Selection of Hidden Layer Neurons and Best Training Method for FFNN in Application of Long Term Load Forecasting

Abstract

For power industries electricity load forecast plays an important role for real-time control, security, optimal unit commitment, economic scheduling, maintenance, energy management, and plant structure planning etc. A new technique for long term load forecasting (LTLF) using optimized feed forward artificial neural network (FFNN) architecture is presented in this paper, which selects optimal number of neurons in the hidden layer as well as the best training method for the case study. The prediction performance of proposed technique is evaluated using mean absolute percentage error (MAPE) of Thailand private electricity consumption and forecasted data. The results obtained are compared with the results of classical auto-regressive (AR) and moving average (MA) methods. It is, in general, observed that the proposed method is prediction wise more accurate.

Energy Demand Forecasting and Optimizing Electric Systems for Developing Countries

“…Primary consideration should be given to the number of hidden neurons and the selection of hidden and output layer activation functions [54]. Another study adopted a new method of selecting the hidden layer neurons of feed-forward neural networks [55]. Azadeh et al [16] proposed a seasonal ANN model for STLF that can easily handle non-linear and complex relationships, use exogenous variables as inputs, and handle corrupted data more effectively than regression models.…”

Multi-Objective Interval Prediction of Load Based on the Conditional Copula Function

“…Load prediction can be generally be divided into the following four categories according to the length of the forecast: long-term, medium-term, short-term, and ultra-short-term. Long-term load prediction is mainly used to predict the load situation in the next several years, generally used for grid planning and reconstruction work [6,7]; medium-term load prediction refers to prediction the load in the next few months to one year, mainly for reservoirs' operation scheduling, unit maintenance, and usage planning for fuel [8][9][10]; short-term load prediction is mainly for the next day to one week of load forecast, often used for optimizing the combination of water and thermal power and control of economic flow [11][12][13]; ultra-short-term load prediction is a hot spot of research in recent years [14,15], and mainly refers to load prediction for the next few minutes. Usually used for power quality control, safety monitoring for online operation, prevention and emergency control, its prediction accuracy directly affects economic dispatch, online safety monitoring, automatic generation control (AGC) frequency modulation, and preventive control emergency situations [16].…”

Application of Digital Filters to Check Quality of the Automatically Scaled Ionograms

“…The FIR filters is based on the moving average (MA) filter, which is usually used to eliminates the extreme deviations in data series(an example of using MA filter is given in [9]). For our purpose, filtered signals are compared with original data.…”