摘要
提出了基于相似性原理的短期负荷预测方法 ,它有较高的预测可靠性 ,尤其适用于预测日天气状况与训练模式天气状况有较大差别的系统。运用相似性原理对人工神经网络的训练模式进行选择 ,使其与预测日有相似的气象特征 ,在此基础上 ,用选择出的相似训练模式对选定的人工神经网络进行训练 ,从而达到提高短期负荷预测精度的目的。对于特殊工作日而言 ,加上一个峰值估计环节后 ,该方法仍然适用。一个实际电力系统的算例证明了该方法的有效性。
A short-term load forecasting method based on analogous theory is proposed. It provides more reliable forecasts, especially when the weather conditions are different from those represented in the training data. To assure a further increase in forecast accuracy, it is necessary to pick out some learning patterns Which have the analogous characteristic as the forecasting day, after that, to use the picked patterns as learning samples for ANN. As for some special days, the method is still available in the case of adding an adjusted process-peak load evaluation. Applying the method to a practical power system, the results prove its competence.
出处
《电力系统自动化》
EI
CSCD
北大核心
2001年第23期45-48,共4页
Automation of Electric Power Systems
