Monitoring of self-excited vibration in grinding process using time-frequency analysis of acceleration signals

Abstract

Grinding is a finish manufacturing process for high precision parts. As the part has high value added at this stage, anomalies and defects are unacceptable. Self-excited vibration or just chatter is an undesirable and complex phenomenon that occurs in grinding, as it may impair the performance of ground parts due to the negative effects such as poor surface quality, intolerable inaccuracy, disproportionate tool wear, among others. This work goal is to present a method for chatter detection in surface grinding of SAE 1045 steel with aluminum oxide grinding wheel by using acceleration signals and digital signal processing techniques. Experimental tests were conducted in a surface grinding machine where a data acquisition system allowed for acquiring the acceleration signals at 2 MHz. The surfaces of the workpieces were visually inspected by means of high resolution images. The harmonic content of the signals along the grinding period was investigated by using the short-time Fourier transform with the purpose of relating the chatter occurrence to it. The results show that the frequency magnitudes of the signals change in time and frequency domain with specific patterns in the spectrogram as the chatter takes place, and they are related to the grinding wheel rotation and its harmonics.