长短时记忆的风电机组故障预测系统

Long and short term memory based fault prediction system for wind turbine

风力发电是一种重要的可再生能源发电方式,其主要设备的故障严重影响风力发电效率与供电的安全可靠性。为控制风电设备主要零部件故障的发生,提出了一种基于长短时记忆(LSTM)的神经网络故障预测方法。该方法主要包括历史数据学习建模和在线数据监测两部分。通过预测风机齿轮箱的磨损故障,验证了该方法具有较好的预测性能,且能较早预测出故障的发生。

Wind power generation is an important method for renewable energy power generation.The efficiency of wind power operation and the reliability of power supply will be affected seriously by its main equipment failure.Therefore,in order to prevent the main component failure of wind power equipment,a fault prediction method based on long and short term memory(LSTM)neural network is proposed.The method mainly includes two parts:historical data learning modeling and online data monitoring.The feasibility of the method is verified by predicting the wear failure of the wind turbine gearbox.The experiment results show that this method has better predictive performance and can predict the failures early.