A Cyber-Physical Production System Framework of Smart CNC Machining Monitoring System: Hidden Semi Markov Model for Tool Effective Residual Life

The micro groove milling experiment was carried out on HSM600U high speed machining center. The tool used in the experiment is a vertical flat end milling cutter with a diameter of 800 microns. The cutting material is steel T4. Each cutting segment lasts for two seconds, and 100 cutting segments are conducted for each group of experiments. The cutting force signal is measured with a Kistler dynamometer, and the sampling frequency is 50KHz.Six different cutting conditions were used in the experiment. Experiments 1, 3 and 5 are used as training samples, and Experiments 2, 4 and 6 are used as test samples. In the experiment, the time sampling rate is fixed, and the space sampling rate corresponding to different rotational speeds is different. In order to keep the spatial sampling rate unchanged, the cutting force signal collected is down sampled and processed, so that the spindle rotates for one circle to sample 100 points. The cutting force coefficients and related dimensions are calculated with 50 cycles of cutting force signals in each cutting segment. Figure8.fig is the evolution process of the error and average value of the predicted tool remaining life (RUL).