Implementing the Analytical Network Process (ANP) in Ranking the Safety and Occupational Health Risks of Cement Plants according to the Combination of the FMEA and William and Fine Methods.

Risk assessment is a logical method to determine the quality and quantity of risks and investigate the potential consequences of probable accidents on people, materials, equipment, and environments. Indeed, the process provides valuable data to make decisions concerning the reduction of risks and threats, the optimization of control systems, and the planning to respond to them. Thus, the present study aimed to determine and estimate, control, and reduce the rates of risks by evaluating and managing the safety and health risks of cement plants using the analytical network process (ANP).  The present study used 2 ANP models, and the pairwise comparisons of the risks were created in each set using the ANP process. Then, the results of the comparisons were handed to a panel of experts that consisted of 10 people. After the matrices of pairwise comparisons were completed, the incompatibility rates were calculated. As the obtained values were all below 0.1, the stability and compatibility of the matrices were confirmed. Then, the pairwise comparisons of the experts were combined using the geometric mean technique, and finally, they were entered into SuperDecision so that their weights could be determined.

The present study found a total of 1184 risks that threatened the safety of the plants and the employees’ health. Evaluating the health risks led to the detection of 352 risks, and the most significant risks were the noise produced by grinding the cement (0.0607), the dust produced by the grinding process (0.0597), and the thermal stress of the baking department (0.0596), respectively. Moreover, evaluating the safety risks resulted in the detection of 529 risks, and the most significant ones were related to falling from the cement mill (0.0601), the fall of items from the packing plant (0.056), and explosions in the baking department (0.0549), respectively. The proposed method in the present study could both distinguish risks more accurately and determine the degree of their relative importance according to their risks degrees.