Chatter detection in milling process based on the combination of wavelet packet transform and PSO-SVM

Chatter is one of the biggest unfavorable factors during high speed machining process a machine tool. It severely affects surface finish and geometric accuracy workpiece. To address this obstacle improve quality efficiency products, it significantly essential to detect chatter machining. Therefore, multi-feature recognition system for detection on basis fusion technology wavelet packet transform (WPT) particle swarm optimization support vector (PSO-SVM) was proposed in paper. Firstly, original vibration signals collected from acceleration sensor were processed through (WPT). The noise irrelevant information remarkably decreased. In addition, packets containing chatter-emerging chosen reconstructed. fourteen time–frequency domain characteristics reconstructed signal calculated as vectors detection. Finally, obtain optimal radial function parameter g penalty C SVM prediction model, algorithms k-fold cross-validation (k-CV), genetic algorithm (GA), (PSO) employed optimizing model parameters SVM. indicated that PSO-SVM improved obviously than others. we applied optimized by PSO detecting state end milling results accurately predicted slight advance.