| Manufacturing practices are generally established by trial and error technique. This process is not only cumbersome but expensive as well. At the same time, computer time to conduct numerical simulations has also become a cost effective factor in the competitive manufacturing environment. The outcome of the process design, that is, optimal solution, is dependent on a number factors including the knowledge and experience of the designer. Artificial Intelligence (AI) tools such Expert Systems (ES), Artificial Neural Networks (ANN) and Case-Based Reasoning (CBR) use this idea of expert knowledge to achieve a workable, economical and optimum solution to the given problem in a faster manner. The time factor has become all the more important due to increased competition, greater demand, and stringent standards in industry.; The primary objective of this research is to develop a methodology and software that will use the AI techniques for the automated design and analysis of various metal forming processes, and an intelligent methodology for the real-time control of a hot rolling processes. An intelligent hybrid system, a combination of ES, ANN and CBR, was used to develop the software for the intelligent design and analysis of metal forming process. This software envisages the implementation by coalescing Computer-Aided Engineering (CAE), Computer-Aided Design (CAD), Material Science theory, Design theory, Optimal Process Control, and AI techniques to achieve a robust and workable system. The required near-net shape is given in a CAD system as an input. The objective is to recommend a metal forming process suitable for forming this shape from a given preform and carry out the analysis and finally develop a process planning report to help in the shop floor for producing this part.; The intelligent design and analysis software was developed to run under the Microsoft Windows{dollar}spcircler{dollar} environment using AutoCAD{dollar}spcircler{dollar} for windows as the CAD system. NeuroShell{dollar}spcircler{dollar}, a neural network shell, CBR Express{dollar}spcircler{dollar}, a windows based case-based reasoning tool, and AutoLISP, VisualBasic and C++ programming languages were used to develop the feature extraction, data generation, design, analysis, reasoning and report generation modules of the software. |
