基于改进PCA-SOM的电静压伺服作动器油滤堵塞故障诊断

针对电静压伺服作动器（EHA）的油滤堵塞故障,提出利用可调式球头油堵预置不同程度的油滤堵塞工况进行数据采集,并在传统自组织映射神经网络(SOM）的基础上,引入主成分分析（PCA）法,利用各元主成分贡献率对神经元竞争域值各维系数进行修订,提出了改进PCA-SOM神经网络对系统堵塞状态进行判识。研究结果表明,与传统SOM神经网络和PCA-SOM神经网络相比,改进PCA-SOM神经网络在提高聚类效果的同时,提高了模型的准确率和训练速度,在EHA的油滤堵塞故障诊断中有更好的适用性。

Identification of Fault Degree of Oil Filter Blockage in Electro-hydraulicActuator Based on an Improved PCA-SOM

In view of the oil filter plugging faults of EHA, to using adjustable ball head oil plug was proposed to preset different degrees of plugging conditions for data collection, and based on the traditional SOM, PCA was introduced to revise each dimensional coefficient of the neuron competition domain values by using the contribution rates of each principal component, as well as proposed an improved PCA-SOM neural network to identify the blockage states of the system. The results show that compared with the traditional SOM neural network and PCA-SOM neural network, the improved PCA-SOM neural network has higher applicability in EHAs oil filter blockage fault diagnosis, which is embodied in that the clustering effectiveness is improved while the accuracy and training speed of the model are also promoted.