Super Short-term Load Forecasting By Considering Grid-connected Wind Power

Wind power is generally treated as"negative load"when considering the characteristics of grid-connected wind power. Based on citing the conception of"equivalent load"which means load minus wind power, this paper studied a method of super short-term"equivalent load"forecasting affected by grid-connected wind power. That is, forecasting the wind power and load respectively, and putting them together to obtain forecasting"equivalent load". At first, this paper presented a new method to deal with non-stationary wind speed series in order to obtain several approximate stationary time series with wavelet decomposition theory. And then, an output power forecasting model of wind farm was built by considering wake effect. Secondly, in order to select forecasting samples in super short-term load forecasting, dynamic cluster was used and LS-SVM (least squares support vector machines) method was also applied to train forecasting samples. At last, an example of super short-term"equivalent load"forecasting by considering grid-connected wind power is provided, and Matlab 6.5 was also used to demonstrate the validity of this method.