Condition-based Maintenance Method for Multi-Component System Based on RUL Prediction: Subsea Tree System As a Case Study

To maintain process safety, condition-based maintenance (CBM) has arouse wide attention as the last part of diagnosis-prediction-maintenance in prognostics and health management. For a multi-component system, perfect maintenance is preferred to be adopted. However, because of the limitations of the special working environment, maintenance technology, or technical level of maintenance personnel, perfect maintenance cannot be realized. Generally, imperfect maintenance is considered and applied in practical engineering application. Reasonable spare part management for the multi-component system need to be scheduled to reduce maintenance preparation time and cost. A novel CBM method based on remaining useful life (RUL) for the multi-component is proposed considering imperfect maintenance and spare part management. Imperfect maintenance model is constructed by adopting virtual age rule. Quantities of spare parts are determined according to practical maintenance requirements, and ordering time and ordering type of spare parts are determined based on prediction results of RUL. In this way, total numbers of spare parts and maintenance preparation cost can be significantly reduced. Maintenance preparation threshold, spare part ordering threshold and total spare parts threshold are three maintenance decision variables that need to be optimized based on genetic algorithm. Subsea tree system for offshore oil and gas development is used to demonstrate the application of the proposed method.