Artificial neural network approach for electric load forecasting in power distribution company/ M. A. Hambali ...[et al.]

Hambali, M.A. and Saheed, Y. K. and Gbolagade, M. D. and Gaddafi, M. (2017) Artificial neural network approach for electric load forecasting in power distribution company/ M. A. Hambali ...[et al.]. e-Academia Journal, 6 (2). pp. 80-90. ISSN 2289-6589

This is the latest version of this item.


Download (17MB) | Preview
Official URL:


In recent years, there have been extensive researches seeking the best methods of improving the load forecast accuracy. Many of these methods are statistical based methods which include time series, regression, Box-Jenkins model, exponential smoothing and so on. However, the statistical models offer limpidity in data interpretation and sensible accuracy in load forecasting but characterized by the problems of limited modeling and hefty computational effort which makes them less desirable than the intelligent techniques. Recently, Artificial Intelligence (AI) has been a better substitute. Among the AI methods, artificial neural networks (ANNs) have got some attention from a lot of researchers in this area due to its flexibility in data modeling. In this paper, ANN for electric load forecasting is proposed. The historical data were collected for three months from Yola power transmission company office along Numan road Jimeta/Yola, Adamawa State, Nigeria. Researchers then performed data preprocessing on the data. Afterwards, data mining algorithms were applied in order to forecast electric load. In doing this, two ANN algorithms (MLP and RBF) and SMO algorithm were employed and compared. The results were then interpreted; the obtained models were analyzed to determine the pattern in load forecasting model. The experimental analysis was performed on WEKA version 3.6.10 environment. Also, 10-fold cross validation test option was used to carry out the experiments. Results obtained showed that multilayer-Perceptron model (MLP) gives an accuracy of 86% with Mean Absolute error (MAE) of 0.016, Radial basis function (RBF) had an accuracy of 76% with MAE of 0.030 and Sequential Minimal Optimization (SMO) accuracy of 85% with MAE of 0.090 which indicated a promising level of electric load forecast.

Item Type: Article
Subjects: H Social Sciences > HD Industries. Land use. Labor > Management. Industrial Management > Forecasting
T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Applications of electric power
Divisions: Universiti Teknologi MARA, Terengganu
Journal or Publication Title: e-Academia Journal
ISSN: 2289-6589
Volume: 6
Number: 2
Page Range: pp. 80-90
Official URL:
Item ID: 21354
Uncontrolled Keywords: ANN; MLP; RBF; SMO; Forecast; Electric load
Last Modified: 18 Dec 2018 07:46
Depositing User: Perpustakaan Cendekiawan UiTM Caw Terengganu

Available Versions of this Item

Actions (login required)

View Item View Item


Downloads per month over past year