Next Day Load Demand Forecasting of Future in Electrical Power Generation on Distribution Networks using Adaptive Neuro-Fuzzy Inference

Abstract

This paper presents the development an Adaptive Neuro Fuzzy inference control system for purpose of improving power system is an application of Artificial Neural Network (ANN) and Fuzzy Logic based hourly load demand forecasting with linear polynomial and exponential equation. The ANN involved is designed using the multilayer back propagation learning. The Fuzzy Logic and the ANN input layer receives information on next day maximum temperature, period and hourly load. The class of day type, the hourly load in present day. The Fuzzy Logic and the ANN output layer provides the predicted hourly load. Test is performed using data of hourly load of Bangkok forecasting for 24 days. The results show that ANN model has mean absolute percentage error of 1.7 % and the Fuzzy Logic model has mean absolute percentage error of 1.5 %. The accurate results of the forecasting will improve the power system security and save generation cost.Index Terms -Hourly load demandforecasting, power system, adaptive Neuro-Fuzzy inference, next day load.

Load demand forecasting: Model inputs selection

“…The model performance has been compared with a multilayer preceptron and Kohonen Classification and Intervention Analysis. Next day demand forecasting in electrical power generation has been developed using ANFIS [4]. The purpose was to improve the power system as an application of artificial neural networks and fuzzy logic based hourly load demand forecasting with linear polynomial and exponential equation.…”