Diagnosing Quality Problems on a Complex CNC Machine via Process-oriented Basis Representations

Abstract

A Process-oriented basis representation (POBREP) can be used to express a multivariate quality vector as a linear combination of fault patterns, plus a residual. Monitoring the estimated coefficients of the linear relationship is especially useful when the quality vector contains the same kind of measurements at different locations on a product. Such an application is the subject of this study. POBREPs are derived for different types of quality problems for a 9-axis CNC drilling machine; complicated geometry involved in the drilling operation makes the derivation process a challenging task. Therefore a thorough treatment of the derivation process is presented. Next, diagnostic power of the POBREP is demonstrated by an application to multivariate process data.