Neural Network Effective Approach for Energy Load Forecasting

T. Rashid and T. Kechadi (Ireland)


Neural Networks, Feed Forward and Feed Back Multi Context Neural Network, Energy Load Forecasting.


This paper presents the feed-forward and feed-back multi context neural network (FFFB-MCNN) to forecast the daily peak load for two large power plant systems. Weather component variables are the key ellements in forecasting because any change in these variables affects the demand of the energy load. Thus, the FFFB-MCNN is used to learn the relationship among past, previous, and future ex ogenous and endogenous variables. Experimental results show that using the change in weather components and the change occurred in the past load as inputs to the FFFB MCNN rather than the basic weather parameters and past load itself as inputs to the same network produce better ac curacy of predicted load. Experimental results also show that using exogenous and endogenous variables as inputs is better than using only the exogenous variable as inputs to the network.

Important Links:

Go Back