Learning expert system for on-line control of manufacturing process

Expert systems are very useful in implementing automation of machining processes where CNC and similar equipments are involved. Frame concept of Turbo Prolog is used to incorporate ‘learning’ capability into automated system. Learning in this context means to add, modify, and delete existing knowledge base. This is required in on-line control of machining process which uses the error signal to update the knowledge base. A knowledge base has been created for a drilling process for simulation. Drilling is defined as a frame. A frame consists of an object and a list of slots describing the object. The ‘has’ predicate is used to describe the relation that drilling has speed, depth of cut, feel rate, etc., i.e., the attribute is in a slot structure rather than being hard coded. The authors can define another frame, ‘drill-2’ as having different attributes plus those inherited from ‘drilling’ (using the rule of inheritance). The ‘is-a’ relation is used to relate a particular drilling process to a more general one. The inheritance also helps in optimizing knowledge base. Other features of Turbo prolog are used to create, modify, and delete frames. The paper also includes interfacing knowledge base with expert system.