Short-Term Load-Forecasting Method Based on Wavelet Decomposition With Second-Order Gray Neural Network Model Combined With ADF Test

Abstract

Improving the accuracy of power system load forecasting is important for economic dispatch. However, a load sequence is highly nonstationary and hence makes accurate forecasting difficult. In this paper, a method based on wavelet decomposition (WD) and a second-order gray neural network combined with an augmented Dickey–Fuller (ADF) test is proposed to improve the accuracy of load forecasting. First, the load sequence is decomposed by WD to reduce the nonstationary load sequence. Then, the ADF test is adopted as the test method for the stationary load sequence of each decomposed component after WD in which the test results determine the best WD level. Finally, because forecasting the wavelet details characterized by high frequencies is difficult owing to its fluctuation, a second-order gray forecasting model is used to forecast each component after WD. Furthermore, to obtain the optimum parameters of the second-order gray forecasting model, the neural network mapping approach is used to build the second-order gray neural network forecasting model. The simulation result of a real load sequence verifies that the method proposed in this paper can effectively improve the load-forecasting accuracy.