基于深度残差收缩网络的风力发电机齿轮箱故障诊断

提出了一种基于深度残差收缩网络的风力发电机齿轮箱故障诊断方法。首先,通过齿轮箱动力学模拟实验平台采集9种工况下的8种故障的振动信号;其次,对所采集的信号进行数据预处理,将其输入至深度残差收缩网络中训练;最后,利用反向传播算法不断优化网络参数,实现变工况下风力发电机齿轮箱故障的识别与分类。实验结果表明,所提方法在变工况场景下,可有效提取齿轮箱的故障特征并具有较高的识别准确率,证明了其在风力发电机齿轮箱故障诊断方面的可行性及有效性。

Gearbox Fault Diagnosis of Wind Turbine Based on Deep Residual Shrinkage Network

In this paper , a gearbox fault diagnosis method for wind turbine based on deep residual shrinkage network is proposed.Firstly , the vibration signals of eight faults under nine operating conditions are collected by the gearbox dynamics simulation experiment platform ; secondly , the collected signals are pre-processed and input to the deep residual systolic network for training ; finally , the back propagation algorithm is used to continuously optimize the network parameters to realize the identification and classification of wind turbine gearbox faults under variable operating conditions.The experimental results show that the proposed method can effectively extract the fault characteristics of the gearbox with high recognition accuracy under the variable working condition scenario , which proves its feasibility and effectiveness in wind turbine gearbox fault diagnosis.