Research on Diagnosis of AC Engine Wear Fault Based on Support Vector Machine and Information Fusion

• DOI: 10.4304/jcp.7.9.2292-2297
• Publication Date: 2012-09-26T23:40:37Z
• Authors: Lei Zhang; Yanfei Dong
• Keywords: 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology

Support Vector Machine (SVM) and information fusion technology based on D-S evidence theory are used to diagnose wear fault of AC engines. Firstly, based on a number of frequently used oil sample analysis methods for detecting engine wear fault, establish corresponding sub SVM classifier. The classifier can reflect the mapping relation between fault symptoms and fault types and achieve the result for a single diagnosis item. And then, use D-S evidence theory to make information fusion over result for a single diagnosis item so as to make fault diagnosis. With diagnosis of AC engine wear fault serving as example, example testing is performed. The result shows that in comparison with conventional methods, the combination of SVM and information fusion technology is fast and effective, suitable for diagnosis of AC engine wear fault.