Transmission loss allocation in deregulated power system using the hybrid genetic algorithm-support vector machine technique

Abstract

This paper proposes a new method to trace the transmission loss in deregulated power system by incorporating the Genetic Algorithm (GA) and Least Squares Support Vector Machine (LS-SVM). The idea is to use GA to find the optimal values of hyper-parameters of LS-SVM and adopts a supervised learning approach to train the LS-SVM model. The proportional sharing method (PSM) is proposed to trace the transmission loss at each transmission line which is then utilized as a teacher in the proposed hybrid technique called GA-SVM method. Based on load profile as inputs and PSM output for transmission loss allocation, the GA-SVM model is expected to learn which generators are responsible for transmission losses. In this paper, 4-bus system and IEEE 14-bus system are used to show the effectiveness of the proposed method.