A knowledge based decision support system for tool changeover in CNCs

Abstract

This paper describes an application of an adaptive planning system for automatic tool changers in flexible manufacturing systems. The conventional models of predictive control usually cannot adapt to a real time dynamic environment. The proposed adaptive control model is capable of self adjusting to changing environments. The algorithm is based on a decision logic, which is constructed by breaking up knowledge and converting them into mathematical form in order to cover all possible conditions that can exist during the implementation phase. Expert thoughts and knowledge from decision logic are stored in the decision tree, which consists of circular nodes, arcs and decision nodes. The suggested system is capable of accepting further rules, new nodes and branches to the tree when additional attributes are needed. This whole knowledge is encoded in the form of production rules and each rule represents a small chunk of knowledge relating to the given domain of tool replacement. A number of related rules collectively respond to highly useful conclusions.The system uses VP Expert development shell, contains an inference engine and, a user interface. The originality of the proposed strategy lies in that a knowledge-based expert system is developed to identify and analyze the current conditions and then readjust the output that reflects the real-time environment. Compared with the various classical models, the approach can synthesize and analyze as many variables as possible to adequately and reliably identify the real-time conditions. Simulation results demonstrate the effectiveness and practicality of this tool-change planning and control strategy.