Behavior analysis of factors affecting safety management to reduce incidents in the pelletizing industry using the system dynamics approach

The purpose of this study was to analyze the behavior of effective factors on safety management to reduce incidents in the pelletizing industry using the system dynamics approach. Risk assessment for the Sechahon pelletizing plant was performed using the FMEA method. A total of 625 risks were identified in this complex of which 286 are high risk RPNs and need to be scrutinized. These risks were categorized according to their nature and consequences in five physical categories, exposure to dust, ergonomics, emergencies and psychosocial risks, so that they can be reviewed in detail. The results showed that decreasing the number of incidents in the risk reduction scenario would reduce the incidence earlier, and is initially more than other policies. The scenario of risk reduction is the best scenario in the short term. The technology improvement scenario shows an incremental growth trend, but the scenario of the current situation can better decrease the number of incidents. The human resource management and safety management scenario will initially be on an upward trend and will continue to decline. The human resource management and safety management scenario is a good way to reduce disasters in the long run.


C. Research Literature
Amin Fathi Biranvand (2017) used the research system dynamic approach to provide appropriate and optimal ways to investigate the causes of incidents and incidents by using time management, cost, and quality method. In this research, the required data were collected using surveying studies. In this research, the required data were collected using field studies using statistical methods and appropriate statistical tests. Then, the dynamic model of the study was plotted and simulated using Vesnim software. Finally, an appropriate strategy for managing time, cost, and quality was presented in the study of the causes of incidents and construction incidents, and then, the issues that led to the reduction of incidents in the construction workshops were expressed. Rauf and Moazeni (2017) in a paper entitled, "Managing the safety of contractors in the control of incidents", discussed and evaluated the problems encountered by the contractors' safety and the reasons for the lack of attention of contractors to safety issues from the technical and economic viewpoints, the training and control of waste and analysis.
In the electricity company, 66 urban development and improvement projects in Asalouyeh have been considered as a case study, the results of which show that the role of the employer in supervising the contractor in various stages of the contract until the end of the implementation is undeniable. Also, the results of the research show that culturalization is essential as a first step in preventing incidents. Obviously, due to the lack of proper culture and safety precautions, even there were events with the purchase of expensive protective equipment. In a study by Chen Yang et al. (2018), a dynamic model was proposed to simulate the chlorine process safety management system, which includes four modules for workers, management, rules (regulations) and equipment, as well as an integrated junction with the degree of corrosion that is comprehensively integrated. By introducing a series of scenarios with different inputs, it was determined that the proposed dynamic model can obtain the penetration pattern in the degree of equipment corrosion. The results of this study showed that work skills have a positive impact on system safety, while work stress mainly affects system safety during the second half of the life cycle. Yeo et al. (2018) conducted a study and theory results showed that a significant senior commitment in terms of human resources and the allocation of costs and suitability of the security manager was a key to the implementation of the safety management system. More than incident and incident costs, improved organizational frameworks and increased ratings of security voting were identified as the main benefits of implementing a safety management system. At the same time, factors such as adequate resources, close coordination and high returns were the key challenges for implementing an effective safety management system in Hong Kong. Dabirian and Safar (2016) in a research entitled "Dynamic Modeling of Building Safety Management System Based on Correctional Measures of the Site", using the dynamical approach of the system examined the structure of management safety management in construction workshops using the dynamical approach of the system. The results of this study show that severity and number of incidents as well as losses considerably decrease by using learning from safety inspection, investigation of incidents and corrective actions based on them and he safety performance of construction projects are improved. This study shows corrective actions as a useful policy to prevent incidents. Mariani et al. (2015) in a paper entitled, " "The dynamics of the system for modeling construction incidents", analyzed the occupational incidents in the construction project due to the probability of the variables affecting the events of construction projects, using the system dynamics method. The model of this research covers the process of occupational incidents and direct and indirect costs. Finally, this model creates OSH cost components that need Industrial Engineering Letters www.iiste.org ISSN 2224-6096 (Paper) ISSN 2225-0581 (online) Vol.10, No.1, 2020 52 to be controlled, as well as improves the supply chain of contracts and supervisors to improve the quality of employees. The results of the research show that the variables of work discipline and work hazards have the greatest impact on work incidents. Therefore, contractors in the construction project fully understand what work and which part of the supply chain should be improved. In addition, showing direct and indirect costs and occupational safety and health costs will provide the benefits of project budget control. Therefore, the project's budget for occupational safety and health costs will not be more than planning.

A. Case study features
Sechahoon Bafgh Pelletizing Complex has been designed to produce 5 million tons of usable pellets in the iron and steel industry by direct reduction method to supply the required pellet in the steel industry. The site of the factory is located near Chaghart iron ore, located 13 kilometers north of Bafgh city, from Yazd province (115 kilometers from Yazd) in the central part of Iran with a longitude of 55028 and a latitude of 31042.

B. FMEA risk assessment
Risk assessment for the Sechahoon pelletizing plant was performed using the FMEA method. A total of 625 risks were identified in this series, of which 286 are high risk RPNs and need to be scrutinized. These risks were categorized according to their nature and consequences in five physical categories, exposure to dust, ergonomics, emergencies, and psychosocial risks, so that they can be reviewed in detail. Figure (1) shows the percentage of each sector's risks. As it turns out, the largest share is related to the physical risks as much as 95%. For this reason, physical risks were divided into various mechanical, electrical, vibration, radiation, sound, heating, and fire sectors. The percentage of each of these risks in Figure 2 shows that the share of mechanical risks exceeds other risks, and then the electrical risks have a higher percentage of high-risk and high RPN risks that need to be addressed more often.

A. Dynamic system model
The number of incidents is one of the problems for each set. Moreover, the number of incidents as well as the lost days of the events should be examined to enhance the safety of existing processes and the factors affecting it, in addition to the risks in order to minimize their management and planning, their damage and the crisis. Figure 3 and 4 show the number of incidents and lost days, respectively. For this purpose, the model of dynamical system was created with information about different months in 2018 and its status was examined in 2019 and based on which different policies were proposed. After executing the simulated model, its validity needs to be measured.

B. Validation of the model
After creating the flow chart of accumulation and simulation of the system and before using the model for analysis and scenario, one or more methods should be used to test the validity of the model. In this research, the validity of the model was investigated by several tests after simulation.

C. Border adequacy test
This test examines the important implications of the problem within the model. In this research, the proposed model has identified the key variables of the model after reviewing the literature. In addition, the necessity and importance of the variables mentioned by experts have also been studied. In response to the question of whether model behavior after the removal of border assumptions is a significant change, the results of the proposed model were studied after removing parts of the model and changing the boundary of the model. Figure 5 shows the graph of the effect of eliminating the "risk control measures" variable. Deleting this factor will reduce safety and increase incidents.

D. Structural Assessment Test
The purpose of the structure test is to determine the model's structure matching with the descriptive knowledge associated with the system and to examine the logic of the decision rules in shaping the behavior of the variables and the correctness of the structure of the model equations. Since the model equations in the software environment have been written in this paper, the correctness of the structure of the model equations was verified by the software (Figure 6).

E. Re-behavior test
One of the most important tests available is the reproduction of behavior (compared with historical data). In this test it is determined which model variables can rebuild the amount of historical data. Figure 7 compares the outputs of the model with the actual data of the past. The matching will ensure modeling results and validate its validity in future prediction.

F. Boundary conditions test
This test examines whether the model behaves appropriately when its inputs are in extreme conditions such as zero or infinity. In other words, in this test, the stability of the model is measured in extreme conditions. To examine this test, the variables "risk control measures", "safety" and "risk" were placed in their boundary condition, the results of which are shown in Figure 8.

G. Sensitivity analysis
In general, the purpose of the sensitivity analysis is to investigate whether changes in the parameters, boundaries, and time intervals result in significant changes in the numerical values, behavior, and policies observed or not.
After simulating and observing the behavior of all components of the model in the desired time period, the change in the variables of the model and their impact analysis on the main variable is investigated. Table 1 shows the results of sensitivity analysis for various variables.  Safety is sensitive and increases with time.
Cost of one working day 1000000 -0 The number of incidents is not sensitive.
Safety has a low sensitivity and increases with time.
Heating risks 1-0 The number of incidents is sensitive and increases with time.
Safety is not sensitive.
Mechanical risks 1-0 The number of incidents has high sensitivity and sensitivity increases with time.
Safety is not sensitive.
Electrical risks 1-0 The number of incidents has high sensitivity and sensitivity increases with time.
Safety is not sensitive.

Audio Risks 1-0 The number of incidents and safety is not sensitive
Vibration Risks 1-0 The number of incidents and safety is not sensitive

Scenario of maintaining the current situation
According to this scenario, the process of safety management is assumed to be in line with the past and without any change or adoption of a new policy. All variables are the same as those previously defined; the actual values of some parameters are as follows: Management measures related to safety are equal to 0.5 (on a scale of 0 to 1), safety measures related to equipment 0.6 (scale 0 to 1), environmental safety measures 0.6 (0 to 1 scale). If such a process continues, the number of incidents will end at the end of the 100th week to 365 incidents. Figure 9 shows the process of the number of events in this scenario.

Scenario of Human Resources Management
The purpose of this scenario is to increase the safety of the human resource so that with its development increases the number of incidents. The human resource does not have the potential to fully cover occupational safety, but it can take a considerable amount of attention. Generally, the increase in human-resource risk control measures in this scenario began after about two months from the start and the use of human resources is considered with more education and more work experience. The number of incidents in this scenario with the values for the scenario of maintaining the current state is compared in Figure 10.

Scenario of Safety Management
The scenario shows that the number of incidents can be reduced by monitoring the performance of safety measures, monitoring the performance of safety measures, enforcing rules, creating guidelines for activities, and establishing monitoring, and development of safe spaces. The results of using this scenario are shown in Figure 12. As can be seen from the figure, the number of events will increase initially with the implementation of this scenario as in previous scenarios, and then it will decrease.

Scenario of Risk reduction
In general, safety management scenarios are conceptually wide and diverse. However, the development of this scenario focuses on reducing the likelihood of incidents and increasing safety. The number of incidents will be reduced by creating these changes that have recently been made and reducing the risks, especially those that are more severe and referred to in the previous sections. Figure 13 shows the results of this scenario in comparison with previous scenarios. According to the figure, the greatest reduction in the number of incidents occurs with this scenario, and then it will increase.

I. Scenario analysis
The number of incidents in different scenarios is shown in Figure 13. In the risk reduction scenario, the reduction of incidents occurs earlier and is initially higher; however, increases again with time. This scenario can be effective in the short term. The scenario of technology improvement represents a slight upward trend, however, it is better than the scenario of maintaining the current situation and can improve reducing the number of incidents in the long run. Both of the human resource management and safety management scenarios are recommended as long-term effective methods. In either short or long term, two scenarios of risk reduction and safety management as two successive scenarios can be considered as the best scenarios in reducing the number of incidents. In the short term, the risk reduction scenario and in the long run the safety management scenario will have the best results in reducing the number of incidents.

V. Conclusion
In general, the results of this simulation indicated that reducing the number of incidents in the scenario of risk reduction is earlier, and is initially more than other policies. The scenario of risk reduction is the best scenario in the short term. The scenario of technology improvement shows an incremental growth trend, but the scenario of maintaining the current situation can better reduce the number of incidents. The scenario of human resource management and safety management scenario will initially be on an upward trend and will continue to decline.
The scenario of human resource management and safety management are a good method to reduce disasters in the long run.
The safety management scenario is better than the scenario of human resources management. The best method to reduce the number of incidents is to successive use of two risk reduction and safety management scenarios. In the short term, the risk reduction scenario and in the long run the safety management scenario will have the best results in reducing the number of incidents.
With regard to the findings and limitations, it is recommended to evaluate the effect of other effective factors on the problem. The cost of running any of the scenarios can be an important issue for managers and its calculation can be helpful. Considering that electrical and mechanical risks have the highest share in increasing the probability of occurrence of incidents, the risk reduction scenario can reduce the number of incidents and be used as an optimal scenario in the short run. It is better to pay much attention to reform electrical and mechanical risks. The methods used to reform the risks will be effective when it reduces the number of incidents. Therefore, it is better to consider these methods more accurately.