Quality of machined products is often related to the shapes of surfacesudthat are constrained by geometric tolerances. In this case, statisticaludquality monitoring should be used to quickly detect unwanted deviationsudfrom the nominal pattern. The majority of the literature has focused onudstatistical profile monitoring, while there is little research onudsurface monitoring. This paper faces the challenging task of moving fromudprofile to surface monitoring. To this aim, different parametricudapproaches and control-charting procedures are presented and comparedudwith reference to a real case study dealing with cylindrical surfacesudobtained by lathe turning. In particular, a novel method presented inudthis paper consists of modeling the manufactured surface via Gaussianudprocesses models and monitoring the deviations of the actual surfaceudfrom the target pattern estimated in phase I. Regardless of the specificudcase study in this paper, the proposed approach is general and can beudextended to deal with different kinds of surfaces or profiles.
展开▼
