Overview Of Allied Signal's XCUT System

XCUT can successfully plan for 80 percent of the features on moderately complex prismatic machined parts and reduce 2 – 4 hours of manual process effort to 15 – 30 minutes, and the automatic recognition and creation of rectangular depression features from the piece part solid model has been addressed.

Abstract

The expert system, XCUT, utilizes Artificial Intelligence (AI) technology for preparing production process plans, automatically selecting cutting tools, and automatically defining machinability parameters and NC tape requirements. XCUT includes the capability to analyze solid model product representations and recognize feature information required by subsequent automated manufacturing processes. A distributed architecture has been implemented where XCUT, the solid model, and the feature extraction interface all run as separate processes and communicate through inter-process mail boxes for integration. XCUT can successfully plan for 80 percent of the features on moderately complex prismatic machined parts and reduce 2 – 4 hours of manual process effort to 15 – 30 minutes. The automatic recognition and creation of rectangular depression features from the piece part solid model has been addressed. Also, XCUT can now automatically classify 80 percent of the feature solid model bodies as a particular instance type in the XCUT feature taxonomy and determines feature attributes (such as, bottom face, depth, width, length). Areas of future work have been identified that will enable the XCUT system to be used for production. These areas where further development is required are: geometric reasoning, representation of manufacturing features, dimensioning and tolerancing, feature extraction, distributed computing architecture, knowledge gathering, and user interfaces.