ALBERT

Description: Since the use of feature-based computer-aided systems became common in production, feature recognition has been a primary method to obtain features that contain specific engineering significance. In feature recognition, engineering significance is extracted from low-level elements and encapsulated into features to facilitate the various engineering tasks including process planning, manufacture and inspection. Due to the various classifications of features and their versatile application areas, there have been many different feature recognition approaches. These feature recognition methods are typically based on the part design models from computer-aided design systems. In this research, a new feature recognition method from computer numerical control (CNC) part programs for milling components is proposed. This approach uses feature recognition algorithms to integrate CNC part programs through the analysis of tool changes, spindle speeds, feed rates, raw material, tool geometry and tool paths to identify the manufacturing process plan. It has a major influence with the ability to extract process knowledge from the shop floor and represent it into a manufacturing feature-level data. This paper focuses on the recognition of 2½D features, but it can be extended to more complex features. Case studies are used to validate the use of the proposed method on typical milling features. Two sample parts are used to illustrate the efficacy and efficiency of the method. In addition, the proposed method is compared against traditional feature recognition techniques, and issues particular to feature recognition from part programs are discussed.