Royal Military College, Kingston, Ontario
University of Southampton, Southampton,
Pages: 14 Published: Jan 1997
The complexity of many materials engineering problems can be drastically reduced by regrouping concepts and facts into a natural human representation provided by object-oriented models. One of the main advantages of using the principles of object orientation (OO) to structure a knowledge based system (KBS) is the compatibility of OO with natural cognitive operations such as classifying, instantiating, and measuring. But knowledge representation goes much beyond conventional data representation and must rely on facts that often possess a halo of fuzziness. In this project, the OO structure was used to organize the complex world of materials degradation and lifetime prediction. A system developed with this principle was found to be able to incorporate the fundamental knowledge derived from failure analysis or from other fields of materials engineering. Such a framework could also constitute the crucible to achieve the data fusion necessary to blend various sources of expertise. This paper describes the key elements of the framework and illustrates its flexibility by providing a guiding structure during the analysis of the results of a computerized search of modern engineering literature.
knowledge representation, lifetime prediction, object oriented modeling, corrosion data, failure analysis
Paper ID: STP13767S