多信号融合的风电齿轮箱异常状态检测
ANOMALY DETECTION OF WIND TURBINE GEARBOX BASED ON MULTI-SIGNAL FUSION
-
摘要: 基于大数据技术能够有效提升风电机组的可靠性状态预测和故障诊断。针对目前的研究对某种单一信号,如电信号等在故障检测中的作用以及信号融合对检测准确性影响还不深入的问题,该文基于BP神经网络,将电信号与振动信号、温度信号融合作为输入层参数,开展基于多信号融合的风电齿轮箱状态预测和故障定位,结合某风场2 MW风力发电机的实际数据,分析不同信号在风电齿轮箱状态预测和故障定位中的作用。结果表明,电信号的融合能够进一步提高风力发电机状态预测与故障定位的准确性。Abstract: Condition monitoring and fault diagnosis based on big data technology can effectively improve the reliability of wind turbines.But current researches on the role of a single signal,such as electrical signal,in fault detection and the impact of fusion signal on detection accuracy is not deep enough. Based on data fusion,the present work establishes a BPNN model for condition prediction and fault location of wind power gearbox,using fusion data of electrical signal,vibration signal and temperature signal as input. Compared with the actual data of a 2 MW wind turbine in a wind farm,the role of different signals in the condition prediction and fault location of wind turbine gearbox is analysed. The results show that electrical signals can further improve the accuracy of condition prediction and fault location.
下载:
