Development of Knowledge-Expandable Ontology-Based Expert System for Process Planning in Cold Forging of Flange Nuts

Abstract This paper proposed a new method of constructing an ontology-based expert system with an expandable knowledge base for process planning in cold forging of flange nuts. There are two essential parts of this method. In the first part, the software Protege is used for expressing the experiential knowledge as a Web Ontology Language (OWL) ontology and Semantic Web Rule Language (SWRL) rules. The inference is executed in Java by means of the API provided by Jess, the rule engine for the Java platform. In the second part, new knowledge is acquired from parameter optimisation experiments, combing process simulation and Taguchi methods. Deform-3D was adopted as the process simulation engine. Then the tool wear of front punch at the 4th stage was predicted for the tool life evaluation. A modified Archard wear model is used for tool wear prediction during the simulation. On the other hand, Taguchi methods is used for designing an optimal tool which makes the tool life longer than that of the existing design. The novelty of the proposed system is that the new knowledge is represented as an OWL ontology and SWRL rules, which makes the knowledge base of the system expandable. The knowledge base can thus be expanded if there are modifications or new knowledge. The system is capable of inferring a workable forging process for flange nuts. It could recommend dimensions of tools after the necessary information, such as the nominal size and specifications of the product, has been input. Users are able to reduce tool design time by taking the results from the system as a process planning decision support. In this way, when the manufacturer gets a new quotation from the client, engineers can search from the company tooling database for feasible forging tools and make the cost estimation based on the forging process plan generated by this system.