Application Of Neural Network On Measurement And Assay System Of Switched Reluctance Motors

This paper discusses mainly on measurement and assay system of switched reluctance motor drives and the applying of neural network.Measurement and assay system is included hardware and software. The hardware includes designing the 80C196KC least system, parts of measuring electric parameters, and communication interface. Software includes PC computer which is built on VB and chip microcomputer which is basic on basic language. This paper is based on VB for developing platform of interface, then have succeeded in series communication between PC and 80C196KC chip microcomputer by RS-232, data acquisition and data processing of Switched Reluctance Motors. 80C196KC chip microcomputer is the cell of data acquisition to transmits the data to PC computer of client, which administer data by SQL sever. VB integrates with MATLAB seamlessly by ActiveX technology, which makes full use of advantage of MATLAB in data acquisition.The applying on nonlinear error compensation and nonlinear model simulation of SR. are discussed in this paper. As result of shortcoming, degree of convergence and precision of convergence are effected. So a joint-optimized method which is to adjust learning factor and momentum factor outline and to adjust shape factor online is proposed.