Real time correction methods of hydrodynamic with H∞ filter

The correction methods of water-level measurements and discharge measurements are a important kind of hydrodynamic model real-time correction methods. The existing hydrodynamic model real-time correction methods are almost designed based on Kalman filter. In which, an implicit condition is that the system noises are obeys a Gaussian distribution with known mean and variance. However, in practice, the statistical properties of system noises are hardly obtained in a short time, especially for the river with few historical data. In this paper, the hydrodynamic model real-time correction methods is studied with the statistical properties of system noises are unknown. Firstly, the Saint-Venant Equations are linearized by the four-point linear implicit scheme. Secondly, two novel hydrodynamic model real-time correction methods with H∞ filter are proposed, based on the integrated state equation and the alternate state equations of water-level state and discharge state. Meanwhile, the validity and feasibility of the two novel hydrodynamic model real-time correction methods are verified the flood process of the Weihe river between Lintong and Huaxian during 9/26/1983~10/2/1983.