"Predictive maintenance and repair policies in oil and gas processing centers"

An appropriate maintenance policy says that repairs should be done when needed. Although preventive repairs can reduce sudden and unexpected repairs, they still reduce availability and increase repair costs. Companies need to develop online and predictive maintenance strategies that can anticipate that any failure could occur at any time, and recognize this need from the signs and symptoms of equipment. This is called predictive maintenance or condition-based maintenance.in this paper we have tried to design a decision support model for predictive maintenance based on conditions based on data mining techniques. This project has been carried out in one of the oil and gas exploitation companies in the south of the country and the selected equipment for this project is gas turbines, which is one of the most basic and critical equipments in oil processing factories.In this project, MPL neural network data mining technique has been used to predict the occurrence of failure in the equipment. Finally, suggestions such as the development of this model for other equipment, controlling the duration of viewing the equipment status and determining the optimal maintenance time for the future are presented.