A Fuzzy Inference System to Improve Maintenance Policies- Case Study: Centrifugal Pumps in Onshore Industry

A Fuzzy Inference System to Improve Maintenance Policies- Case Study: Centrifugal Pumps in Onshore Industry

The goal of this study is to develop a fuzzy inference system for pump diagnosis to improve the maintenance policies of centrifugal pumps. In order to achieve the objective, for knowledge acquisition, first the equipment conditions and operational data of the centrifugal pumps and theP&ID of the plant are simulated by ASPEN HYSYS (a chemical-process simulation package). Then according to the OREDA handbook classification the most critical failure modes of the centrifugal pumps and the relevant maintainable items to these failure modes are considered. Next, based on the mechanical trouble shootings, interpretation of the domain experts & the process simulation, through a Failure Modes & Effects Analysis (FMEA), the impact of these failure modes on the operational conditions of the centrifugal pumps are categorized. Then based on the obtained information, the linguistic rules are extracted. And finally, the rules are passed to a Mamdani-type inference engine The verification of the proposed approach is based on the previous failure and maintenance reports.