A prognostic driven dynamic predictive maintenance decision-making model for offshore wind turbine systems

The existing researches on predictive maintenance for offshore wind turbine systems primarily focus on optimizing maintenance costs over a long term, lacking a more detailed assessment of the balance between utilizing the remaining useful life value of degraded components and optimizing maintenance costs. Moreover, the current methods on spare parts management and maintenance scheduling for wind turbines have poor dynamic response characteristics and cannot effectively utilize the prognostic information. To achieve the optimized health management of wind turbine systems, this paper explores a prognostic driven dynamic predictive maintenance decision-making method. Considering two typical health management activities of wind turbine systems, an individual maintenance decision-making model and multi-component maintenance decision-making model are constructed, where two new expected net operation and maintenance cost functions are defined. Based on these two models and the dynamical prognostic information, a system-level dynamic maintenance decision-making model is established to obtain the optimized maintenance schedule of wind turbine systems. Finally, the effectiveness and superiority of the proposed method are validated through an actual offshore wind turbine dataset and several simulation cases.