Short term load forecasting using artificial neural network
2017 Fourth International Conference on Image Information Processing (ICIIP)
Forecasting of electrical load is very crucial to the effective and efficient operation of any power system. This is achieved by obtaining the most accurate forecast which help in minimizing the risk in decision making and reducesthe costs of operating the power plant. Therefore, the comparative study of time series and artificial neural network methods for short term load forecasting is carried out in this paper using real time load data of Covenant University,withthe moving average,
... average, exponential smoothing (time series method) and the Artificial Neural Network (ANN) models. The work was done for the day-to-day operation of the soon-to-becompleted power station of the university. For each of the methods, models were developed for the load forecast. The Artificial Neural Network proved to be the best forecast method when the results are compared in terms of error measurementswith amean absolutedeviation(MAD) having 0.225, mean squared error (MSE) having 0.095 and the mean absolute percent error(MAPE)having 8.25. www.iosrjournals.org 73 | Page management and is important for the electricity industry in the deregulated economy. It has many applications which includes energy purchasing, generation, load switching, contract evaluation and infrastructure development. A large variety of mathematical methods have been developed for load forecasting  . An accurate load forecast can be very helpful in developing a power supply strategy, finance planning, market research and electricity management  . For every forecast, there are different factors to be put into consideration. These factors to a great deal determine how accurate the forecast will be, as well as determine the load demand and hence, affect the load curve. These factors include calendar effects, seasonal variations, weekday variations, weekend-days variation, weather and temperature amongst others. Calendar effects include the effects of working days or trading days and holidays. Growth in the economy, population, extreme weathers may also contribute to annual variations. These variations need to be considered in forecasting energy intended for domestic use  . Accurate load forecasting holds a great saving potential for electric utility corporations  . The goal of any forecast is to obtain the forecast with the least error. During forecasting, an underestimation in energy demand may result in a limited supply of electricity at the consumer end, which leads to energy quality reduction in system reliability. On the other hand, an overestimation may cause unnecessary investments to the establishments therefore resulting in uneconomical operating conditions  . In  , two time series models (multiplicative decomposition model and the smoothing techniques) were used for the short term load forecast. Moving Average and Exponential Smoothing Techniques were used for the load forecast of UniversitiTeknologi PETRONAS (UTP), Malaysia. In an ANN model was developed for the short term load forecast of the 132/33KV sub-Station, Kano, Nigeria. The Levenberg-Marquardt optimization technique which is one of the best for training the network was used as a back propagation algorithm for the Multilayer Feed Forward ANN model. In  , the short-term load pattern for the University of Ibadan was investigated and a multi-layered feed-forward artificial neural networks (ANN) model was developed to forecast the time series half-hourly load pattern of the system.