A case-based reasoning system for identifying failure mechanisms

Correctly identifying the mechanism responsible for a failure is a major step in failure analysis. Today, human experts normally perform this task. In the problem-solving process, human experts often recall similar cases to help identifying the mechanism involved. This has motivated the use of case-based reasoning to develop a computerized system for failure-mechanism identification in this study. Major issues and the methods applied are discussed. To determine its accuracy, the system is subsequently evaluated using historical cases, which are classified into two categories: standard and exceptional. The test results show that 100% accuracy can be achieved for standard cases, and that exceptional cases also attain accuracy as high as 71.25%. It is thus concluded that case-based reasoning is a viable approach for the identification of failure mechanisms. (C) 2000 Elsevier Science Ltd. All rights reserved.