目前感潮河段水位预报模型基于当前站水位、上游径流水位和下游海洋潮位数据进行预报,导致预见期受上游来流时间制约、上下游站数据缺失情况下无法进行预报、数据采集成本高等问题.文章假设感潮河段水位由径流因素和潮汐因素共同线性作用,基于单站等时距水位数据,建立了包含一个径流动力修正系数和若干个潮汐调和分潮系数的非线性优化预报模型,并结合曲线拟合最小二乘法、潮汐调和分析法、梯度下降法,给出一个寻优的迭代算法,对模型参数进行优化辨识.模型应用于长江感潮河段部分观测站,水位预报精度较好,具有可行性和实用价值.%The current water level forecast models of tidal reach are subject to the observation data of water lev-el from the current tidal reach station, the runoff stations upstream and the tide stations downstream. It leads to the problems as follows:the leading time is subject to the time of the flow from the upstream;if the data at the upstream or the downstream are missed, the forecast would not be able to proceed;the acquisition of the data is high in cost. Assuming that the water level of tidal reach was linearly controlled by the runoff and tide, the paper built a nonlin-ear optimal forecast model with respect to one dynamic correction coefficient of runoff and a set of tidal harmonic co-efficients, which was based on the water level at time knots with constant time interval from a single station. Then, combining the curve fitting least square method, the harmonic analysis method and the gradient descent method, an iterative algorithm was proposed for solving the forecast model. The model was applied to some stations of the tidal reach of Yangtze River for test. Numerical experiences show the feasible and applicable of the model.
展开▼
