| In metal cutting processes, chatter has been recognized as one of the main factors that limit machining productivity and affect product quality. Two different categories of chatter were classified by researchers, i.e., regenerative chatter and non-regenerative chatter, and in this thesis the former is mainly studied. Over the past few decades, though various chatter detection and suppression methods have been developed, their industrial acceptance is still very limited. This research work presents a new system for on-line chatter detection and suppression. Its detection module implements a statistical index to identify chatters by performing wavelet transform and conducting statistical analysis of positive wavelet transform modulus maxima (WTMM). To suppress chatter, two versions of fuzzy control modules, i.e., plain fuzzy control and self-regulating fuzzy control have been implemented. Unlike the previous chatter suppression systems, the new suppression module features two-way adjustment, i.e., both increasing and decreasing the amount of adjustment. Along with the use of single or multi-output control variables to suppress chatter, productivity is preserved as much as possible.; The proposed system is implemented on a SERVO 2000 milling machine. Extensive tests have been carried out. The experimental results show that the wavelet-based chatter detection index can not only detect the existence of chatters but also distinguish the severity levels. The new chatter suppression module works reasonably well in most tests. However, its performance is adversely affected in the presence of non-regenerative vibrations due to the lack of workpiece or clamping rigidity. Further improvements need to be carried out for industrial applications. |
