Abstract: Typical turning experiments on tool wear are systematically done, the sound and vibration signals of tool wear are thoroughly and meticulously analyzed. These analyses include the time domain simple statistical parameter analysis, filter cepstrum analysis, correlation analysis and AR parameter analysis etc. Furthermore, multigrade on-line monitoring models of tool wear condition are effectively refined which are more accurate and adaptive, and more resistant to noise environment.

Key words: Multigrade monitoring model, Sound and vibration identification, Tool condition