The success of most companies in the machining industry relies on a high level of efficiency to remain competitive. This is achievable through fully automated monitoring and diagnostic systems that indicate the present condition of cutting tools. Before accomplishing this, the measurement and diagnostic technique must illustrate high reliability.;In this study, vibration signals were acquired to detect cutting tool failure on a milling machine, due to the practical advantages of vibrations and the fact that milling is a highly versatile and widely applied operation.;A three-insert face mill machined workpieces of nominally identical dimensions. The various cutting insert configurations applied during the tests contained sharp inserts and inserts of known fracture magnitude. The vibration signals generated by the cutting operation were analysed by quantitative parameters extracted from the frequency domain and amplitude probability functions to determine whether insert fracture could be identified. Two parameters were investigated using the frequency domain data: they were area under the frequency band, and the overlapping area of the frequency spectrum for two adjacent inserts. With the amplitude probability distribution (a.p.d.), the skewness, overlapping area, and the mean parameters were looked at as potential fault features. Using a classification scheme, the area under the frequency band and the a.p.d. vertical mean showed great potential for being implemented into an automated diagnostic system.
展开▼
