A Fuzzy Approach to the Study of Human Reliability in the Petroleum Industry

This work presents a methodology for characterization of human reliability based on fuzzy sets concepts, aiming at reducing the possibility of human errors in oil refineries. It is suited for operation, maintenance and inspection activities in oil production and distribution units and is based on the API-770 guide for reduction of human errors, which identifies 64 performance factors. In order to assess the possibility of a human fault, the model analyses the elements that interact with each operator. It is possible, by this methodology, to obtain a human reliability index, to find out the problems that may constitute causes of human errors and to devise strategies for the control of potentially adverse impacts of interactions that add uncertainty and complexity to processes.