摘要
分析了燃气负荷的形状相似、趋势相似和特征相似,根据节假日燃气负荷的特点,提出基于相似日的节假日负荷预测方法。以前趋势相似性和日特征相似性为原则,在燃气负荷历史数据库中选择若干相似日,以相似日作为样本采用神经网络方法进行预测。实例证明该方法具有良好的预测精度。
The shape similarity, trend similarity and characteristic similarity of gas load are analyzed, and according to the feature of gas load in holidays, a method of gas load prediction in holidays based on similar days is put forward. Taking the former trend similarity and daily characteristic similarity as principle, some similar days are selected in the history database of gas load, and taking the similar days as sample, the prediction is made with the neural network . The example demonstrates that this method has better prediction precision.
出处
《煤气与热力》
2006年第5期10-14,共5页
Gas & Heat
