Design of Screw Performance Evaluation System Based on DSP and Wavelet Neural Network

An embedded system of screw performance evaluation based on DSP and wavelet neural network is designed for evaluating screw performance rapidly and accurately in the CNC processing field.This paper deduces that monitoring the vibration of screw and researching the relationship between load force and the vibration can effectively evaluate the trend of screw performance.The vibration of screw is served as analysis signal,the eigenvalues extraction algorithm and the wavelet-neural-network based performance evaluation algorithm are put forward,and the system hardware is presented.According to the experiments on the NC CINCINNATI V5-3000,the embedded system has features such as reliability,accuracy and easily algorithm updating,and the design is ideally suitable for the development and testing of the performance evaluation algorithms.