数据驱动的中小河流智能洪水预报方法对比研究

发布时间：2018-03-22 17:09

本文选题：数据驱动模型　切入点：中小河流洪水预报　出处：《中国科学技术大学学报》2016年09期 　论文类型：期刊论文

【摘要】：近几年在洪水预报中,数据驱动洪水预报模型得到了广泛的应用,并取得了良好的效果,但是数据驱动模型大都用于大流域,很少用于小流域.中小河流大多位于资料短缺的山丘区,洪水具有突发性强,汇流时间快,预见期短的特点.为此分别构建了SVM模型、BP神经网络模型、RBF网络模型、极限学习机(ELM)模型,并利用所构建的模型对昌化流域进行预报;结果表明,SVM模型和RBF网络模型在低流量区段预测较准确,而且模型预报稳定;BP神经网络模型在高流量区段较准确,但是模型预报结果不稳定;ELM模型预报误差较大,而且预报不稳定;于是采用组合模型方式:低流量区段采用SVM模型或RBF网络模型,高流量区段采用BP神经网络模型,实验结果表明组合模型预报效果更好.
[Abstract]:In recent years, the data-driven flood forecasting model has been widely used in flood forecasting, and has achieved good results, but the data-driven model is mostly used in large watershed. It is seldom used in small watersheds. The small and medium rivers are mostly located in the hilly areas where the data is short. The flood has the characteristics of sudden strong, fast confluence time and short forecast period. Therefore, the SVM model and the BP neural network model are constructed respectively. The RBF model is used to predict the Changhua River Basin, and the results show that the SVM model and the RBF network model are more accurate in the prediction of the low flow area. Moreover, the model prediction is stable and the BP neural network model is more accurate in the high flow region, but the model prediction results are unstable and the prediction error of ELM model is large and the prediction is unstable. So the combination model is adopted: the SVM model or RBF network model is used in the low flow section and the BP neural network model is used in the high flow area. The experimental results show that the combined model is more effective.
【作者单位】：河海大学计算机与信息学院;
【基金】：公益性行业科研专项(201501022) 江苏省重点研发计划(BE2015707)资助
【分类号】：P338;TP183

【相似文献】