The hydraulic system leakage failure is the most common failure of the hydraulic system.The hydraulic system leakage can cause problems such as poor pressure holding performance,insufficient output of the hydraulic cylinder,reduced working efficiency,unstable movement speed,low load stiffness,environmental pollution,etc.May endanger the safety of life and property.According to the current situation that most of the current hydraulic system leakage fault diagnosis researches adopt simulation research and simulated leakage experiments,a hydraulic system leakage failure experimental platform including a reversing circuit and a speed regulating circuit is built,and a one-way valve leakage research based on a real prototype is carried out.The main work contents are as follows:1.Analyzed the research status of the leakage mechanism,feature extraction,and pattern recognition of hydraulic system leakage faults at home and abroad,systematically elaborated the cause and form of leakage caused by one-way valves,and summarized the mathematical models of leakage calculation for different leakage forms.2.Build an experimental platform for hydraulic system leakage fault diagnosis,develop a data acquisition system based on LabVIEW,conduct an experimental study on the check valve of the hydraulic system,and obtain experimental data samples of different leakage failure modes.3.A wavelet packet transform with better time-frequency localization effect on the signal is used to establish a multi-source multi-domain feature set of one-way valve leakage faults that integrates the time domain and the time-frequency domain and contains pressure and vibration characteristics.26 characteristic parameters in the time-frequency domain.Principal component analysis is used to study the dimensionality reduction of feature parameters,and a compressed low-dimensional feature matrix is obtained.4.Based on the in-depth analysis of the hydraulic system leakage failure mode identification method,a diagnosis method of one-way valve leakage failure based on pressure and vibration signal processing is proposed.Determine the structure of BP neural network,learn samples and training parameters and train them,and input test samples for verification,which can effectively identify the four types of leakage faults of the one-way valve,verify the effectiveness of the method,and have certain engineering practical value for leakage fault diagnosis. |