论文部分内容阅读
应用系统响应函数识别方法推求谢尔曼单位线,是水文学家感兴趣的课题。但是,应用一般的无约束识别方法会导致单位线纵座标的正负振荡;二约束条件(水量平衡及纵坐标非负)下的系统识别方法,在理论上和实践中仍可能出现在正区间的振荡。因此,建议在三约束条件《增加无振荡约束)下,进行系统识别,算法则采用在理论上证明是收敛的A·V·Fiacco的松弛无约束极小化法。经在20多个流域150多条单位线应用,证明上述方法稳定性好,基本上不出现振荡。并且克服了传统水文学对于长历时降水,连续洪水难于分析单位线的障碍。
It is a topic of interest for hydrologists to apply the system response function identification method to infer the Sherman unit line. However, the application of general unconstrained identification method leads to the positive and negative oscillation of the ordinate of the unit line. The system identification method under the two constraints (water balance and non-negative ordinate) may still appear in the positive interval in theory and in practice Oscillation. Therefore, it is suggested that the system be identified under the condition of three constraints: “increase no oscillation constraint”, and the algorithm uses the theory of loosely bound unconstrained minimization of AV Fiacco, which is theoretically proved to be convergent. The application of more than 150 unit lines in more than 20 basins proves that the above method has good stability and basically does not oscillate. It also overcomes the obstacles of traditional hydrology to long-term precipitation and difficult continuous analysis of unit hydrographs.