新安江模型参数线性化率定研究

Research on the Linearizing Method of Parameter Calibration for Xin anjiang Model

包为民教授提出了非线性函数参数的线性化率定方法,解决了以误差平方和为目标函数求解非线性函数参数增加不相关的局部优值问题。为此,研究将参数线性化率定方法应用于新安江模型,并通过理想流域和实际流域验证了该方法的合理性和高效性。选取湖北乌溪沟流域43场洪水对次洪8个敏感参数同时进行线性化率定,各参数取不同的初始值均可以较快地收敛到稳定的最优值。31场洪水参与模拟,全部合格;12场洪水参与检验,全部合格。次洪模拟的合格率达到100%,确定性系数为0.923,预报等级为甲级。

A linearizing calibration method of non-linear function parameters proposed by Professor Bao Weimin can solve the non-related local optimums problems of traditional non-linear function parameter calibration method based on minimum error square sum objective function.To this end,the application of proposed linearizing calibration method in the parameter calibration of Xin anjiang Model is studied,and the rationality and validity of this method is verified through the comparison of ideal basin and actual basin.43 floods in Wuxigou Basin in Hubei Province are selected to linearize the eight sensitive parameters of hourly model by the linearizing calibration method at the same time,and the calibration results show that every parameter can quickly converge to a stable optimal value by taking different initial values.31 floods participated in the simulation and all qualified,and 12 floods participated in the inspection and all qualified.The qualified rate of hourly model simulation reaches 100%,the uncertainty coefficient is 0.923 and the forecasting grade is Class A.