| Now, the machine manufacture is developing to the flexibility and automation. The key to ensuring smooth production is to monitor tool wear condition in the large-scale automation production. It is very important to extending no-fault running time of machine tool and improving machining efficiency, production quality. Based on the technology status quo of tool condition monitoring, this paper has carried out research on the tool condition monitoring by cutting sound signals, primarily works for the following:(1) Established sound signals collection system. The cutting sound signals were collected on the NC machining center and analyzed by statistics analysis and power density spectrum. The characteristic component which changes with the increase of tool wear degree has been discovered by above analysis. So it has been proved feasible that the tool condition could be monitored well by cutting sound signals.(2) About the monitor of tool wear condition, based on that the tool wear condition is corresponding to the cutting sound singnals, this paper presents an on-line tool wear condition monitoring measure by cutting sound spectrum HMM (Hidden Markov Model) recognition. This paper has carried out tool wear condition recognization experiments by establishing HMM and extracting LPCC (Linear Prediction Cepstrum Coefficient) as the characteristic parameter, and achieved the tool wear degree recognization in grade finally by above experiments.(3) About the monitor of tool breakage condition, based on the characteristic of instantaneous change with the tool breakage, the characteristic parameter was extracted by wavelet analysis, and the similar signals also were got rid of by wavelet analysis. Then the tool breakage can be detected by setting suitable threshold value.(4) Completed hardware design of the monitoring system. This paper selected ARM as the CPU of monitoring system to satisfy the need of recognized arithmetic and completed the peripheral electrocircuit design.
|
PDF Full Text Request