Due to the complication and nonlinearity of power load forecasting,it is difficult to obtain accurate results by using the traditional mid-long term forecasting model. To improve the forecasting accuracy,a multivariable time series inversion self-memory model is constructed. The proposed model uses grey correlation analysis to select the main influ?encing factors of the power load,and adopt them to perform the dynamic equation inversion for the variation of power load. Moreover,the fitting and forecasting of power load is realized by combining the self-memory model. In this way, the forecasting precision is improved and the forecasting results can reflect the inherent variation characteristics of his?torical power load data to the maximum extent,which improves the stability of fitting and forecasting. To verify the effec?tiveness of the proposed model,the total electricity consumption data in a certain region from 1986 to 2002 are used as training samples to conduct fitting analysis,and further forecast the total electricity consumption in years 2003-2006. The fitting and forecasting results prove the validity and feasibility of the proposed model in the mid-long term power load forecasting.%电力负荷预测的复杂性、非线性使传统的中长期预测模型难以获得精确的结果.为了提高中长期电力负荷预测准确度,构建了多变量时间序列反演自记忆模型.该模型使用灰色关联分析选取电力负荷变化主要影响因素,采用主要影响因素对电力负荷自身变化过程进行动力方程反演,并结合自记忆模型,实现对电力负荷数据的拟合与预测.在提高预测精度的同时,使预测结果最大程度地体现历史电力负荷数据的内在变化规律,提高拟合和预测的稳定性.为了验证模型的效果,使用1986—2002年某地区全社会用电量数据作为训练样本,进行拟合分析,并预测2003—2006年全社会用电量.拟合和预测的结果证明了该模型在中长期负荷预测中的有效性和可行性.
展开▼
