Applications of artificial intelligence and the design of fixtures for automated manufacturing

Abstract

It is generally acknowledged that workholding is a very time consuming operation to be performed during any machining process. While a considerable amount of research and development effort has been centered on the automation of the cutting process and part manipulation within a work cell, less attention has been spent on the part setup and clamping problem. It is proposed that currently, there is a lack of knowledge about the more "craft" oriented information and the subsequent decision making made by the expert machinist or tool designer when clamping the work. The paper being presented establishes a framework to codify the workholding process. A method of developing a database of part, fixture, and process information is described. A rule set based on geometry and knowledge obtained from the machinist is then developed. Eventually, it is desired to incorporate the data base and rule set into a production system language called the Cell Management Language (CML) in order that the system's performance may be compared with that of an expert machinist. Finally, in line with the automation aspect of machining and workholding, a new device that allows heavy duty clamps to be automatically loaded onto a special fixture bed, remotely locked, and actuated to clamp down on the work is presented.