Research On Case-Based Reasoning For Fault Diagnosis

Case-based reasoning technology is the emerging field of artificial intelligence as a method of problem solving.With the Electro-hydraulic equipment system to large, complicated, systematic and in the direction of automation, maintenance personnel need a higher demand to analyse Fault Sources and locate Faults. Therefore, case-based reasoning on the fault diagnosis will play very important role in improving the skills of maintenance personnel, shortening the training cycle, saving maintenance costs.Based on the analysis of the content and method of cases, this article uses a new object-oriented representation method of cases which combined with database technology. The hierarchy model of fault case has been built by using the concept of abstract cases. Made a detailed design on case'structure and organization. Used three-level-structure to organize case-base: system-level case-base, typical case-base and specific sub-case-base. Adopted maximum-similarity-method to classify the original case-base to construct each specific sub-case-base; used calculating"max sum among cases"to search the typical case in each sub-case-base.Based on Euclidian Distance, using normalized function to improve the similarity calculation method.A method based on the Rough Sets for is applied and improved in order to overcome problems of the existing algorithms such as poor objectivity and high complexity.The key technologies of the system have been discussed comprehensively, which contains: case retrieving, modifying, learning and case-base maintaining. Among those key technologies, case retrieving is the core step. The system used phasic-nearest-neighbour strategy to retrieve cases in case-base. This strategy combining with two-level-structure highly reduced retrieving times, so that improved the efficiency of search.On the platform of Windows XP + Microsoft SQL Serveer 2000 + ADO database technology, finished Fault diagnosis system based on Rough Sets and CBR in Visual C++6.0 and MATLAB mixed programming environment. The reasoning algorithm and operating of the prototype system is verified. Finally, the whole ariticle is summarized as well as the next necessary work is pointed out.