基于缓冲修正的灰色插值负荷预测建模与仿真

Mid-term Load Forecasting and Simulation Based on Buffer Operator and Modified Grey Model

研究电力负荷准确预测问题。由于电力负荷影响因子多,如气象条件、社会影响、系统运行状态和增容决策等因素,并且上述影响因子含有噪音信息,故而电力负荷与影响因子之间呈现复杂非线性关系,传统预测方法无法刻画其变化规律,导致电力负荷精度低。为了解决上述问题,在传统GM(1,1)模型的基础上,根据未来负荷趋势的判断,利用平均弱化缓冲算子(AWBO)对历史数据进行修正,并对区间内的负荷分量运用组合插值方法对进行重构,构建改进的灰色预测模型可以更好的应对电力负荷影响因子带来的非线性干扰。最后给出改进预测方法的建模步骤。通过实例验证表明,改进算法在预测精度、泛化能力及收敛速度方面较突出,具有良好的预测效果。

Study power load forecasting. An important feature of smart grid is the intelligent power distribution function based on electricity consumption forecasting with high accuracy. Accurate prediction of electricity consump tion is the key indicator of power intelligence. This paper aimed at the shortcomings of traditional GM （ 1,1 ） model, and corrected the historical data by average weakening buffer operator （AWBO） according to prediction of electric load trend. The combinative interpolation was used to rebuild the background value and form the improved GM （ 1, 1 ） prediction model. The improved model overcomes the instability of traditional prediction model. Finally, the mod eling procedure of the proposed model was presented. In the experiment research, the experiment result demonstrates the feasibility and validity of the proposed model.