Electricity Load Forecasting Using Artificial Neural Networks

Load forecasting is an essential part of an efficient power system planning and operation. This research work is on short term electricity load forecasting using Artificial Neural Network (ANN) and Ogbomoso a city in Nigeria is considered as a case study. Input variables considered are past loads history, hours of the day and days of the week, while the output is the forecasted load for 24 h ahead. The training tool Neurosolution was employed in simulating and designing the feed forward back propagation forecasting network. Result obtained shows that electricity load can be predicted ahead of time also, it can also be inferred from this research that, load forecast using neural network is somewhat intelligent in that it gives real values even when the past load history is of zero value. This shows that areas without constant supply of electricity can still forecast future loads with a reasonable error margin so as to help in better load distribution and effective load shedding planning.