Blade-based maintenance policy of offshore wind turbine with the presence of covariate under random shock
Résumé
The blade is the most expensive part of a wind turbine whose failure could be hazardous. The damage of blade is motivated by a mechanism of fatigue and affected by random factors. The technique and price of remote surveillance equipment of rotor blade are becoming well proven and affordable, which makes the monitoring of blade possible. In this paper we model the crack of blade via Paris- Erdogan law considering the intrinsic randomness of the fatigue crack propagation, random shock, random load and dynamic environment. We propose two reactive maintenance policies taking advantage of the monitoring signal as well as obeying the periodic maintenance. Between two periodic maintenances if the control signal exceeds a threshold or a shock occurs, a reactive maintenance will occur. Each maintenance action has to choose one of the two actions: imperfect repair and replacement. The numerical simulations prove that the proposed policy can reduce cost.