Chatter prediction using merged wavelet denoising and ANFIS

Inexpedient vibration between cutting tool and work piece promotes regenerative chatter in turning process. Generally, acquired chatter signals are contaminated with ambient noise. In this study, experimentally recorded raw chatter signals have been denoised using wavelet transform. Further, in order to quantify the chatter severity a new parameter called chatter index has been evaluated considering aforesaid denoised signals at different levels of cutting parameters as inputs. Moreover, these input and output parameters have been considered to train the chatter phenomenon using adaptive neuro-fuzzy inference system. Developed technique has been validated by performing more experiments. Hence, a new idea for identification and quantification of chatter has been proposed which will be very efficient in suppression of chatter.