Title :
Wind power ramp events classification and forecasting: A data mining approach
Author :
Zareipour, Hamidreza ; Huang, Dongliang ; Rosehart, William
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Calgary, Calgary, AB, Canada
Abstract :
Available wind power forecasting tools predict the future values of wind power production. System operators use those predictions to estimate the severity of wind power ramp up/down events, and determine the set of actions needed to manage those events. In this paper, a direct approach for predicting the severity of wind power ramp events is presented. Ramp events are categorized into `classes´, and available data are used to predict the class of future ramps. Support vector machines (SVM) are used as classifiers and an elaborate model for forming the set of inputs to the classifier is proposed. Numerical results based on the wind power data in Alberta, Canada, is presented.
Keywords :
data mining; power engineering computing; support vector machines; wind power; data mining; forecasting; support vector machines; wind power ramp events classification; Educational institutions; Forecasting; Mutual information; Support vector machines; Training; Wind forecasting; Wind power generation; Wind power; classification; data mining; feature selection; ramp event;
Conference_Titel :
Power and Energy Society General Meeting, 2011 IEEE
Conference_Location :
San Diego, CA
Print_ISBN :
978-1-4577-1000-1
Electronic_ISBN :
1944-9925
DOI :
10.1109/PES.2011.6039625
