【作者】 李历波；

【导师】 孙才新；

【作者基本信息】 重庆大学 ， 电气工程， 2002， 硕士

【摘要】 在配电网规划中，不仅要预测负荷的量，还要预测负荷的位置，即负荷的空间分布，只有在确定负荷空间分布的基础上，才能进行配电网的布点和布线规划。本文利用地理信息系统（Geographic Information System，）空间信息分布的功能，研究了配电网空间负荷短期预测方法。将GIS的空间信息分析功能应用于配网空间负荷预测的研究：综合利用灰色理论及遗传规划（Genetic Programming，GP）的优点，提出了一种根据馈线的历史峰值负荷进行外推的组合SLF法：首先采用负荷耦合回归法来修正负荷历史，消除由于负荷转移引起的误差；然后根据实际馈线路径和配网GIS图形分层，运用图层叠加分析得到馈线的实际供电范围；接着采用灰色关联度聚类方法对馈线负荷增长曲线进行聚类分析；最后采用GP来对灰色聚类结果进行符号回归，分别得到每一类曲线的最佳拟合曲线形式。实际工程应用表明该种方法优于其它趋势法，特别适用于计算资源有限的情况。更多还原

【Abstract】 The quantity and its location of loads should be forecasted in the distribution network planning, which also means loads spatial distribution. The nodes and lines planning could be done only after determine the spatial loads distribution correctly.Based on the Geographic Information System ( GIS ) spatial information analysis function, we study on the SLF. We propose a combined SLF method to extrapolate feeder load growth by using feeder’s history peak value and the merits of gray theory and Genetic Programming ( GP ). At first, we adopt load transfer coupling method to correct load history and its error for load transfer. Secondly, we get the real power-supply area by using layer overlap analysis, based on practical feeder path and distribution GIS map layer. Thirdly, we adopt gray incidence degree method to cluster the feeder load growth curve. At last, we get the optimized curve fitting of every cluster by using GP method to process the gray cluster results. The real instance shows that this method has the better result than other trending methods, and it is fit for the limited computer resource especially.更多还原