基于小波变换的短期风速预测综合模型

Integrated models of short-term wind speed forecasting based on wavelet transform

针对风速序列具有非平稳性、非线性、异方差性的特点,首先利用db3小波对原始风速序列进行多分辨率分析,并对其系数进行单支重构,得到表征风速序列内在特性不同频段上的概貌风速与细节风速;其次对不同频段上的风速序列建立时间序列主模型,采用LM检验法分析所建模型的残差序列,提出用ARCH模型和GARCH模型进行改进,更贴近实际地反应了风速变化的规律;最后通过实例验证该文方法能够有效提高预测精度。

Aiming at the characteristics such as non-stationary,non-linear and heteroscedasticity of wind speed sequence, this paper firstly adopts db3 wavelet to add multi-resolution wavelet decomposition to original wind speed sequence and single branch reconstruction to its parameters, and as a result, approximate wind speed sequence and detailed wind speed sequence on different frequency bands representing internal characteristics of wind speed se- quence are acquired. Secondly, this paper builds time-sequence master model for wind sequences in different fre- quency ranges, adopts LM test to analyze residual series of the established model, and eventually come up with the idea that ARCH and GARCH model can be used to improve the exsiting methods, so the change pattern of wind speed is well reflected. Lastly, a practical example is used to test the improved method which has proved that the method in this paper is an effective way to improve prediction accuracy.