人民珠江2024，Vol.45Issue(12)：114-121,8.DOI:10.3969/j.issn.1001-9235.2024.12.012
基于APSO和SVM的滨海河网冲淤需水量预测模型研究
Prediction Model of Water Demand for Scouring Siltation in Coastal River Networks Based on APSO and SVM:A Case Study of Doulong Port
摘要
Abstract
Siltation in coastal river networks is a serious water security issue,and balancing siltation through water diversion is an important measure in ecological water resource dispatching.Taking the Doulong Port in the Lixiahe region of Jiangsu Province as the research subject,this paper redefined the water demand for scouring siltation based on the characteristic of scouring siltation in the area,so as to reflect the water demand for the net change in sediment monitored.By analyzing the in-situ monitoring data for the downstream channel of Doulong Port from 2001 to 2018,the paper extracted key variables such as initial riverbed volume,siltation amount,time duration,rainfall volume,and the frequency of sluice openings.A predictive model of water demand for scouring siltation was constructed,which combined adaptive particle swarm optimization(APSO)algorithm with support vector machine(SVM)and optimized the model parameters of the SVM through the APSO algorithm,enhancing the prediction accuracy of the APSO-SVM model.Furthermore,Sobol sensitivity analysis indicates that siltation amount,initial riverbed volume,and the frequency of sluice openings are the main variables affecting the water demand for scouring siltation,while rainfall volume and time duration show significant interactive effects.This study provides a theoretical basis and practical guidance for water resource management and channel maintenance,contributing to the improvement of water diversion efficiency in sea-entry channels and channel safety.关键词
引水冲淤/冲淤需水量/支持向量机/自适应粒子群优化/敏感性分析Key words
water diversion for scouring siltation/water demand for scouring siltation/support vector machine/adaptive particle swarm optimization/sensitivity analysis分类
建筑与水利引用本文复制引用
马朱桐,向龙,闫珂..基于APSO和SVM的滨海河网冲淤需水量预测模型研究[J].人民珠江,2024,45(12):114-121,8.基金项目
国家重点研发计划重点专项(2023YFC3207503) （2023YFC3207503）
