Peak load forecasting of Electricity Generating Authority of Thailand by Gaussian Process

Author

Ploysuwan, Tuchsanai ; Atsawathawichok, Pramukpong ; Teekaput, Prasit

Author_Institution

Dept. of Electr. Eng., Siam Univ., Bangkok, Thailand

fYear

2014

fDate

19-21 March 2014

Firstpage

1

Lastpage

4

Abstract

This paper use Gaussian Process: GP to present about the peak electricity (Peak load) forecasting for the highest demand of electricity since 2011 to 2012 of “Electricity Generating Authority of Thailand” (EGAT) by use the data since 2000 to 2010 as a the training data. The four important variables are 1) time per month, 2) peak load per month, 3) GDP 4) GNP and present about how to compute the hyper - parameter θ which is the important variable that cause an efficiency forecast. The result of experiment has shown that the process which give few error and has more efficiency than Neural Network (NN).

Keywords

Gaussian processes; electric power generation; learning (artificial intelligence); load forecasting; neural nets; power engineering computing; power generation economics; EGAT; GDP; GNP; Gaussian process; NN; electricity demand; electricity generating authority of Thailand; hyperparameter computation; neural network; peak electricity load forecasting efficiency; peak load per month; time per month; training data; Economic indicators; Electricity; Indexes; Xenon; Gaussian Process; Load Forecasting; Neural Network; Peak Electricity Demand;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical Engineering Congress (iEECON), 2014 International

Conference_Location

Chonburi

Type

conf

DOI

10.1109/iEECON.2014.6925858

Filename

6925858