The factor analysis of short-term load forecast based on wavelet transform

In recent years, the wavelet analysis method, which can help people acquire the needed signals components with different frequencies by decomposition in different scales, has been applied comparatively widely to many fields including power systems. By using the wavelet transform, the short-term load forecast based on wavelet transform extracts different components those are the loads that change in varied way, namely low-frequency components and high-frequency components. For low-frequency components, loads are analyzed and forecasted by the time-sequence method. For high-frequency components, after the key influential factors are acquired by correlation analysis, the corresponding model is built for BP forecasting. Then the whole forecasted loads are gained by combining those foregoing results. By analysis and comparison, the method can enhance the. feasibility and validity of load components decomposed, and then increase the accuracy of load forecasting.