| In metal cutting processes, excessive vibration or chatter has an adverse effect on productivity and product surface quality. Various studies have been reported in the literature over the past few decades. However, the real application of the outcome of these studies has been very limited. A new system has been developed in this study for chatter detection and chatter suppression. The coherence function values of the frequency spectra from two accelerometers in orthogonal directions were used as a chatter indicator. The vibration energy was used to offset the over-vigilance behaviour of the coherence function. A fuzzy logic control approach was used for chatter suppression based on both the coherence function value and vibration energy level. To improve the adaptability of the fuzzy controller, a self-learning algorithm has also been developed for on-line updating the fuzzy rule base. A direct output tuning method was also proposed to improve the responsiveness of the system.; The proposed system has been tested using both steel and aluminium workpieces with and without thin-walls. The experimental results show that the proposed system worked reasonably well for on-line chatter detection and suppression.; The thesis also explored the possibility of using the coherence function for chatter prediction. The verification of its feasibility may be carried out in the future. |
