Research On Knowledge-Based Engineering (KBE) System Of Camshaft Casting Based On Neuro-Optimization

Up-to-date, development of computer application system of CAX (CAD,CAE,CAM) is based on queuing interfacial methodology. This result in difficulty of flexible knowledge analyzing in design and manufacturing process. Also it is uneasy for us to express varieties of technological knowledge. The project applied knowledge-based engineering (KBE) technique to design of casting process. By means of object oriented modeling and expert system method, process knowledge was organized with clarified structure. The system developed can simulate effectively ways of problem solving by expert of casting field. In the system, construction and analysis of process knowledge is based on neuro-optimization.In this paper, design and implementation of system developing is discussed in detail. Based on connection mode of neuro-optimization to KBE system, the general function of system and special functions corresponding to a number of sub-modules were schemed. The system includes two parts, i.e. neuro-optimization and parametric design. In development of design system, knowledge analyzing and updating provides data support to parametric design. Knowledge mining and structuralizing is conducted according to technological references and experience of domain experts.In the first place, expression and inference mode of specialized domain knowledge (about technological conditions and design parameters) was modeled, in which object oriented modeling method is applied. Reasoning strategy is in two-way, comprising data driving and goal driving. Based on neuro-optimization, camshaft casting CAD system is divided into two sub-system, i.e., casting process/mould CAD and neuro-optimization. In the casting process/mould CAD, fine classification can be conducted by process CAD and mould CAD. In process CAD, some tasks including casting geometric molding, parting surface determining, gating system and chill design should be achieved. Mould CAD includes design of casting model, molding plate, and standardized components, check of interfering etc. To sub-system of neuro-optimization, there are four parts, including acquisition and inference of knowledge, network modeling, learning of network, structure updating etc. In the system, modeling efficiency is the key point of neuro-optimization, which assure successful process design.In application, step gating system of ironed-casting camshaft is designed by use of neuro-expert system developed. With AutoCAD, 3-D modeling of gating system is achieved based on Object ARX technique including choice of process conditions, design of gate and molding plate etc.The research testifies that it is feasible to apply neuro-optimization to casting KBE scheming. Combination of both enhance integrity and attributes featured by re-utilization and updating capability. This is the focus of the study achievements.