Vibration Status Monitoring of Machine Center Based on EMD and LTSA

Abstract

As the abnormal conditions of manufacturing process seriously affect the machining performance of machine tool, the vibration signal of spindle is selected to monitor the manufacturing process of machine center. The vibration signals are decomposed by empirical mode decomposition (EMD) method, and the first five intrinsic module functions components are picked out to calculate the power spectrums. Then, the local tangential space arrangement (LTSA) method is developed for dimension reduction, and the one-dimensional feature vector indicating the vibration state is obtained. A support vector machine model is used to classify vibration states based on one-dimensional critical features of three different manufacturing processes. The classification result indicates that the EMD-LTSA method is an efficient feature extraction method for vibration status monitoring of machine tools.