Research On Fault Diagnosis Of Hydraulic System For Continuously Variable Transmission Of Hydraulic Machinery

The hydraulic mechanical continuously variable transmission(HMCVT)in tractors is a highly automated mechatronic-hydraulic integrated system.Once a failure occurs,not only will it be difficult to diagnose the problem,but it will also delay farming operations,and in severe cases it may even threaten the safety of operation and driving.However,due to the late start of research in China,there is a lack of experience in its manufacturing,control and use.In order to solve the problem of fault diagnosis and improve the reliability of the shifting process,this paper conducted research in the following four aspects:(1)To address the problem of lack of fault data for the hydraulic machinery continuously variable transmission(CVT),a fault diagnosis test bench was built to collect fault data in real time.In response to the need for sample segmentation,an improved dynamic time warping(DTW)algorithm was proposed.By analyzing the common fault mechanisms in the hydraulic system of the hydraulic machinery CVT,such as oil channel leakage,oil channel blockage,seal ring damage,solenoid valve spool jamming,and clutch piston jamming,pressure data were collected under different working conditions and subjected to denoising and integration.To address the problem of data acquisition breakpoints,a top-down segmentation method was proposed to speed up the traditional DTW algorithm,and the DTW distance calculation formula was improved based on the similarity calculation to achieve data sample segmentation.This method greatly improves the accuracy and speed of sample segmentation.(2)Aiming at the high-dimensional,nonlinear,and redundant characteristics of the fault feature of the hydraulic system under the control of heavy-duty clutch,a Gaussian naive Bayes classification method based on time window and principal component analysis(PCA)was proposed.The method uses time windows to intercept clutch pressure data,extracts fault characteristics by PCA,and completes fault diagnosis through a Gaussian naive Bayes classifier.The experimental results show that this method reduces unnecessary dimensions and data redundancy.Compared with traditional machine learning methods,it has higher average accuracy and recall rates,which are 97% and 87%,respectively.This method has important significance for quick fault location and repair,improving equipment operation stability and safety,and can provide reference for similar research fields.(3)Aiming at the gradual development of faults,the huge amount of fault data,and the hidden fault characteristics of the hydraulic system under the control of light-duty clutch,a fault classification method based on an improved echo state network(ESN)was proposed.This method utilizes the reservoir structure of the Echo State Network(ESN)to induce rich echo state features from time sequences.Multiscale convolution is used as a replacement for the linear regression algorithm in ESN,overcoming the limitation of the linear regression algorithm in decoding high-dimensional features.Additionally,an attention mechanism is introduced to improve the model’s sensitivity to sensitive time periods.Experimental results demonstrate that the proposed improved network,Attention＿Conv ESN,achieves a fault diagnosis rate of 96.98%.The proposed method not only effectively improves the accuracy of fault diagnosis but also significantly reduces the training parameters and improves the model operation speed,with better fault diagnosis performance.(4)In order to further achieve visual analysis and precise management,a working condition monitoring system for hydraulic machinery CVT was designed and developed based on actual needs.The system has functions such as data visualization display,intelligent fault diagnosis,and monitoring device management,and can achieve online real-time storage,display,and alarm processing of working condition data.The monitoring system developed in this study has the characteristics of good interactivity,strong real-time performance,and high reliability,and can provide an efficient and intelligent supervision system for users who manage equipment installed with hydraulic machinery CVT,greatly improving work efficiency and management level.This system has important significance for improving the stability,reliability,and safety of equipment.