| Title |
24-Hour Load Forecasting Algorithm Using Artificial Neural Network in Summer Weekdays |
| Authors |
Kyeong-Hwan Kim ; Rae-Jun Park ; Se-Won Jo ; Kyung-Bin Song |
| DOI |
http://dx.doi.org/10.5207/JIEIE.2017.31.12.113 |
| Keywords |
Load Forecasting ; Artificial Neural Network ; Artificial Intelligence |
| Abstract |
The short-term load forecasting is essential for efficient operation of electricity market, economical operation plan of generators, and prevention of outage events. 24-hour load forecasting algorithm using neural network algorithm is proposed. The input parameters of the artificial neural network are composed of time index, hourly load data, hourly temperature data for the day before the forecasting day and hourly load data, hourly temperature data for two days before the forecasting day. The output parameters are hourly load data on the forecasting day. The artificial neural network training is performed using the past 28 training cases. The min-max normalization is used as a normalization method of input parameters such as time index, hourly load data, and hourly temperature data during training. The case studies show that the average percentage errors of the proposed algorithm are improved comparing with errors of the exponential smoothing method. The proposed algorithm is expected to contribute to the efficient operation of power system and electric power market by providing more accurate predictive load value of day ahead electricity demand. |