STEP-NC based reverse engineering of in-process model of NC simulation

This paper proposes a comprehensive process for reverse engineering and feature recognition from an in-process model (IPM), represented in triangle meshes which results from NC simulation. IPM can be rather useful in redesign, process planning, machining inspection, etc., in addition to visualizing machining processes. For example, we may carry out in-process corrections accordingly by comparing the rebuilt model from an IPM with the original one. However, until now, IPM reverse engineering, especially manufacturing feature-based, is seldom researched. First, after systematically summarizing the IPM characteristics and taking advantages of them, a novel region segmentation method based on a shape descriptor-called the shape index, which is close to the visual effect and is capable to identify local shapes accurately, is proposed. The shape index, which is derived from principal curvatures, is calculated by discrete differential geometry methods. The segmentation is carried out gradually from simple surface types, such as planes, cylinders, to more complex surface types; second, the recognition of manufacturing features defined in the ISO 14649 standard is explained in detail. The approach is graph-based, and highly relies on the presence of concave edges in the model. In the recognition algorithm Euler characteristic and curvedness of the model are applied, which are novel attempts. Case studies to verify the proposed approaches are provided.