Fault Diagnosis of Electro-Hydraulic Servo Valve Based on Multilayer Perceptron

Abstract

The electro-hydraulic servo valve is the core component of the hydraulic system, so the electro-hydraulic servo valve is the key object in the maintenance process of the hydraulic system. Therefore, it is very important to study the fault of electro-hydraulic servo valve. Set up an experimental platform to measure the data under 3MP, 3.5MP, 4MP, 4.5MP and 5MP. After analysis and tradeoff, it was finally decided to use Adam, Momentum and RAdam optimization algorithms to realize the fault diagnosis of electro hydraulic servo valve. The data are organized into 37 dimensional arrays as samples for training. The first 32 dimensions are the characteristic parameters of the electro-hydraulic servo valve, and the last 5 dimensions correspond to five fault modes. The test set is 32 dimensions, and the diagnosis result is 5 dimensions corresponding to a fault mode. The diagnosis result is in line with expectations.