【作者】 唐琼玲；

【导师】 李如琦；

【作者基本信息】 广西大学 ， 农业电气化与自动化， 2006， 硕士

【摘要】 本文首先对了电力负荷预测的定义、目的和意义等相关方面进行简单地介绍，然后分析了国内外中长期电力负荷预测的发展现状，指出了目前中长期电力负荷预测方法存在的不足，以及导致预测精度不高的原因，并针对这些原因提出了基于反馈的中长期电力负荷组合预测方法。本方法由五个部分组成：①历史资料的搜集：本文提出了两种方案分别解决历史资料中数据缺失以及数据不平滑的情况；②根据预测电网的情况选择多种单一负荷预测方法进行预测，采用组合预测方法求解预测值。本文提出了三种求解组合预测模型的方法：线性规划、人工神经网络、遗传算法。并对每一种方法都进行了详细地推导、分析、比较以及算例分析；③考虑到了由于不确定因素导致负荷预测结果存在的误差。根据能否搜集到不确定因素的资料提出神经网络法求解或可变权组合误差法求解。然后再将求解出的误差结果反馈到组合预测结果中。这一方法的使用，使得预测结果的精度有了很大程度的提高；④通过概率性分析，确定中长期负荷预测结果的波动范围，也就是置信区间；⑤对本文提出的“基于反馈的中长期电力负荷组合预测方法”模型进行校验。选用的校验方法有：绝对误差、相对误差、均方根误差、后验差检验等。 最后，用本文提出的“基于反馈的中长期电力负荷组合预测方法”对广西电网进行中长期电力负荷预测，并对预测结果进行分析、校验。证明了该方法能较好的提高中长期电力负荷的预测精度，具有较高的实用性、科学性。更多还原

【Abstract】 This paper first gives a brief introduction of the definition, goals and significance of electricity load forecasting, then makes an analysis of the current conditions and prospects of mid-long term electricity load forecasting both at home and abroad and also points out the insufficiency in the method of mid-long term electricity load forecasting and the reasons of the low accuracy. This paper puts forward the method of mid-long term electricity load synthetic forecasting based on feedback based on the above reasons and consists of the following five parts, namely, I. Collection of previous data: two schemes are mentioned to solve the problems of the data insufficiency and rough from the previous data; II. On the basis of the current situations of forecasted power network, the paper points out the method of taking advantage of some single load forecasting method and synthetic forecasting method to get the forecasting value. There are three methods of obtaining synthetic forecasting model, namely, linear programming, ANN, GA with detailed derivation, analysis, comparison and example analyzing; III. Taking the error of load forecasting result caused by some uncertain factors into consideration and according to the fact whether the uncertain factors can be collected, two plans-ANN and variable weight synthetic error method are proposed in the paper. Then the result will be checked as the feedback used in the synthetic forecasting, which can increase the更多还原