基于多尺度空间分辨率的空间负荷预测误差评价方法

Method of Spatial Load Forecasting Error Evaluation Based on the Multi-scale Spatial Resolution

针对如何评估空间负荷预测(spatial load forecasting,SLF)中空间误差的难题,提出了基于多尺度空间分辨率的空间负荷预测误差评价方法。该方法考虑到SLF的局部预测误差会对整个预测区域产生影响,从如何计算局部预测误差给其他位置预测值带来影响的问题入手,将极端场景作为边界条件分析SLF结果的误差地图,揭示了在不同空间分辨率下,对相同的空间预测误差的感知程度多数情况并不一样,甚至还可能完全相反的演变规律,建立了针对既有网架结构、供电容量和规划目标年负荷水平的SLF误差评价机制,给出了SLF误差评价指标。运用该方法对东北某市的一个行政区进行空间负荷预测的工程实例表明,所提出的SLF误差评价方法是行之有效的。

Faced with the difficult problem how to evaluate spatial error in spatial load forecasting（SLF）, the error evaluation method of SLF based on the multi-scale spatial resolution was proposed. Considering local forecasting error impacts on the SLF error of the whole, how to calculate local forecasting error impact on the predicted load value of another location was studied firstly. The extreme scenarios were used as boundary conditions to analyze the error map of SLF results. Under different spatial resolutions, it is brought to light that the degree of perception is different for the same spatial error and the evolution regular changes to be even completely opposite. Aiming at the structure and the supply capacity of existing power grid, and the load level of planning target-year, spatial error evaluation mechanism was established. At the same time, the evaluation index for SLF error was provided. The method of spatial load forecasting error evaluation based on the multi-scale spatial resolution was applied to spatial load forecasting in an administrative region of a northeast city, and is proved to be effective.