Research On Wind Turbine Fault Prediction Based On Monitoring Information

In recent years, wind power is becoming the world’s fastest growing renewable energy, according to government plan. By2020China’s wind power installed capacity will reach30GW. The complexity of wind turbines and wind power plants always located in remote, even at sea, make the equipment maintenance costs to be one of the largest expenses for the current wind power plant. Therefore, more effective maintenance strategy is badly needed for the maintenance of wind power generation equipment. If the forecasts made before a failure occurs, we can advance the development of maintenance plans, reasonable arrangements for maintenance personnel and supplies, reduce downtime caused by a sudden wind turbine damage, and ensure the unit normal continuous operation. Therefore, the study of key components improving wind turbine fault prediction algorithm to reduce the wind farm operation. maintenance costs. which is of great significance for wind turbine reliability.This article describes the composition of the wind turbine and its monitoring status, highlighting the wind turbine’s gearbox operation. maintenance. and the main form of failure. On the base of effective real-time monitoring of the wind turbines, two prediction algorithms are introduced-the time series algorithm and improved grey algorithm. and their characteristics and application of prediction are analyzed. At first we improve the gray algorithm and apply it to the gearbox’s prediction, based on the real-time monitoring information of the wind turbine. The future temperature data is predicted in advance to show the temperature changes, combining with the experience threshold, the decision of maintenance will be made. Then, this paper introduces time series analysis method for temperature prediction, and decision-making techniques-statistical process control method, combining the two algorithms together to make wind turbine gearbox failure prediction identification. According to the statistical control chart. reliable gearbox fault information and diagnosis results can be showed. Finally, theories and algorithms described in this paper are applied in a North China wind farm, analyzing the actual turbine performance monitoring data. And the desired results are achieved, proving that the proposed algorithms are effective.