Improvement of risk assessment in the FMEA using nonlinear model, revised fuzzy TOPSIS, and support vector machine

In every organization, performing accurate risk assessment along with consideration of increasing accidents is a necessary tool to prevent and reduce the fatal and non-fatal consequences of their occurrence. One of the most popular methods of risk assessment is Failure Mode and Effects Analysis, which evaluate failure modes in a system by using risk priority numbers (RPNs). These methods have been criticized for including several deficiencies such as the effect of personal 'opinions, the same importance of the factors and risk rating. The present work utilizes a hybrid approach based on support vector machine and fuzzy inference system to decrease the effect of personal's opinions in determining the factors of the severity and occurrence. Also, Logarithmic Fuzzy Preference Programming is used to determine the crisp weight of dependent factor of FMEA and revised fuzzy TOPSIS used for more accurate ranking of risks. One main feature of the proposed model is that it can be used to evaluate safety risks in all organizations. To investigate the suitability of this approach, the proposed model was presented in the Copper leaching factory, Kerman, Iran. The results showed that this model has the ability to predict severity and occurrence refers to occupational accidents which occurred in a 5-year period (2012-2017) with accuracy of 87% and 95%, respectively. Also, based on the results, it was found that the weights of severity, occurrence, and detection were 0.479, 0.335, and 0.186, respectively. Results of the ranking process showed that the risk of fall from height and stucking between the objects had the highest and the lowest priority, respectively.