Expert System for Process Selection and Operation Optimization

Abstract

The working of an Expert System, PRODUCER, conceived as part of a CIM system for a plant manufacturing discreteparts, is presented. PRODUCER starts out by determining if the desired part can be produced. Based on part attributes it then decides on a specific manufacturing process, e.g., Welding, Casting, Forging or Machining. Having selected the process, PRODUCER establishes the particular operation. With the operation decided, PRODUCER proceeds to find all the feasible combinations, of equipment and tools, that could produce the desired component. The turning operation has been chosen to demonstrate PRODUCER's capabilities. PRODUCER then sets about the task of identifying the most optimal pair of machine-tool and cutting-tool, which will provide the highest Metal Removal Rate, MRR. This is accomplished at two levels. At the higher level, PRODUCER generates constraints, representing physical limitations of the cutting process, for each machine and tool combination. These constraints are then passed on to an Optimization program. This is a Fortran program, which operates at a lower level, and returns the optimum values of the process control variables, for each machine-tool and cutting-tool combination. PRODUCER finally yields the highest maximization of the MRR. In doing so it also identifies the particular machine-tool and cutting-tool associated with this global optimum. PRODUCER, essentially a knowledge-based production system, implemented in the First Order Predicate Logic language of Prolog, also enables intelligent adaptive control.