Research Article

The represenaion of metallurgical knowledge for alloy design

I. Hulthage^a1, M. A. Przystupa^a2, M. L. Farinacci^a3 and M. D. Rychener^a1

^a1 Robotics Institute, Carnegie–Mellon University, Pittsburgh, PA 15213

^a2 P.O. Box 39, Redondo Beach, CA 90277

^a3 ALCOA Laboratories, ALCOA Center, PA 15069, U.S.A.

Abstract

This article gives an overview of the metallurgical database of ALADIN, an expert system that aids metallurgists in the design of new aluminum alloys. Declarative structured representations in the form of schemata are used for metallurgical data and concepts. The representation is very general, as the goal has been to create a representation for all knowledge about aluminum alloys and metallurgy relevant to the design process. The alloy database and the architecture of the microstructure database is discussed in detail. The microstructure of alloys is described by an enumeration of the types of microstructural elements present along with their characteristics.

Ingemar A. E. Hulthage is a Research Associate in the Robotics Institute of the Carnegie-Mellon University. He earned his PhD in Theoretical Physics from Stockholm University 1981. Before he joined the Robotics Institute in June 1984 he had done research in physics at CMU, SUNY at Stony Brook, Niels Bohr Institute in Copenhagen and at Stockholm University and authored or coauthored eight publications in Theoretical Physics. During 1984-1987 Dr Hulthage worked on The Aluminum Alloy Design Inventor or ALADIN. His main research interest is now the use of scientific models in Artificial Intelligence.

Marek A. Przystupa received his MS in mechanical engineering from Technical University of Wroclaw, Poland in 1973 and a PhD in metallurgical engineering from Michigan Technological University in 1980. After his doctorate he undertook postdoctoral study at University of California, Los Angeles, then joined Alcoa Laboratories and is currently working as an independent consultant. He has been active in the areas of materials behavior simulations, computerized methods of materials design and quantitative methods of microstructure characterization and consulting on texture, anisotropy and modern materials design and processing methods. He is a member of the Metallurgical Society of AIME and American Society for Metals.

Martha Farinacci is a Staff Computer Scientist in Alcoa Laboratories. She holds a BS degree in mathematics from Rensselaer Polytechnic Institute and an MS degree in applied mathematics from Carnegie-Mellon University. She has extensive experience in the application of optimization, simulation, engineering analysis and numerical methods to problems in engineering and manufacturing. She has worked with the Robotics Institute of Carnegie-Mellon University on the application of artificial intelligence to aluminum alloy design. She is currently developing other knowledge based systems to support engineering design and manufacturing planning.

Michael Rychener is a research scientist in the Engineering Design Research Center at Carnegie Mellon University. He received a PhD in Computer Science from Carnegie-Mellon, preceded by an MS from Stanford University and a BA from Oberlin College. He has held research faculty positions at Carnegie-Mellon in Computer Science and in the Robotics Institute. Main research areas include knowledge-based expert systems for engineering design, rule-based languages, machine learning, user interfaces, and extensible tools for artificial intelligence applications.