Hybrid Based Artificial Intellegence Short –Term Load Forecasting
Kayode O. Adebunmi
*
Department of Electronic and Electrical Engineering, Ladoke Akintola University of Technology, Nigeria.
Temilola M. Adepoju *
Department of Computer Engineering, Federal Polytechnic Ede, Nigeria.
Gafari A. Adepoju
Department of Electronic and Electrical Engineering, Ladoke Akintola University of Technology, Nigeria.
Akeem O. Bisiriyu
Department of Electronic and Electrical Engineering, the Polytechnic Iree, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
Electrical power load forecasting, which forms a key element in the power industry's electricity preparation, is used for providing required data for day-to-day system management activities and power utility unit participation. Since the statistical method is a linear model, and the load and meteorological parameters have a nonlinear relationship, the statistical method for load forecasting involves a great calculation time for parameter recognition. Using this tool for load forecasting often results in a major mistake in prediction. Due to the disadvantages of the statistical method of load forecasting Neuro-fuzzy model was used in this work. Three models: Adaptive Neuro-Fuzzy Inference System (ANFIS), Artificial Neural Network (ANN) and Multilinear Regression (MLR) were simulated in MATLAB environment and their output results were compared using root mean square error (RMSE) and mean absolute error (MAE). The ANFIS model outperforms the other models with least errors of RMSE and MAE of 2.2198% and 1.7932% respectively.
Keywords: Load forecasting, electrical load, electricity, neuro-fuzzy model, and artificial intelligence