实时洪水抗差预报系统研究

Application of Robust Estimation in Real Time Flood Forecasting System

把抗差估计理论引入实时洪水预报系统,提出适合于实时洪水预报系统的抗差特征函数,建立降雨误差的动态监控模式,研究实时修正模型参数的动态抗差估计,结合洪水预报模型,形成具有水文学特点、抗差性强的实时洪水抗差预报系统。将此系统和传统实时洪水预报系统运用于闽江七里街流域,比较当降雨和流量资料存在异常误差时两系统的预报效果。结果表明,实时洪水抗差预报系统能抵御异常值的影响,获得比较稳定且高精度的预报结果。

Robust estimation theory was introduced to real-time flood forecasting system, and the theories and methods that were not adaptable to flood forecasting have been modified and perfected. This paper proposed robust characteristic function for real-time flood forecasting system, and established the dynamitic monitor module for rainfall error, and researched dynamitic robust estimation for AR model parameters. A real-time flood robust forecasting system with hydrologic characteristics and good robust ability has been formed.