浙江大学学报（理学版）2011，Vol.38Issue(2)：234-238,5.DOI:10.3785/j.issn.1008-9497.2011.02.024
钱塘江河口盐度的神经网络模拟
Neural network modeling of salinity in Qiantang Estuary
摘要
Abstract
Saltwater intrusion may have serious impacts on drinking water sources in tidal estuaries. Reasonable predictions of the estuarine salinity may help to keep the drinking water sources safe and also to better controll the estuarine salinity by scheduling the discharge of reservoirs. A model is presented that based on the correlation between the non-linear time series of salinity and the forcing signals of freshwater input and tidal range. With the input data normalized, the model is trained with the first half data set, and then gives the hindcast of salinity of the second half. Then the output data is anti-normalized and the result matched well with the measured data, which indicates that the ANN model is capable of simulating the salinity change in tidal estuaries. The model is also applied to simulate the variation of salinity responding to different discharges. The result shows that descending flow in the process of discharging water in reservoirs can better inhibit the invasion of salt water and reduce the harmful influence on drinking water sources in tidal estuaries more efficiently.关键词
盐度/神经网络/SCG算法/钱塘江河口分类
海洋科学引用本文复制引用
许丹,孙志林,潘德炉..钱塘江河口盐度的神经网络模拟[J].浙江大学学报（理学版）,2011,38(2):234-238,5.基金项目
国家科技重大项目(2009ZX07424-001) （2009ZX07424-001）
国家自然科学基金资助项目(40776007). （40776007）
