基于深度置信网络风电机组变桨系统的故障诊断

针对风电机组变桨系统常见故障,提出一种基于深度置信网络(DBN)的故障诊断方法。设计出基于DBN的变桨系统故障诊断框架;通过堆叠多层受限玻尔兹曼机(RBM),对比重构数据与原始输入数据差异,研究了DBN故障特征自提取能力;将堆叠RBM提取的故障特征输入到顶层分类器中进行训练,得到故障诊断模型;最后采用风场真实故障数据集进行了验证测试。实验结果表明,采用该方法进行风电机组变桨系统故障诊断相比其他方法具有更高的准确率。

Fault Diagnosis of Wind Turbine Pitch System Based on Deep Belief Network

In order to eliminite the common faults of wind turbine pitch system,a fault diagnosis method based on deep belief network (DBN)is proposed.The fault diagnosis steps of pitch system based on DBN are discussed.By stacking the multi-layer restricted Boltzmann machine (RBM) and comparing the difference between the reconstructed data and the original input data,the self extraction capability of the fault featare of DBN was studied.The features extracted were input into the top classifier for training,and the fault diagnosis model was obtained.The wind field real fault data set verified the results.The experimental results show that the method is more accurate than other methods in fault diagnosis of wind turbine pitch system.