On-line Tool Wear Monitoring via Sparse Coding Based on DCT and WPD

Abstract

The adaptive and on-line tool wear monitoring is of great importance to improve the milling precision and efficiency. In traditional tool wear monitoring, feature extraction of cutting force signal by time-frequency method was usually off-line and needed signal reconstruction. In this paper, a novel online tool wear monitoring method is proposed. In the method, the sparse coefficients is measured by sparse coding based on DCT and WPD and then utilized to indicate the tool wear level without signal reconstruction. Experiments of tool wear monitoring are conducted for high speed CNC manufacturing. The simulation results show that the proposed method is capable to indicate tool wear level and robust to cutting conditions