Development of a novel Electrical Industry Safety Risk Index (EISRI) in the electricity power distribution industry based on fuzzy analytic hierarchy process (FAHP)

Many workers are exposed to electrical energy during the fulfillment of their tasks. It is necessary to identify the potential risk factors for electrical damages. The present study aimed to develop a novel Electrical Industry Safety Risk Index (EISRI) in the electricity power distribution industry based on fuzzy analytic hierarchy process (FAHP). In this study several different safety risk assessment methods were analyzed. Then, common activities in the electricity distribution industry were classified into ten occupational groups. To identify the general structure of risk assessment and determine three main components, including personal, environmental, and organizational a three-stage Delphi study was conducted with the participation of 30 experts. The fuzzy analytic hierarchy process approach was used to weight the components and parameters in each job group. Finally, the results of the EISRI were compared with the failure mode and effect analysis (FMEA) method. The most effective component in determining the risk level was the personal component (PC), with a 0.537 weighted average. Cronbach's alpha values for each of the personal, environmental, and organizational components and the entire model were 0.90, 0.85, 0.82, and 0.86, respectively, and model reliability was confirmed. The results obtained from the EISRI method were compared with the FMEA method, the results of both methods were very close to each other (p < 0.05). The results of this study revealed that the highest weighted average was related to the personal component due to the high impact of the human factors in carrying out activities in various occupations. The EISRI can be applied as a substitute for general risk assessment methods due to the suitability of this method with the nature of activities in this industry. The present technique can be a practical step toward developing suitable risk management algorithm.

Many workers are exposed to electrical energy during the fulfillment of their tasks. It is necessary to identify the potential risk factors for electrical damages. The present study aimed to develop a novel Electrical Industry Safety Risk Index (EISRI) in the electricity power distribution industry based on fuzzy analytic hierarchy process (FAHP). In this study several different safety risk assessment methods were analyzed. Then, common activities in the electricity distribution industry were classified into ten occupational groups. To identify the general structure of risk assessment and determine three main components, including personal, environmental, and organizational a three-stage Delphi study was conducted with the participation of 30 experts. The fuzzy analytic hierarchy process approach was used to weight the components and parameters in each job group. Finally, the results of the EISRI were compared with the failure mode and effect analysis (FMEA) method. The most effective component in determining the risk level was the personal component (PC), with a 0.537 weighted average. Cronbach's alpha values for each of the personal, environmental, and organizational components and the entire model were 0.90, 0.85, 0.82, and 0.86, respectively, and model reliability was confirmed. The results obtained from the EISRI method were compared with the FMEA method, the results of both methods were very close to each other (p < 0.05). The results of this study revealed that the highest weighted average was related to the personal component due to the high impact of the human factors in carrying out activities in various occupations. The EISRI can be applied as a substitute for general risk assessment methods due to the suitability of this method with the nature of activities in this process (FAHP).

Method
This cross-sectional and practical study was conducted in 2022 using a semi-quantitative safety risk assessment technique in Gilan Power Distribution Company based on the fuzzy analytic hierarchy process (FAHP). This study was conducted according to the following steps.

The first phase
In this phase, FMEA (failure mode and effect analysis) and JSA (job safety analysis) methods were used to identify and primary assessment of the risks of the electricity distribution industry. A panel of experts grouped the jobs to determine the risks of different task and sub-tasks in the electricity distribution industry and, subsequently, more accurate risk assessment.
Before conducting the study, an informed consent form was obtained from the participants to participate in the study. Participants were able to leave the study at any stage of the research if they were not satisfied. In this study, no intervention or study was done on human or animal samples. This study were conducted according to established ethical guidelines of the electricity distribution company of Gilan province.

Delphi study
To prepare the list of measured parameters according to the industry characteristics the Delphi method was utilized. The Delphi method is a type of regular communication technique developed at first for forecasting according to the experts opinions. This methodology is a process for collecting the knowledge and wisdom existing to a group of experts, which is performed by distributing questionnaires among these population and giving response on the answers received. Participants in the Delphi study included 5 to 20 experts [15].
In the present study, a Delphi questionnaire was designed after determining the main parameters affecting the risk index. To capture the majority of the country's specialists cooperation, the opinions of 30 experts, including Ph.D. and MS graduates in the fields of occupational safety and health engineering, industrial management, and electrical engineering, employed in 25 universities and 15 large power plants and manufacturing industries, in three phase of Delphi study were gathered.
In the first phase of the Delphi study, experts were requested to comment on the method's overall structure. In the second phase, experts were requested to ranking components and parameters according to their importance. Then, to determine the importance of each of the criteria and sub-criteria in each of the job subgroups and to select the weight of each parameter, the analytic hierarchy process (AHP) was applied.

Content validity
Content validity ratios (CVR) and content validity index (CVI) were used to evaluate the content validity quantitatively. In this regard, experts were asked to review each item based on a three-part range of "necessary," "useful but not necessary," and "not necessary" to determine CVR. The responses were calculated according to Eq. (1) [23].
where, nE is the number of specialists who have answered the "necessary" option, and N is the total number of specialists. The CVI index was also calculated by Eq. (2) [24].
Where, ∑ 1 n CVR is the sum of the approving points for each item that scored " relevant but need review" and "completely relevant," and R is the total number of specialists. After obtaining the tables from the experts, the CVR and CVI were calculated for each item and the items that did not get the minimum required score, were removed from the tables. Content validity ratio (CVR) and content validity index (CVI) were used to determine the content validity of the model. The acceptable limit value of the CVR was considered 0.33 according to the Lawshe table and proportional to the number of participants in the Delphi study (30 experts) As shown in Table 1, the minimum validity score was calculated according to the number of expert panel [25]. Also, the acceptable limit value of the CVI was considered to be 0.79 [26].

Reliability
Cronbach's alpha method was used to evaluate the reliability of the new method [27]. An alpha coefficient of 0.7 or higher was considered as the minimum score required to confirm the model's reliability.
In this stage of the study, after determination of the general structure of risk assessment, the most effective causes of accidents were classified into three groups, including personal component, environmental component, and organizational component according to the results of previous studies and investigation of the accident database of electricity distribution industry in the last ten years (from 2012 to 2022).
At the end of this phase, after holding two sessions with panel members and using the information extracted from previous studies, the jobs in electricity distribution companies were divided into ten groups according to the nature of the job, stages of work, and similarity of risks, as described in Table 2. After job grouping, the hierarchical task analysis (HTA) and tabular task analysis (TTA) methods was used for the job analysis [20,22]. In order to use the participatory approach and also to identify all existing tasks and sub-tasks, after three training course, tabular task analysis (TTA) was also completed by the workers of the studied industry.

Table 2
Grouping of main job activities (Tasks) related to each job.  Thomas L. Saaty presented this methodology in 1983. The aim of this method is to prioritize multiple criteria. The Analytic Hierarchy Process is a multi-criteria decision-making tool for weighting the parameters and choosing the optimal option. These criteria are paired based on study goals, and related weight is specified. Ultimately, the options are paired comparisons based on each criterion [28]. The main purpose of the hierarchical analysis process method is to opt the best option based on various criteria by developing a pairwise comparison matrix [29].
In the hierarchical process, the parameters of any level are compared in pairs with their respective parameters, and related weights are calculated (relative weights). In the Next phase, the final weight of each parameters is determined, which is called the absolute weight. In these comparisons, decision-makers will applied verbal judgments [15]. The comparison and weighing of factors are performed in a K × K matrix (K = number of rows and columns of the pairwise comparisons matrix). The pairwise comparison is performed according to the valuation of the row factor relevant to the column element. For valuation, an interval scale from 1 to 9 is usually utilize based on language phrases: a higher value shows the preference of a row element over a column element so that a value of 9 means the most valuable element, and a value of 1 means the least valuable element.
To enhancement the reliability of the findings of the questionnaires analysis, the consistent rate of the system is controlled. The reliability values of the expert panel questionnaire were considered the same as the adjustment rate. In this study, a value of 0.1 or less was considered as the acceptable compatibility limit of pairwise comparisons.

Fuzzy logic
Fuzzy logic is a type of the multi-valued area in which the creteria's accurate values may be any real number between zero and one. This logic is use to apply the concept of partial correctness so that its values can be between fully accurate and fully false [30]. This method is a strong tool to deal with the uncertainty of human judgment in decision-making.
Different studies have used combining the AHP method with fuzzy logic to weight the criteria and sub-criteria. Various tools for performing FAHP have been suggested. This study was based on the method suggested by Chang because it is more comfortable to coduct than other approaches and supplies accurate outputs [31].
The Likert scoring method was used for the users scoring the sub-criteria during the activity (Table 3). With this range or scale, the respondents determine their level of agreement with the item. These levels are sequential and show the lowest to the highest agreement levels.

Formation of the risk matrix
The next step is to form a risk level matrix using the results obtained from the previous steps for each occupational group in the power distribution industry. The risk matrix is obtained by combining different levels of personal component, environmental component, and organizational component of varying levels of risk.
The risk matrix was divided into three levels by the ALARP (as low as reasonably practicable) principle and main criteria, including acceptable risks, tolerable risks, and unacceptable risks. ALARP" is short for "as low as reasonably practicable". Reasonably practicable involves weighing a risk against the trouble, time, and money needed to control it. Thus, ALARP describes the level to which we expect to see workplace risks controlled. The ALARP concept arises within a regulatory framework. Increasingly, it is used by companies worldwide as it provides a reasonable basis for managing risks [32].
For this purpose, instead of forming a general risk matrix for all occupations, a separate risk matrix was drawn in each occupational group and used to determine the three levels of the risks according to experts' opinions. After receiving the final information, ten risk matrices with three levels were designed for occupational groups.

The third phase
In this phase of the study, in each job group, two most frequent tasks were evaluated using both the FMEA and the new method in Gilan Power Distribution Company to assess the efficiency and validity of the new method by caparison of the results.
Some occupational group activities such as the electrician with climbing height task, a technical with a panel-loading task, service personnel with the general cleaning task, the administrative expert during work on computer, and driving were selected for risk assessment using the FMEA method. Each activity was observed and evaluated in three different periods, then the average of the results of three observations was considered as the final score. After determining the hazards, the possible consequences and their root causes were identified. Then, the risk priority number (RPN) was determined based on the intensity, repeatability, and risk detection Table 3 Calculation of the Likert spectrum to assign a risk rating in EISRI.
Finally, in order to provide descriptive indicators, average, standard deviation and frequency statistics were used. In order to check the relationship between the risk levels of the new method and the FMEA method, chi-square test was used at a significance level of 0.05. All analyzes were performed in the environment of SPSS software version 25.

Results
A total of 30 experts were involved in the present study. The expert panel's mean age and work experience were 41.63 ± 8.09 and 8.12 ± 5.13 years, respectively. 51% of the experts were Ph.D., and 49% had an MS degree.
As shown in Table 4, by conducting an initial risk assessment, the most important causes of accidents were identified and categorized into three groups: personal, environmental, and organizational components.
The results of CVR and CVI for parameters affecting personal, environmental, and organizational components are presented in Table 5. In this table, 16 effective factors are listed that have an acceptable result for their content validity (see Table 6).
Cronbach's alpha values for each of the personal, environmental, and organizational components and the entire model were 0.90, 0.85, 0.82, and 0.86, respectively, and model reliability was confirmed.

Electical Industry Safety Risk Index (EISRI)
The numerical value of Electical Industry Safety Risk Index (EISRI) was estimated using Eq. (3): where PC is the personal component, EC is the environment component, and OC is the organizational component. Equations No. 4, 5, and 6 were used to obtain each of three main components. To calculate the effective weight of the sub-criteria, the score of the main criterion was multiplied by the weight of the relevant sub-criterion. Then by summing the effective weight of the sub-criteria, the weight of the main criteria was calculated.
W PC = P 1 + P 2 + P 3 + P 4 + P 5 (4) There are risk identification and risk assessment programs 4. Design, planning, organized responsibilities and supervision exist 5. Lack of concern for safety and the tendency of supervisors and the organization to time and economic issues 6. Monitoring, inspection, and auditing exist where P is sub-criteria of personal component, E is sub-criteria of environmental component, M is sub-criteria of organizational component, W PC is the weight of the personal component, W EC is the weight of the environmental component, and W OC is the weight of organizational component, and the score is the weight of the main criterion. The schematic of the new risk assessment method is shown in Fig. 1.

Risk matrix
The risk matrix was divided into three levels by the ALARP (as low as reasonably practicable) principle and main criteria, including acceptable risks, tolerable risks, and unacceptable risks. For maximum accuracy in estimating the risk level, instead of forming a general risk matrix for all occupations, a separate risk matrix was drawn in each occupational group. For instance, the electrician matrix is shown in Table 7, and the other matrix tables are in the appendix (see Tables S1-S10).

Analytical results
The results revealed that the most important main component in determining the level of risk in all ten occupational groups was the PC, with an average weight of 0.537.
In personal component, the average weight of the most effective sub-criteria were according to following. The results obtained from the developed risk assessment method were compared with the FMEA method and presented, the results of both methods were very close to each other (p < 0.05).

Discussion
The electricity distribution industry is one of the most important industries in the world, and many accidents occur in this industry every year, which causes many human and financial damages to various organizations and societies. Therefore, formulation, development and application of a safety index evaluation method in different parts of this industry and according to the nature of activities in this industry, for the implementation of corrective and control measures, is of great importance.
In the present study, the jobs in electricity distribution companies were divided into ten groups according to the nature of the job and related activities. This study also considered several criteria in job grouping, including similarity of job activities, the risk of the jobs, the importance of jobs, degree of involvement in activities of the company, and the experts' comments. Then, the tasks of each occupational group were categorized, and the risks and possible events were determined using the Hierarchical Task Analysis (HTA) and tabular task analysis (TTA) methods. The Delphi method and the fuzzy analytic hierarchy process (FAHP) approach were used to determine the criteria and sub-criteria affecting the level of risk and related weighting.
Previous studies have also illustrates that the use of the Delphi study approach and the fuzzy analytic hierarchy process (FAHP) method can be an effectual step in developing semi-quantitative and quantitative risk assessment methodologies based on the characteristics of different work environment situations [15]. A study performed by Pinto et al. showed that health and safety risk assessment is a complex task for employees and requires considering multiple and appropriate parameters for each job [33].
The findings of Nasrollah Nejhada et al. study in an electricity distribution company of Iran revealed that the majority of the accidents occurred in summer. Workers' negligence was the cause of 75% of deaths. Also, their findings indicated that occupational injuries are going to have an increase in the future [34]. These findings show the need to use risk assessment methods appropriate to the nature of this industry.
The present study showed that the most influential criterion in the risk level is personal component (PC) in all ten occupational groups. The reason for this may be the nature of the job in electricity distribution companies and the important role of personal behaviors in the risk occurrence. By studying work-related accidents, it can be concluded that more than 90% of accidents are due to   human behavior and error, which confirms the results of the present study [35]. The present study indicated that the most influential factors in the occurrence of accidents.
1. The personal component includes working individually, not using personal and group protection equipment, haste while doing work, and lack of necessary knowledge and skills to perform activities. 2. The environmental component includes dangerous work situations and harmful environmental factors, worn out and old equipment and facilities, unfavorable work atmosphere, and complex and unfamiliar equipment and facilities. 3. The organizational component includes excessive pressure on workers to perform their duties, lack of risk identification and effective risk assessment program, lack of supervision and organized tasks and responsibilities, lack of safety concerns, the tendency of managers to economic issues, and the lack of adequate and regular inspections and audits.
The results of this study are consistent with the results of some studies about the risks of low-voltage networks [35,36]. The results of Tiju Baby et al.'s study demonstrated that personal factors, safety climate factors, and health conditions of the workers were significantly associate to occupational electrical accidents. In personal component, age, job role, education, and experience are significantly effecting the safety behaviour of the electrical workers. The results of the Baby et al.'s study highlighted the need for interventions to reduce personal parametes in the electrical industry [5]. In the current study, the highest weighted average was related to the personal component, which can be an effective step in highlighting the role of this component in the risk score and finally implementing control measures.
The results revealed that the most important main component in determining the level of risk in all ten occupational groups was the PC, with an average weight of 0.537.
In the present study, after weighing the criteria and sub-criteria through pairwise comparisons, experts' opinions were quantitatively summed, and the data were normalized. The results showed that the weight of haste while working in the job groups of the electrician, technical expert, driver, dispatching technician, service, and warehouse were 0.27, 0.12, 0.27, 0.15, 0.13 and 0.18, respectively and has the most significant effect in the personal component. The reason for this issue can be considered as having high concentration as a job requirement in the mentioned jobs. When people are in a hurry, their attention is affected, significantly increasing the risk of hazards.
The sub-criterion of physical competence to work in the job group of the administrative employee with an average weight of 0.14 has the most significant impact on the risk level in this job group. Office employees spend most of their time sitting and working with a computer. Suppose the ergonomic principles are not observed or the work station is not appropriate, including the fit of the desk and chair with the person. In that case, it can lead to musculoskeletal disorders over time. One of the essential factors in creating a fit for the person in the work environment is being in the standard and ergonomic design range. If the ergonomic principles is not observed, the complications will appear over time. Previous studies conducted to determine the risk factors associated with the prevalence of musculoskeletal disorders, revealed that the incompatibility of the person and the work station leads to unfavorable ergonomic conditions and increased risk of ergonomic disorders [37,38]. Also, in a study conducted by Suri et al. to investigate the relationship between quality of life and the incidence of musculoskeletal disorders in car assemblers, it was found that the physical condition of individuals is effective in the occurrence of ergonomic risks [39].
The sub-criterion of personal and group protection equipment in the occupational group of technical technicians with an average weight of 0.21 was determined as the most effective sub-criterion in the risk occurrence. This job group has good knowledge and skills due to specialized job activities. On the other hand, the workload in this job group is usually pre-determined, and sub-criteria such as haste during work have negligible effects on the risk level. In this occupational group, personal and group protection equipment as the main barrier between risk and individuals is an essential factor in preventing accidents.
The sub-criterion of barriers and personal and group protection equipment in the printing and copying group with an average weight of 0.15 was identified as the most effective sub-criterion in risk occurrence. In this occupational group, exposure to harmful chemical factors such as gases and vapors is possible during paper copying. They should use personal protective equipment such as respirators, gloves, and general ventilation to prevent health risks. Performing work individually in the security group with an average weight of 0.15 was determined as the most effective sub-criterion in the risk of accidents.
Ultimately, the results of the new risk assessment method were similar to the results obtained from the FMEA, which shows the reliability of the results obtained for specialized risk assessment in the electricity industry.

Limitations and strengths of the study
Among the limitations of the current method is the lack of a quantitative approach to calculate and evaluate the effective parameters of the electricity distribution industry's probability and severity of risks. Thus, it is suggested that researchers in the future develop and apply quantitative tools according to the conditions of using safety management systems via international management guidelines.
Another limitations of this study was not considering some of the activities done in different ways in different electricity distribution companies in the country. It is necessary to continue this study with a full review of jobs in this industry and develop this risk assessment approach for all activities in electricity companies all over the world.
The present study was performed for the first time to introduce and implement a unique approach to assess electricity distribution industry safety risk and accordance with the particular specification of this industries and activities. Among the strengths of this study are the risk assessment at the place of activity and the presentation of controls immediately to take rapid measures to reduce the risk level. Also, control measures are presented separately in three personal, environmental, and organizational components, making the rules more effective and easier to implement.
Another Strength of the present study is forming a specific risk matrix for each job, considering the nature of the job and the importance and severity of accidents in different occupations, ultimately leading to a more accurate estimate of the risk level, which differs from many existing safety risk estimation methods.
This study determined the most important factors affecting occupational accident frequency and severity. The present technique can be a practical step toward decreasing occupational accident risk levels in the electricity distribution industry and developing control plans, especially in developing countries, due to low safety performance and related inspections.

Conclusion
The present method was developed under the title of "Electical Industry Safety Risk Index (EISRI)" based on three personal, environmental and organizational components and using 16 sub-criteria or parameters using the fuzzy analytic hierarchy process (FAHP) approach. The results of the present study showed that the highest weighted average was related to the personal component due to the high impact of the human factors in carrying out activities in various occupations. The EISRI can be used as a substitute for general risk assessment methods due to the suitability of this method with the nature of activities in this industry. The present technique can be a practical step toward decreasing occupational accident risk levels in the electricity distribution industry and developing control plans, especially in developing countries, due to lower risk management performance.

Credit author statement
Mohsen Sadeghi-Yarandi: Conceived and designed the experiments; Wrote the paper.Salman Torabi-Gudarzi: Performed the experiments; Contributed reagents, materials, analysis tools or data; Wrote the paper.Nasrin Asadi, Hamedeh Golmohammadpour: Analyzed and interpreted the data.Vahid Ahmadi-Moshiran: Performed the experiments; Analyzed and interpreted the data.Mostafa Taheri, Aysa Ghasemi-Koozekonan: Performed the experiments.Ahmad Soltanzadeh: Contributed reagents, materials, analysis tools or data.Bahare Alimohammadi: Conceived and designed the experiments.

Funding statement
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability statement
Data included in article/supp. Material/referenced in article.

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.