Wind power generation technology is a clean energy power generation technology with broad prospects and great development potential. Because repairing wind turbines requires a lot of cost. Therefore, it is of great significance to find an effective maintenance strategy. Predictive maintenance is to prevent the system from malfunctioning by checking the equipment system or replacing it with a new one at a fixed time. In wind farms, traditional predictive maintenance is based on time. This method has a definite inspection rate, which allows maintenance to be planned, but it cannot accurately reflect the true state of the system, and it usually results in overmaintenance or under-maintenance. This paper combines the operating principles of wind turbines, establishes a generalized stochastic differential equation model based on the traditional ordinary differential equation model, and simulates and compares time-based preventive maintenance and conditionbased predictive maintenance based on actual operating data.

预测性维修,基于时间的预测性维修,风电机组,基于状态的预测性维修