Research on optimal transformer maintenance scheme based on lLS-SVM

Jian Liu,Kaipei Liu,Shijie Zhou,Wei Guo

2012-01-01

Abstract:Optimization of transformer maintenance scheme can enhance substation reliability and lower maintenance cost. Based on least squares support vector machines (LS-SVM) method, combing with cost-effect effectiveness method, research on optimized transformer maintenance scheme was carried out. The paper built a transformer deficiencies tree from statistical data. Aggregating deficiency data, outage duration, maintenance operation deficiency checking out rate and etc as basis, a two layer dynamical adjustment optimization method was used to choose LS-SVM parameters, then LS-SVM algorithm was used to predict deficiencies before and after scheme optimization, and cost effect measure was used to determine the best scheme. This algorithm has been used to evaluate transformer status in a power supply company and modify transformer maintenance scheme, and ideal result is achieved. © 2012 Chin. Soc. for Elec. Eng.