C/S Structure-based Rotating Machinery Fault Diagnosis System Design

The failure of the rotatory machinery leads to tremendous economic losses. Detecting the type of failure and locating the position of failure can avoid the mechanical fault at the beginning of the rotatory machinery failure, and provide advice for the maintenance of the rotatory machine, the replacement of the failure parts. For complex machine, which is away from other, Networked fault diagnosis system is needed to diagnose fault remotely, such as wind turbine, hydroelectric generator, etc. Compare to B/S structure, C/S structure has better data transmission capabilities, less load for database server, better security of information. To realize the remote fault diagnosis, this thesis aims at the design of rotary machinery networked fault diagnosis system, which is based on the C/S structure. The outline of the work is listed as below:Full implementation of the thorough project design of the C/S structure-based networked fault diagnosis system. Analyze the total demand of the system, especially for the database module demand and fault diagnosis module demand. Do Scheme design of all modules of the system based on the principle of modulation, including networked data acquisition: Connecting other modules via LAN for data transmission; Remote status monitoring: Monitoring of rotating machinery real-time in LAN, to ensure the normal operation; networked fault diagnosis: The client access database server for failure sample to realize fault diagnosis of abnormal data; database server: To store data and related information, accessed by other modules in the LAN. The main contents of this thesis are database module and the design of fault diagnosis module.Design the C/S structure- based fault diagnosis system database module. The data source of the database is included, the data is acquired and transferred to the database server by acquisition module via local area network; the creation of the fault sample base, the remote monitoring module performs the control of the acquisition/ management of database server/creation of the outlier data sample base. Perform the database module structure design and logic structure design for the database, to get access to the database module via internet remotely. Design the datasheet for the database, design the all the modules of the database’s operation, such as management of database at the client port, data browsing, user-management and etc.Design the function of the C/S structure-based fault diagnosis system client module. There are some different function modules, and integrate the different signal processing methods,fault recognition methods to the corresponding function modules to realize the networked fault diagnosis. Pre-processing function includes wavelet de-noising, EMD de-noising. Time –frequent domain features, wavelet pack power spectrum feature, and mixed domain feature are integrated into feature extraction. Dimension reduction function includes PCA, KPCA, manifold learning algorithms. Neurotic network, support vector machine and k-nearest neighbors are integrated in pattern recognition. It’s convenient to choose different methods to perform fault diagnosis, and then process the signals via different methods, which leads to the best solution.Develop the C/S structure-based fault diagnosis system which has independent intellectual property rights. Its ability evolved from networked monitoring to the creation of fault database and fault diagnosis based on the database. The C/S structure-based fault diagnosis system will be verified through power transmission fault diagnosis test-rig.