Study on the voltage qualification rate prediction from the voltage contour situation picture based on support vector machine

A novel picture-based situation assessment method is proposed to predict the voltage qualification rate (VQR) of substation-centralized distribution network (SCDN) by using the feature vector of voltage contour situation picture (VCSP). At first, the feature vector denoted in X of VCSP is proposed, and the VQR of SCDN denoted in Y is calculated by counting the total number of load node inside the upper bound and lower bound voltage contour lines, which is an accurate method through the diagram information of VCSP picture; then the relationship between X and Y is explored by support vector machine (SVM) method. Genetic algorithm (GA) is proposed to select the parameters of SVM algorithm. The clustering of VCSP samples is used to improve the accuracy of SVM processing. Case study has proven the effectiveness and efficiency of provided method.