APPLICATION OF HYBRID MODEL IN DAILY PEAK POWER LOAD PREDICTION PROBLEM

Abstract

The problem of prediction of P_max and P_min of power daily load is one of fundamental tasks in power system operation and control. There are already a number of different models and solutions including linear and nonlinear ones. In this paper we propose a new hybrid approach including both linear and nonlinear estimators in order to be able to model the complex dependencies between future power daily loads and the data from the past. The Singular Value Decomposition algorithm will be used to estimate the linear relationship between signals where the classic neural network MLP will be used to estimate the nonlinear features. The proposed solution was test on real data taken from Thai Nguyen Power Company.