TIME-SERIES BASED ANALYSIS FOR PRIMARY CHATTER IN METAL-CUTTING

Primary chatter is investigated experimentally using both traditional and frequency domain techniques. The experiment was conducted on a CNC lathe using coated carbide cutting inserts. The tool vibrations and c utting forces were measured and analyzed for different sets of cutting conditions. The experimental results indicated that the tool vibratio ns are more sensitive to the cutting geometry than the cutting forces. Also, the feed rate was found to be the most critical parameter which leads the system into chaotic behavior. In particular, the wavelet tr ansform technique was used to analyze chatter signals. It clearly iden tified the transient boundary and the occurrence of the built-up edge (BUE). Based on the frequency information contained in the wavelet tra nsforms, the cutting process is probably quadratically non-linear. In addition, for high feed rate, chaotic behavior was observed. Tradition al techniques were also used to study the chaotic behavior of the proc ess. These techniques include phase plane, power spectra, Poincare map and fractal dimensions. They all agreed with the observation obtained by the wavelet transform regarding chaotic motion.