【摘要】 电力负荷具有分形市场假说(FMH)的主要特征。根据分形拼贴定理,由分形插值方法求取一个吸引子与电力负荷历史数据相近的迭代函数系统(IFS),以建立分形预测模型,实现电力负荷预测。预测结果表明,该方法不存在收敛问题,数据收集简便,具有较好的实用价值。用电力负荷时间序列 分形分析进一步表明了预测值的合理性。更多还原

【Abstract】 According to the fractal collage theorem and with the fractal interpolation, the authors presentan iteration function IFS). Its attractor is similar to the history data of electric consumptionsoasto establish fractal forecasting ( model to realize electric power consumption. It is showed in the one swatch and 4 checkpoints that the maximal relative error is 3.96% and the average relative error is 0.015%. The fractal analysis of the electric load time series further shows that the forecasting value is reasonable and the electric load in the following 20 to 30 years will maintain a sustained and stable increase.更多还原