基于SSA-PPR模型的澜沧江枯季通航流量预测

Navigable Runoff Prediction in the Dry Season of the Lancang River Based on SSA-PPR

随着国家"一带一路"战略的推进,云南省要充分发挥连接东南亚国际水上大通道的作用,为促进澜沧江航运的正常发展,对其通航保障能力进行研究具有重要的现实意义。足够的来水量是河道通航的重要条件之一,开展河道的最枯月径流量预测,对航运保障具有重要的参考作用。为提高预测精度,笔者提出首先使用SSA分析法对周期性较强影响因子进行重构,与其他因子混合后,再与PPR相结合形成SSA-PPR预报模型,并对澜沧江的年最枯月径流量进行预测验证,结果表明SSA-PPR模型训练期平均相对误差的绝对值为2.5%,检验期平均相对误差的绝对值为6.2%,精度为甲等,相对于未使用SSA混合因子前,其预测效果更好,可以满足澜沧江枯季通航流量预测的需要。

With the promotion of China＇s ＂the Belt and the Road＂strategy,Yunnan Province is supposed to play its role in connecting Southeast Asian international channels. To promote the development of the Lancang River＇s shipping,it＇s necessary to study its navigation support ability. Sufficient runoff is one of the most important conditions for a river to be navigable,the prediction of the minimum monthly runoff in the dry season of rivers is an important reference for the protection of shipping. In order to improve the accuracy of prediction,we select the impact factors with strong periodicity firstly,then reconstruct themselves by using SSA and mix them with other factors. And then,we can turn both of them into a new model called SSA-PPR. Finally,we predict the minimum monthly runoff in the dry season of the Lancang River by using this special model and verifiy it. The result of using the SSA-PPR model shows that one absolute value of the average relative error is 2.5% during the training and the other is 6.2% during the test,the accuracy is Class A. It is better than PPR without mixing factor by the SSA. It can meet the needs of the dry season runoff prediction for the Lancang River＇s navigation.