should be noted that the first two load cases were static in nature, whereas the inertial forces occurring during the braking can be
considered as a dynamic, impact load. This is important fact, since it explains why the maximum stresses mentioned above resulted in
crack initiation, despite being slightly lower than the maximum stresses in the supports in some other cases. Despite all of this, it was
the combination of dead weight and working load that proved to be fatal for the bucket wheel excavator.
This was due to a combination of a number of factors, all of which were human in nature, mainly due to breaking rules and
procedures. These factors were already mentioned in the introductory part of the paper, and are generally related to improper
operation of the equipment in question, which resulted in loads significantly greater than the ones which should normally occur under

5
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

Fig. 7. Equivalent (von Misses) Stress distribution in the bucket wheel excavator model for the load case with inertial forces (left), the critical joint
with highest stresses (top right), and magnitudes (bottom right).

Table 2
Von Misses stresses, MPa.
The first load case(dead weight) The second load case(dead weight and work load) The third load case(inertial forces)

8.38 19.29 136.9

10.87 18.63 97.31
9.67 19.70 139.1
10.2 15.34 100.3

the given conditions. From this, it can be clearly seen that the risks related to human factors in the mining industry are extremely
dangerous, as these factors can completely change the way in which a structure behaves during its exploitation. In the next section, a
detailed analysis of these risks, along with the suggestions on how to eliminate them (or at least reduce them as much as possible).

3. Risk management in open pit mines

The plan of implementing the risk assessment includes organization and coordination, list of experts and the timelines, ways of
gathering documentation and information from employee surveys, consulting with employee representatives and informing them
about the results of risk assessments and measures taken, and all other activities required.
Identification of hazards should take into account: routine and non-routine activities of the entire staff that has access to the work
place, the human factor, working equipment, outside effects on work health and safety, events in vicinity of the work place, changes of
management systems, including temporary ones, data about previous incidents, accidents and emergency situations related to op­
erations/activities, processes, products /services and work place.
Re-examining of hazards and harms takes into account all changes in operations/activities of the mine and in the environment that
could lead to increase or decrease of hazard levels (refitting of objects, installations, plants, changes in processes, technology,
introducing of new processes, products/services, energy, changes in organization structure, laws and regulations, serious incidents,
accidents, emergency situation). During the re-examining, checklists, interviews, measures, direct control, results of previous system
management checks and other re-evaluations are used depending on the nature of the operation/activity being examined.

6
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

The life cycle of machines comprises design, manufacturing, delivery, commission, exploitation, revitalization and decommission,
as given in an example of a bucket wheel excavator, Fig. 8.
Analysis of risk for machines in open pits requires data about the intended use of the machine and assumed misuses of machines
during its life cycle, analysis of qualifications and competence of machine operators, exposure of other individuals to the effects of
machine’s work, as well as limitations (operator - machine, installing, maintenance), time limits (machine, service intervals) and
environmental limitations (inside/outside, dust, temperature, landslides). Identifying of hazards (dangers) include mechanical,
electrical, heat, noise, vibration and radiation. Risk analysis includes risk assessments, i.e. assessment of maximum possible injuries/
damages and the probability form them to occur and these are essential information typical for every risky situation. Risk diagrams and
matrices are used in practice to display the risks of a dangerous situation.
During the probability assessment, analysis of historical data is applied, when data about previous incidents exist, and evidence of
previous malfunctions, types of damage/ failures, properties of relevant equipment and machines are taken into consideration. Risk
matrix can be modified and used to define priority of maintenance which leads to a period during which the equipment needs to be
repaired, Table 3. Potential consequences are the same as in the “big-main” RAM matrix, Table 4.

4. Influence of personnel reliability on exploitation safety of excavator units

Research methodology

The complexity of factors that affect employee behavior in respect to risk is investigated by conducting a comprehensive survey.
This also involves the factor analysis which implies that multiple variables have similar response scheme, [34], and use of statistical
methods to determine the main factors influencing the risk of human factor, being applied to a group of 33 variables. The obtained
factors were then used as input variables for binary logistic regression to establish the predictive model of employee behavior
regarding rules. As a software tool, IBM SPSS Statistics 25 was used, whereas the results were presented using MS EXCEL.
At the time of the research, the total number of employees was 2121. The sample size represents 22% of the total number of
employees and included all represented educational categories as well as representatives of all work processes at the open pit mine. The
survey questionnaire was constructed specifically for this purpose.
Randomly chose 476 employees have been chosen as the sample in open pit mines, involved in coal extracting. Data and personal
attitudes of employees provided relevant information for the problem analysed here, related to age and mining experience, as pre­
sented in Table 5.

4.1. Determining of the main factors

The following variables are observed: age; years of service; the description of my job is clear to me; I have enough knowledge for my
job; the nature of my workplace is at increased risk for me; the nature of my workplace is at increased risk for others; my manager
supports me; my manager controls security; the manager tells me if I’m doing well; I work in a group / team of colleagues; when I
notice something that may lead to a problem or an accident, my reaction depends on my assessment of the danger; I feel the support of
my colleagues at work; I communicate well with my colleagues and there is no problem to understand each other; Colleagues are
generally predictable, competent and well-meaning; We all strive to work safely;rules and regulations are important for my safety; I
know that people violate rules and regulations; sometimes it is necessary to break the rules to get the job done; executives are aware of
violations of rules and regulations; being careful will reduce the chance of an accident; communication about rules and regulations is
generally pretty good; Managers explain why rules or regulations are necessary; improved training and introduction to job will help in
understanding and implementing rules and regulations; I consider myself effective; I have a high degree of self-esteem; I am always
focused on work; I am social; I am ready to work together; I am an extrovert; I feel good in my skin; I am happy; I am happy with my
overall life I plan to work at the mine for the next 5 years.
Factor analysis is applied to establish the main factors affecting the risk behaviour, starting with checking if the data set is
appropriate for the factor analysis: since KMO = 0.841 > 0.6 and the level of significance, Sig = 0.000 less than 0.05, the justifiability
condition is fulfilled.
Based on the criteria of eigenvalues, Catel criterion and the rule of retain any eigenvalue that accounts for at least 5% of the
variance, it was decided to use four factors for further research, named as Satisfaction with life, Supportive leadership style, experience

Fig. 8. The life cycle stages of a bucket wheel excavator.

7
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

Table 3
Priorities.
RAM: Priority Suggested time for completion of work

x Reject the request –

6 Normal priority 12 weeks
5 Normal priority 8 weeks
4 Normal priority 6 weeks
3 Normal priority 2 weeks
2 Emergency priority 1 week
1 Urgent priority Immediately + overtime

Table 4
Risk matrix.
Potential consequences on Failure probability
people property environment reputation A B C D E
≥ 18 3 to 18 2 to 3 2 days to 2 ≤2
months months weeks weeks days

0 No injury No consequences No consequences No consequence x x x x x

1 Slight injury Slight Slight Slight 6 6 5 4 3
consequences consequences consequences
2 Minor injury Minor Minor Minor 6 5 4 3 2
consequences consequences consequences
3 Severe injury Severe Severe Severe 5 4 3 2 1
consequences consequences consequences
4 Permanent Significant Significant Significant 4 3 2 1 1
disability consequences consequences consequences
5 Fatality Huge Huge Huge 3 2 1 1 1
consequences consequences consequences

Table 5
Sample description.
Frequency Percent

Age 20–29 24 5,0

30–39 126 26,5
40–49 205 43,1
50–59 107 22,5
over 60 14 2,9
Service (years) less than 5 29 6,1
5–14 50 10,5
14–24 158 33,2
25–34 182 38,2
over 35 57 12,0

and risky job. They can explain 48,86% of the variance, with few details given in the following text.
Personality characteristics, related to items: (I’m ready to work together, I’m always focused on work, I’m an extrovert, I consider
myself effective, I feel good in my skin, I’m social, I am happy, I am happy with my overall life, I have a high degree of self-esteem, I
plan to work at the mine for the next 5 years, Being careful will reduce the chance of an accident, Rules and regulations are important
for my safety, When I notice something that may lead to a problem or an accident, my reaction depends on my assessment of the
danger, I have enough knowledge for my job, Improved training and introduction to job will help in understanding and implementing
rules and regulations.
Supportive leadership style, related to items: My manager controls security, The manager tells me if I’m doing well, My manager
supports me, I feel the support of my colleagues at work, Colleagues are generally predictable, competent and well-meaning, I
communicate well with my colleagues and there is no problem to understand each other, We all strive to work safely,I work in a group /
team of colleagues, Managers explain why rules or regulations are necessary, communication about rules and regulations is generally
pretty good; The description of my job is clear to me.
Mining experiance, related to items: Years of service, Age, I know that people violate rules and regulations, Executives are aware of
violations of rules and regulations, Sometimes it is necessary to break the rules to get the job done.
Risky job, related to items: The nature of my workplace is at increased risk for me, The nature of my workplace is at increased risk
for others.

8
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

4.2. Rules and regulations

Analysis has shown that 109 (22.9%) claimed that they break rules and regulations, whereas 367 (77.1%) claimed that they do not.
Fig. 8 shows answers of the mine workers about reasons for risk taking: “the pressure of director”, 18%, “simply, people take risk”,
17%, “bad rules and regulation“17%, “it is easiest way to do it”, 13%, “people get tired” 9%, “male behaviour” and “the pressure of
honor, each 2%. It is important to note that 22% of respondents answered “other reason for risk taking.” This means that the list of
reasons needs to be improved. In comparison to similar research conducted by Laurence (2005) [9] in several different mines in
Australia, obtained results are different. In his research the most important reason was “people get tired”, 25%, and “it is easiest way to
do it”, 21%. He had less than 4% “other” answers.
The results, Table 6, show that work procedures and rules are mostly violated by workers who have enough experience and should
be the safest in terms of competencies and skills. Due to the importance of following the rules and procedures in order to improve and
maintain the desired level of safety in the mine, we researched the attitudes of employees about the existing rules and procedures.
Obtained results show that 39% of workers consider that rules do not define the real situation on the job; 10% state “rules are too old“,
8% think the rules are not understood, 8% think “rules are bad and contains errors“, 7% think that rules are too rigid, 5% state “too
many things to remember“, 4% condiser rules “too complex“ and another 4% say “they are simply bad“, and finally, 15% say
“something elase“. Results presented by Laurence [9] are different: about 19% of workers means that there are lack of real world in
rules, 18% thinks “there are too many to remember“ in rules, 18% consider rules to complex, 16% respondents think rules are to rigid.
Differences in workers’ attitudes about good rules, depending on whether they break them or not, are shown in Table 7.
The obtained results show that a significant number of workers who violate the rules and procedures believe that they should be,
above all, practical. Taking into account the data obtained above, that this is mostly done by experienced workers in the best part of
their career, it can be concluded that they consider themselves skilled and competent to assess that they can break the rules for
practical reasons.
Finally, as for the attitude to the rules and regulations, 57% state “to reduce the risk”, 22% state “The problem with the rules and
regulations”, 10% say “to save time”, 8% claims “Lack of management commitment”, and 3% say “to save energy”
Action taken after violation of rules and procedures were identified are presented in the Table 8. Based on the answers of the
workers, it can be seen that disciplinary actions are taken in less than 50% of cases (46.8%), while in 26% only a report is made. In 8%
of cases, the situation is analyzed, while in 5.5% of cases no action is taken (see Table 9).

4.3. Prediction model

Examples of categories on binary variables that are sometimes used as criterion variables could be the following: folows the rules –
does not folow the rules, has learning problems - no learning problems, leave the work place – do not leave the work place, etc. Binary
logistic regression enables the examination of outcome prediction models given in two categories and IBM SPSS Statistics 25 Binary
Logistic Regression technique was applied in aim to create the predictive model.
The model consists of an independent variable “deviation of rules“ and of predictor variables: personality characteristics (related
mostly to satisfaction with life), supportive leadership style; mining expirience; risky job. The predictor variables were obtained in the
previous factor analysis procedure.
By default, SPSS logistic regression makes a list of missing data, so that if there is missing value for any variable in the model, the
entire case will be excluded from the analysis. Thus, out of 476 cases, 467 were included in analysis. The Omnibus Tests of Model
Coefficients is used to check that the new model (with explanatory variables included) is an improvement over the baseline model
(without predictors) Omnibus Tests: − 2LL = 80.104 = Model χ2 df = 4, p < .001. In this case there is a significant difference between
the Log-likelihoods of the baseline model and the new model (sig less than 0,001). Grouping cases together was proposed in [35]
according to the predicted values from the logistic regression model. Specifically, the predicted values are arrayed from lowest to
highest, and then separated into several groups of approximately equal size. Goodness-of-fit (GOF) tests help decide whether your
model is correctly specified. They produce a p-value - if it’s low (say, below 0.05), the model is rejected. If it’s high, the model passes
the test, Table 10.Since Sig = 0,249 > 0, 05, the nonsignificant chi-square is indicative of good fit of data with linear model. For each of
ten groups the observed number of deviate and not deviate events is calculated, as well as the expected number of deviate and not
deviate events, Table 11. The expected number of deviate event is just the sum of the predicted probabilities over the individuals in the
group. For each bin and each event, we have number of observed cases and expected number predicted from the model. Table 11 shows
the stacking of the empirically obtained (Observed) categorical affiliation of observation units on a criterion variable and their pre­
dicted (Predicted) categorical affiliation based on a logistic model containing all the predictors introduced in block 1.
The classification table presented in the Table 12 is based on the model that includes explanatory variables and it correctly classifies
the outcome for 81, 4% of the cases. Table 13 contains the logistic coefficients estimates for the model with the predictors introduced in
block 1 (column B). Data given in the “S.E.” column presents the asymptotic standard errors for the individual logistic coefficients are
shown. The Wald column contains Wald ’s H2 statistics, the df degree of freedom column, and the Sig column (to test the hypothesis
that the logistic coefficient for the predictor variable vj is zero). Column exp (b) contains exponential logistic coefficients that are very
important for interpreting logistic regression outcomes.These are the values for the logistic regression equation for predicting the
dependent variable from the independent variable. Therefore, the logistic model estimated from a given sample looks like this:

9
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

Table 6
Mining experience of those who violate rules and procedures.
years Frequency Percent Valid Percent Cumulative Percent
les than 5 4 3.7 3.7 3.7
5-14 12 11.0 11.0 14.7
15-34 35 32.1 32.1 46.8
25-34 43 39.4 39.4 86.2
over 35 15 13.8 13.8 100.0
Total 109 100.0 100.0

Table 7
Good rule or regulation.
Frequency Percent Valid Percent Cumulative Percent

I deviate from the rules Involvement of the workforce 6 5,5 5,5 5,5
simplicity 9 8,3 8,3 13,8
Intelligibility 10 9,2 9,2 22,9
Practicality 25 22,9 22,9 45,9
When the data on the basis of common sense 3 2,8 2,8 48,6
Accuracy 11 10,1 10,1 58,7
Everything mentioned 31 28,4 28,4 87,2
Something else 14 12,8 12,8 100,0
Total 109 100,0 100,0
I do not deviate from the rules Involvement of the workforce 50 6,8 6,8 6,8
simplicity 50 6,8 6,8 13,6
Intelligibility 104 14,2 14,2 27,8
Practicality 90 12,3 12,3 40,1
When the data on the basis of common sense 74 10,1 10,1 50,1
Accuracy 70 9,5 9,5 59,7
Everything mentioned 278 37,9 37,9 97,5
Something else 18 2,5 2,5 100,0
Total 734 100,0 100,0

Table 8
Actions taken after violations of the rules were identified.
Frequency Percent Valid Percent Cumulative Percent

Report is made 124 26,1 26,1 26,1

Analysis and consulting 38 8,0 8,0 34,0
Retraining 15 3,2 3,2 37,2
Disciplinary action 223 46,8 46,8 84,0
Without action 26 5,5 5,5 89,5
Something else 50 10,5 10,5 100,0
Total 476 100,0 100,0

Table 9
Case summary.
Case Processing Summary

Unweighted Casesa N Percent

Selected Cases Included in Analysis 467 98.1

Missing Cases 9 1.9
Total 476 100.0
Unselected Cases 0 0.0
Total 476 100.0
a
If weight is in effect, see classification table for the total number of cases.

Table 10
Goodness-of-fit test.
Hosmer and Lemeshow Test

Step Chi-square df Sig.

1 10.240 8 0.249

10
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

Table 11
Goodness-of-fit-contingency table.
Contingency Table for Hosmer and Lemeshow Test

Deviation from rules = Not deviate Deviation from rules = Deviate Total

Observed Expected Observed Expected

Step 1 1 44 45.660 3 1.340 47

2 44 44.344 3 2.656 47
3 42 42.538 5 4.462 47
4 41 40.936 6 6.064 47
5 40 39.145 7 7.855 47
6 37 37.456 10 9.544 47
7 36 35.347 11 11.653 47
8 40 32.248 7 14.752 47
9 24 27.739 23 19.261 47
10 14 16.587 30 27.413 44

Table 12
Classification table.
Classification Tablea

Observed Predicted

Deviation from rules Percentage Correct

Not deviate Deviate

Step 1 Deviation from rules Not deviate 352 10 97.2

Deviate 77 28 26.7
Overall Percentage 81.4
a
The cut value is, 500

Table 13
The model variables.
Variables in the Equation

B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B)

Lower Upper

Step 1a Personality characteristics -0.612 0.118 27.054 1 0.000 0.542 0.430 0.683
Leadership -0.292 0.118 6.071 1 0.014 0.747 0.592 0.942
Experience 0.506 0.135 14.116 1 0.000 1.658 1.274 2.159
Risky job 0.797 0.134 35.280 1 0.000 2.219 1.706 2.886
Constant -1.547 0.140 121.767 1 0.000 0.213
a
Variable(s) entered on step 1: Personal characteristics, Leadership, Experience, Risky job.

( )
P(deviate)
ln(ODDS) = ln
1 − p(deviate)
= − 1, 547 − 0, 612*Personalitycharacteristics − 0, 292*Leadershipstyle + 0, 506*MiningExpirience + 0, 797*Riskyjob

If we expose the logistic coefficient for “Personality characteristics“, we get the value − 0,612 in the “B column“ and the column Exp
(b) of the Variables in the Equation table:
exp(b1) = exp( − 0, 612) = 0.542isaoddsratio(2nodeviateresponsesforevery1 ˝deviatefromrules˝, p = 0, 33),
Thus, the chances of answering the question with “deviated from rules“(according to non-deviated) are decreased about twice
when the “value” on the “Personality characteristics“predictor variable is “increased” by 1 and the other three predictors in the model
are kept constant. In this particular case, personality characteristics are mainly described through level of socialization, satisfaction
with life and competence at work. This means that by increasing satisfaction with life and work, as well as a higher level of social­
ization, they are affected in a way that reduces violations of rules and procedures.
Likewise, if we expose the logistic coefficient for Risky JOB, we get the value in the row Risky JOB and the column Exp (b) of the
Variables in the Equation table: exp (b4) = exp(0,797) = 2.219. It actually means that the chances for the answer “deviated from rules“
is 2.219 times higher for those who has risky job.

11
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

5. Discussion

From previous empirical research (Swira, Chen, Kirin), it is evident that employee behavior plays a key role in preventing accidents
in the mining industry. The importance of respecting the rules and procedures was emphasized. It was determined that there is a need
to improve the existing rules and procedures and / or to raise the trust of employees in them.
The research presented in this paper was conducted to identify the major human risk factors that then served as predictors for
predicting workers’ behavior with respect to adherence to rules and procedures. The obtained results indicate that leadership plays an
important role in worker behavior. A supportive leadership style results in more responsible employee behavior and a lower proba­
bility of deviation of rules and procedures. The result also indicates that the safe behavior of workers is strongly influenced by the
difficult to control factor of “personal characteristics” highly related to life satisfaction and socialization level. It has also been found
that riskier work and longer work experience increase the likelihood of breaking the rules. Reasons for risk taking are “the pressure of
director”, “simply, people take risk”, “bad rules and regulation“, “it is easiest way to do it”, and other reasons which needs to be
explored more deeply. Obtained results are compared to results from symilar survay, Laurence (2005) who got that the reasons for risk
taking are “people get tired “and “it is easiest way to do it”, 21%.
Workers’ views on rules and procedures were explored. Obtained results show that 39% of workers consider that rules are not
adequate and do not define the real situation on the job; 8% think the rules are not understood, 8% think “rules are bad and contains
errors“, while 7% of workers thinks that rules are too rigid. Laurence (2005) got that the main problems with the rules and procedures
were as follows: lack of real world in rules, too many to remember, complexity, too rigid rules.
In addition, case studies involving failure of mining equipment were analysed, both from risk and fracture mechanics points of
view, and in the presented cases, the failure was a direct consequence of human behaviour related to working procedures during the
operating of said equipment. This further reinforces the aforementioned points about the need to adjust and change the rules in a way
that makes them easier to understand and follow properly.

6. Conclusions

Regulatory requirements for workplace safety represent factors that force companies to devote considerable attention to consid­
ering the human factor in risk management and improvement of safety at work.
Nowadays, when quality standards introduced become more and strict, safety of technical systems raises everything to a higher
level, while humans with their characteristics cannot keep up with such a trend, the reduction of risks related to the human factor
represents the essentially most important step in risk reduction in industrial systems.
The introduction of rules and procedures raises the level of security in high-risk systems, but risk management should be a “live”
process that must not be rigid but open to innovation in the formation of more effective, efficient and clearer policies and procedures,
with the aim of increasing safety. As can be seen from the practical examples, this approach could also increase the working life of
mining equipment, since it was demonstrated how the human factor influences the potential for failure, and better compliance with the
rules would ensure lower levels of dynamic loads during exploitation. As a result, the life of this equipment would be extended, as it
could withstand a greater amount of work cycles.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

References

[1] A. Tubis, Sylwia Werbinska-Wojciechowska,

́ Adam Wroblewski, Risk Assessment Methods in Mining Industry—A Systematic Review, Appl. Sci. 10 (2020) 5172,
https://doi.org/10.3390/app10155172.
[2] S. Kirin, M. Božic, M. Brzakovic, I. Vucetic, Challenges of future research in the area of industrial safety, Structural Integrity and Life 15 (2) (2015) 71–78.
[3] J. Bonsu, W. van Dyk, J.-P. Franzidis, F. Petersen, A. Isafiade, A systemic study of mining accident causality: an analysis of 91 mining accidents from a platinum
mine in South Africa, Journal of the Southern African Institute of Mining and Metallurgy (2017), https://doi.org/10.17159/2411-9717/2017/v117n1a9.
[4] S. Kirin, A. Jovanović, P. Stanojević, A. Sedmak, E. Džindo, Risk analysis in structural integrity – application to a large company, Structural Integrity and Life 11
(3) (2011) 209–212.
[5] S. Kirin, P. Stanojevic, I. Miljanovic, A. Sedmak, T. Peric, P. Ilic, Influence of the human factor on risks in an open-pit mine, Structural Integrity and Life 15 (2)
(2015) 117–128.
[6] N. Xia, M.A. Griffin, X. Wang, X. Liu, D. Wang, Is there agreement between worker self and supervisor assessment of worker safety performance? An examination
in the construction industry, Journal of Safety Research 65 (2018) 29–37.
[7] Jens Rasmussen, Human errors, A taxonomy for describing human malfunction in industrial installations, Journal of Occupational Accidents 4 (2–4) (1982)
311–333.
[8] T. Golubović, A. Sedmak, V. Spasojević Brkić, S. Kirin, I. Rakonjac, Novel risk based assessment of pressure vessels integrity, Tehnički vjesnik 25 (3) (2018)
803–807.
[9] Jimmy Swira, 2018, Top 10 causes of accidents on opencast mines, African Mining Brief, February 2018.
[10] M. Bourrier, C. Bieder, Trapping Safety Into Rules?: How Desirable or Avoidable is Proceduralization? CRC Press, 2013.
[11] J. Bell, N. Healey, The Causes of Major Hazard Incidents and How to Improve Risk Control and Health and Safety Management: A Review of the Existing
Literature, (2006) HSL/2006/117.
[12] D. Laurence, Safety rules and regulations on mine sites – The problem and a solution, Journal of Safety Research 36 (2005) 39–50.
[13] D. Danicic, S. Sedmak, I. Blacic, S. Kirin, Scenario of fracture development in bucket wheel excavator, Structural Integrity and Life 13 (3) (2013) 189–196.

12
S. Kirin et al. Engineering Failure Analysis 126 (2021) 105456

[14] Hong Chen, Hui Qi, Ruyin Long, Maolong Zhang, Research on 10-year tendency of China coal mine accidents and the characteristics of human factors, Safety
Science 50 (4) (2012) 745–750.
[15] Snežana Kirin, Milorad Pantelić, Igor Miljanović, Violeta Talović, Risk management of machines in open pit mines, Structural Integrity and Life 13 (3) (2013)
197–202.
[16] Vladislav Kecojevic, Dragan Komljenovic, William A Groves, Mark Radomsky, An analysis of equipment-related fatal accidents in U.S. mining operations:
1995–2005, October 2007, Safety Science 45(8):864-874, DOI:10.1016/j.ssci.2006.08.024.
[17] P. Jovančić, D. Ignjatović, M. Tanasijević, T. Maneski, Load-bearing steel structure diagnostics on bucket wheel excavator, for the purpose of failure prevention,
Engineering Failure Analysis 18 (4) (2011) 1203–1211.
[18] J. Gottvald, Analysis of Vibrations of Bucket Wheel Excavator Schrs1320 During Mining Process, FME Transactions 40 (2012) 165–170.
[19] D. Daničić, T. Maneski, The Structure Failure of the Discharge Boom of Bucket Wheel Excavator C 700 S due to Dynamic Effects, Structural Integrity and Life 12
(1) (2012) 43–46.
[20] Ž. Lazarević, I. Aranđelović, S. Kirin, An Analysis of Random Mechanical Failures of Bucket Wheel Excavator Structural Integrity and Life, 15(3) (2015)
143–146.
[21] D. Polovina, S. Ivković, D. Ignjatović, M. Tanasijević, Remaining operational capabilities evaluation of bucket wheel excavator by application of expert
assessment method with empirical correction factor, Structural Integrity and Life 10 (1) (2010) 31–41.
[22] D. Daničić, T. Maneski, D. Ignjatović, Structural diagnostics and behaviour of bucket wheel excavator, Structural Integrity and Life 10 (1) (2010) 53–59.
[23] S. Bošnjak, Z. Petković, N. Zrnić, M. Pantelić, A. Obradović, Failure analysis and redesign of the bucket wheel excavator two-wheel bogie, Engineering Failure
Analysis 17 (2) (2010) 473–485.
[24] A. Petrović, T. Maneski, N. Trišović, D. Ignjatović, M. Dunjić, Identification of crack initiation cause in pylons construction of the excavator SchRs630, Technical
Gazette 25 (2) (2018) 486–491.
[25] B. Petrović, A. Petrović, D. Ignjatović, I. Grozdanović, D. Kozak, M. Katinić, Assessment of the maximum possible extension of bucket wheel SchRs740 boom
based on static and dynamic calculation, Technical Gazette 23 (4) (2016) 1233–1238.
[26] S. Bošnjak, N. Zrnić, Dynamics, failures, redesigning and environmentally friendly technologies in surface mining systems, Archives of Civil and Mechanical
Engineering 12 (3) (2012) 348–359.
[27] S. Bošnjak, M. Pantelić, N. Zrnić, N. Gnjatović, M. Đorđević, Failure analysis and reconstruction design of the slewing platform mantle of the bucket wheel
excavator O&K SchRs 630, Engineering Failure Analysis 18 (2) (2011) 658–669.
[28] S. Bošnjak, Z. Petković, P. Matejić, N. Zrnić, S. Petrić, A. Simonović, Analysis of Stress-Strain State of Bucket Wheel Excavator Revolving Platform Structure –
Fundament of Efficient Reconstruction, Structural Integrity and Life 5 (3) (2005) 129–142.
[29] S. Bošnjak, Z. Petković, N. Zrnić, G. Simić, A. Simonović, Cracks, repair and reconstruction of bucket wheel excavator slewing platform, Engineering Failure
Analysis 16 (5) (2009) 1631–1642.
[30] S. Bošnjak, S. Savićević, N. Gnjatović, I. Milenović, M. Pantelić, Disaster of the bucket wheel excavator caused by extreme environmental impact: Consequences,
rescue and reconstruction, Engineering Failure Analysis 56 (2015) 360–374.
[31] S. Bošnjak, Z. Petković, A. Simonović, N. Zrnić, N. Gnjatović, ‘Designing-in’ failures and redesign of bucket wheel excavator undercarriage, Engineering Failure
Analysis 35 (2013) 95–103.
[32] T. Maneski, D. Ignjatović, Repair and Reconstruction of Bucket Wheel Excavators, Structural Integrity and Life 4 (1) (2004) 3–7.
[33] T. Maneski, D. Ignjatović, Structural performance diagnostics, Structural Integrity and Life 4 (1) (2004) 3–7.
[34] S. Kirin, W. Li, M. Brzakovic, I. Miljanovic, A. Sedmak, Rules of risk management - Case study of open pit mine, Procedia Structural Integrity 28 (2020) 764–769.
[35] Hosmer and Lemeshow, Goodness of fit tests for the multiple logistic regression model, Communications in Statistics - Theory and Methods, 9(10) (1980) 1043-
1069.

13