基于胶囊网络的风电机组主轴承故障诊断研究

Research on Main Bearing Fault Diagnosis of Wind Turbine Based on Capsule Network

为了实现对直驱式风电机组发电机轴承的故障诊断,根据传热学理论建立发电机轴承温度动态模型,通过模型获得发电机轴承产热和散热特征参数。结合一定周期内机组多传感器测量数据和轴承产热和散热特征,组成设备状态矩阵。使用胶囊网络,从状态矩阵中提取了轴承故障特征,实现了对风电机组发电机轴承的故障诊断。在针对实际风电机组的故障诊断试验中,模型取得了95.3%的准确率。该研究为故障样本较少且不安装振动监测设备的直驱式风电机组组成的风电场主轴承远程监测和故障诊断提供了参考。

To realize the fault diagnosis of generator bearings of direct-drive wind turbine,a dynamic model of generator bearing temperature is established according to the heat transfer theory,and the parameters of generator bearing heat production and heat dissipation characteristics are obtained through the model.Combining the multi-sensor measurement data of the turbine within a certain period and the heat production and heat dissipation characteristics of the bearing,the equipment state matrix is formed.The bearing fault characteristics are extracted from the state matrixby using capsule network,and the fault diagnosis of the generator bearing of the wind turbine is realized.The model achieved 95.3%accuracy in the fault diagnosis experiments for actual wind turbines.This study provides a reference for remote monitoring and fault diagnosis of main bearings in wind farms composed of direct-drive wind turbines with fewer fault samples and no vibration monitoring equipment installed.