A New Approach of the Dimensional Inspection Based on Dimensions Assembly Identification

Abstract

In the industry practice are often meet cases when is necessary to identify piece surfaces regarding another surfaces already machined.
In present, for these cases, is used the 3-2-1 method in order to determine the reference trihedron, and, after this, to identify the piece surfaces. The reference surfaces identified by this method can’t regard the deviations which emerge at actual surfaces.
In this paper is proposed a new identification method based on the determination of dimensions assembly, which contain the dimensions reciprocally restricted by tolerances and after this the identifying of this dimensions assembly. The new method is based on neural network modeling.