CAI system for inheritance of maintenance expertise

In complex technological systems, inheritance of maintenance expertise from veterans to novices is a serious problem, because no effective methods are known. Maintenance experts require expertise that enables them to identify causes of trouble and to detect relevant symptoms before trouble actually occurs. The aim of this study is therefore to propose a scheme of knowledge representation that can be used for supporting inheritance of maintenance expertise based on a clear framework of design knowledge. We adopted a representation of design knowledge where design specifications are derived from design requirements and engineering knowledge. Another important information source for maintenance expertise is past cases of troubleshooting. We represented a trouble case as a sequence of abnormal conditions of the target component. We propose a representation scheme that can deal with both design knowledge and trouble cases in the common framework. We develop a prototype CAI system, on which users can add hypertext documents in relation to the background knowledge using the framework. This enables users to grow the system by accumulating technical know-how. The system was tested using design data and trouble case data of the main coolant pump of a nuclear reactor to verify the functionality of the system