关键词: 神经网络, 特征选择, 刀具磨损, 仿真

Abstract:

An information fusion from multiple sensors (acoustic emission,vibration and spindle motor power) via back propagation algorithm to estimate tool wear is presented.A simple efficient method,relevant analysis using grey model theory,is used to select the features.The influence of parameters on performance of network and the robustness with respect to noisy and inaccurate sensor information are investigated.

Key words: artificial neural network, feature selection, tool wearestimation　simulation