Online Tool Condition Monitoring of Milling Machining Based on Time–frequency Analysis of Vibration Responses

In the milling process, the vibration signals show strong non-stationarity due to the influence of various symbiotic factors. And time–frequency analysis method is an effective means to analyze and deal with time-varying and non-stationary data. Therefore, this paper focuses on analyzing the vibration responses with the representative time–frequency analysis method short time Fourier transform (STFT). It has been found that STFT shows a high processing effect and good time–frequency characteristics in association with the signal dynamics. In particular, it can more clearly enhance the repetitive impact signal features due to the intermittent interactions between the cutter and workpiece. At the same time, vibration signals during the milling process are superimposed on periodic sinusoidal signals, impact signals and other strong noise signals, etc. Such understandings lead to new effective features including the Gini index, the root-mean-square (RMS), Kurtosis, and average mean value to form a set of measure characteristics and achieve multi-feature evaluation of tool wear state. The experimental results of milling manufacturing demonstrate that the proposed method can identify the early tool wear in real time with fast response and good robustness, which meets the requirements of online monitoring.