Remaining Useful Life Prediction Method for Multi-Component System Considering Maintenance: Subsea Christmas Tree System as A Case Study

Maintenance is an important technical measure to maintain and restore the performance status of equipment and ensure the safety of the production process in industrial production, and is an indispensable part of prediction and health management. However, most of the existing remaining useful life (RUL) prediction methods assume that there is no maintenance or only perfect maintenance during the whole life cycle; thus, the predicted RUL value of the system is obviously lower than its actual operating value. The complex environment of the system further increases the difficulty of maintenance, and its maintenance nodes and maintenance degree are limited by the construction period and working conditions, which increases the difficulty of RUL prediction. An RUL prediction method for a multi-omponent system based on the Wiener process considering maintenance is proposed. The performance degradation model of components is established by a dynamic Bayesian network as the initial model, which solves the uncertainty of insufficient data problems. Based on the experience of experts, the degree of degradation is divided according to Poisson process simulation random failure, and different maintenance strategies are used to estimate a variety of condition maintenance factors. An example of a subsea tree system is given to verify the effectiveness of the proposed method.