Fault diagnosis of electro‐hydraulic servo valve using extreme learning machine
Chao Liu, Yunfang Wang, Tianhong Pan, Gang Zheng
Abstract
An electro-hydraulic servo valve (EHSV) is a core component of an electro-hydraulic servo system, the function of which directly affects the reliability and performance of the system. To distinguish the fault of an EHSV, a fault diagnosis model using an extreme learning machine (ELM) is proposed in this article. First, the structure and working principle of the EHSV are described. Next, a fault diagnosis model constructed using an ELM is proposed, in which the no-load flow characteristic curve is taken as the input and the corresponding category label is taken as the output. Using a brute force method, the activation functions and number of hidden layer nodes in the ELM are set. Compared with a model built using a support vector machine with a genetic algorithm, the proposed algorithm achieves a faster training speed and higher classification accuracy.