Application of Bayesian networks in corrective maintenance safety

The emphasis of this exploratory paper is on some maintenance related accidents occurred during previous decades and the aim is to develop a model for preventing such accidents by understanding the major causes and the influencing factors among them. The model is intended to explain the affecting factors in a fatal accident and how they are related to one another. To fulfil the aim of study, a Bayesian network model was developed using the GeNIe 2.0 software and further on the relations between different factors were assigned. The model was developed based on a case study of Tosco Avon refinery accident in Martinez, California occurred on February 23rd, 1999 where four workers were killed in a fiery accident. The proposed model could predict the final result of the execution of a corrective maintenance process to make a safer decision. Since the model was developed based on Tosco Avon refinery accident case, validation has been carried out regarding some other accidents. The results showed that the proposed model is a suitable tool to predict the outcomes of an action in corrective maintenance. In addition the model can be adjusted to different applications.