ABSTRACT
The energy produced by a wind turbine is directly proportional to wind turbine availability and wind farm benefits can increase with an appropriate management of preventive and corrective maintenance. The operational conditions of wind farms are quite different from traditional power stations. Thus, a specific feature of wind power generation is the stochastic behavior of wind velocity, what determines the energy produced, and also influences the turbine degradation process due to the stochastic load suffered by the wind turbine. Nowadays, the maintenance plan scheduled in wind turbine farms includes routine checks to minimize the degradation effects. Moreover, the maintenance plan is determined by the meteorological conditions. In this context, this paper presents a specific methodology to manage maintenance planning, which selects the optimal preventive maintenance taking into consideration the weather conditions of wind, the effect of maintenance activities on the state of the wind turbine, the actual energy produced and the costs associated with maintenance and loss of power. The methodology is based on simulation Monte Carlo (MC) and Genetic Algorithms (AAGG) to perform a multi-objective optimization of maintenance intervals using as decision criteria the capacity utilization ratio and the maintenance cost.
Finally, an application case is performed to optimize the maintenance plan of a wind turbine of 2 MW. A generic failure database has been used. This database contains information about, failure date, failure description and duration of corrective and preventive maintenance. Daily wind velocity data have been obtained from a Spanish database to determine the wind behavior.
