Condition Assessment of Power Supply Equipment Based on Kernel Principal Component Analysis and Multi-class Support Vector Machine

The power supply equipment is very complicated, which makes the process of assessment too long. The traditional method of assessment is also not comprehensive enough, which induces low accuracy of assessment. A method based on kernel principal component analysis and fast multi-class support vector machine is introduced in this paper: kernel principal component analysis, as the preprocessor of the index system, analyses the most important factors which influence equipment condition. Then multi-class support vector machine, as the assessment tool, can classify power supply equipments as per the requirements of condition based maintenance. The result of experiment shows that the method can reduce the complex of assessment and is more comprehensive. It also improves rapidity and accuracy of traditional assessment.