Development of a knowledge-based system for the control of 319-aluminum melt quality and prediction of casting characteristics.

摘要

Knowledge Based Systems are computer programs used in situations where the use of heuristic knowledge and reasoning offers the only possibility of a solution mainly because the problem being studied is too complex to be represented in an algorithmic form. In general terms, such systems address and solve knowledge-intensive problems that can have many acceptable answers. The interpretation of Thermal Analysis results is a suitable task for artificial intelligence technology since it is a knowledge intensive activity in which several factors interact and modify the characteristics of the cooling curves. In the present work, an Aluminum Thermal Analysis System (AlTAS) developed at the University of Windsor was implemented at a casting plant and after verification of its performance in an industrial scenario a Case-Based Reasoning System coupled with a Statistical Process Control module were developed to upgrade the capabilities of the AlTAS. The system will be used to control, on-line and prior to casting, the quality of the W319-Al melt used to produce aluminum engine blocks under the Cosworth casting process, and predict their tensile properties and microstructure.Dept. of Industrial and Manufacturing Systems Engineering. Paper copy at Leddy Library: Theses u26 Major Papers - Basement, West Bldg. / Call Number: Thesis2001 .P45. Source: Masters Abstracts International, Volume: 40-03, page: 0777. Advisers: Reza Lashkari; Jerry Sokolowski. Thesis (M.A.Sc.)--University of Windsor (Canada), 2001.