Tool breakage detection using statistical moments

摘要

On-line monitoring and detection of tool breakage is essential for highly automated machining centers. This paper presents a tool breakage method that utilizes statistical moments of the Beta distribution. The method has been applied successfully by the first author for the detection of tooth cracks in gears. It is extended here to the detection of chipped or broken teeth (inserts) in milling cutters. The method is assessed using cutting tests of steel workpieces employing Taguchi's approach for statistical design of experiments. Results of these experimetns show the Kurtosis of the Beta distribution to be a reliable tool breakage indicator that is insensitive to operating conditions and for typical amounts of tool radial runouts.