Case Study of Weather Maintenance in Wind Power Generation

Present paper shows a new procedure to obtain control algorithms based on weather conditions to fault forecast in a real wind farm during a year. In an initial study, results showed a correlation between central hydraulic level and gearbox level and asymmetric currents and the emergency stop. A second statistical study of one-way analysis of variance was selected to define groups of errors under the same weather conditions. In consequence, a clear relationship between groups of maintenance problems and weather conditions under which they happens was obtained. It was obtained that the more important weather variables for maintenance were the mean outdoor temperature and the wind velocity. Mean outdoor temperature was related with freezing of anemometers and excessive orienting time faults. On the other hand, brake shoe temperature, emergency stop, low voltage and asymmetric currents were related with wind velocity and wind direction faults. Finally, relative humidity resulted as the worst weather parameter to predict maintenance problems. Furthermore, the average values of each weather condition at which faults are expected were defined due to these values can be employed in future research works about fault forecast based on local weather forecast.