基于广义简约梯度算法的落差指数法参数优选  

Parameter optimization of fall index method based on generalized reduced gradient method

作　　者：周绍阳 王悦 龚朝海 ZHOU Shaoyang;WANG Yue;GONG Chaohai(Lower Changjiang River Bureau of Hydrological and Water Resources Survey,Bureau of Hydrology of Changjiang Water Resources Commission,Nanjing 210011,China)

机构地区：[1]长江水利委员会水文局长江下游水文水资源勘测局,江苏南京210011

出　　处：《水利水电快报》2023年第11期35-39,共5页Express Water Resources & Hydropower Information

摘　　要：传统落差指数法计算中存在多变量、非唯一解的问题,为此,提出了将落差指数法参数优选问题视为非线性约束优化问题。以长江大通水文站为例,利用广义简约梯度算法求解落差指数法中的参数,并将计算结果与传统方法进行了精度对比。结果表明:该算法准确度高且实用性强,实现了参数的全部优化并得到最优解,提高了流量整编精度,简化了计算步骤且不需要人工干预。该方法可为落差指数法的参数求解提供新思路。To solve the problem of multi-variable and non-unique solutions in traditional fall index method calculations,this study proposed that the parameter optimization problem of the fall index method can be solved as a nonlinear constrained optimization problem.Taking Datong Hydrological Station as an example,the generalized reduced gradient method was used to solve the parameters in the fall index method,and the accuracy of calculation results was compared with traditional methods.The results showed that the algorithm was correct and practical,which could achieve full optimization of parameters and obtain optimal solutions,improve the accuracy of discharge processing and simplify the calculation process without human intervention.The proposed method can provide a new approach for solving the parameters of the fall index method.

关 键 词：水位流量关系 广义简约梯度算法 落差指数法 单值化 大通水文站 

分 类 号：P332.4[天文地球—水文科学]