基于数据变换技术和无偏NGM(1,1,k)模型的电力负荷预测

Power Load Forecasting Based on Data Transformation and Unbiased NGM(1,1,k) Model

针对电力负荷数据以近似非齐次指数规律增长时采用GM(1,1)及其各类改进模型模拟误差较大问题,基于无偏NGM(1,1,k)模型的建模机理,采用函数变换方法作为数据变换技术处理负荷数据,并以预测精度作为数据变换技术优选条件。实例应用结果表明,该方法有效可行。

GM （1, 1） model and its improved model have large modeling errors for load forecasting when load data appear approximation non-homogeneous exponential growth. Based on unbiased NGM （1,1, k） modeling mechanism, the function transformation method is used as a data conversion technology to deal with load data and the predict accuracy is taken as the optimal conditions for data transformation. The instance results show that the proposed method is feasible and effective.