Recently, there has been a further demand for precise monitoring of milling process with a machine tool by a simple and cost-effective method. One of the ways for monitoring is to install acceleration sensors to the tool spindle and estimate cutting forces from the model of the tool spindle structure. This method generally allows stable monitoring of cutting forces for tools with large diameters. However, in case of using tools with small diameters, it is difficult to estimate the cutting force due to higher frequency vibrations generated near the tool center point. Therefore, to solve the problem, we propose a new monitoring method by signal analysis of acceleration sensors in this research. The problem is that the acceleration sensor signals contain two types of signals, such as ‘acceleration change due to mechanical displacement of a tool spindle generated by cutting load’ and ‘high frequency self-excited and forced vibration’. In our method, we separate these two signals by using an approximation of sequential quadratic regression. From the former, cutting forces are estimated from equation of motions which are obtained in advance from frequency responses of the tool spindle, and from the latter, intensities of vibrations due to milling are estimated. This method was tested several milling patterns such as large cutting loads, fluctuations of cutting loads, and cutting during abnormal vibrations. As a result, we have achieved in-process monitoring of milling process not only in the X and Y directions, but also in the Z direction with a small radius ball end mill.
Read full abstract