A dynamic programming-based maintenance model of offshore wind turbine considering logistic delay and weather condition

There are several challenges toward maintaining offshore wind turbines such as ensuring the intrinsic performance of the wind turbine, maximizing short- and long-term profits, and optimizing maintenance grouping. These criteria should be optimized in respect to several practical constraints. Moreover, issues such as decisionmaking in uncertain environments where reliability-oriented field data are insufficient and information regarding the state of health state also be tackled.
This paper proposes a failure model that includes multiple failure modes with different failure consequences. Logistic delays and weather conditions are also obstacles for performing maintenance activities. To save logistic cost and downtime loss, a two-level maintenance policy considering both periodic maintenance planning and reactive maintenance is developed. The first level is aimed at failure modes whose degradation process is blackbox by grouping individual maintenance activities according to their own maintenance period. The second level is aimed at those failure modes whose failure can be indicated through two options: 1) Integration with existing maintenance tasks of the first level and 2) creating a new maintenance opportunity. The effects of decision thresholds on the choice of maintenance options are studied through simulations.