Forecasting Solar Photovoltaic power production at the aggregated system level

Author

Yue Zhang ; Beaudin, Marc ; Zareipour, Hamidreza ; Wood, David

Author_Institution

Dept. of Electr. & Comput. Eng., Univ. of Calgary, Calgary, AB, Canada

fYear

2014

fDate

7-9 Sept. 2014

Firstpage

1

Lastpage

6

Abstract

Solar Photovoltaic power production has grown significantly over the past few years. California ISO is the first system operator in North America to make the data for aggregated system-level solar power production across its territory available on a regular basis. In this paper, we demonstrate the application of three well-established forecasting models to 24-hour-ahead prediction of solar power at the system level. The models investigated in this paper include Auto Regressive Integrated Moving Average (ARIMA), Radial Basis Function Neural Network (RBFNN), and Least Squares Support Vector Machine (LS-SVM). Numerical results and discussions are provided based on California ISO solar power data.

Keywords

autoregressive moving average processes; economic forecasting; photovoltaic power systems; power markets; power system analysis computing; radial basis function networks; regression analysis; solar power stations; support vector machines; ARIMA; California ISO solar power data; LS-SVM; North America; PV market; RBFNN; aggregated system level; auto regressive integrated moving average; least squares support vector machine; photovoltaic solar power production systems; radial basis function neural network; solar photovoltaic power production forecasting; solar power prediction; system operator; Accuracy; Biological system modeling; Data models; Forecasting; Predictive models; Production; Training; ARIMA; Forecasting; LS-SVM; Prediction; RBFNN; Solar Power;

fLanguage

English

Publisher

ieee

Conference_Titel

North American Power Symposium (NAPS), 2014

Conference_Location

Pullman, WA

Type

conf

DOI

10.1109/NAPS.2014.6965389

Filename

6965389