卡尔曼滤波校正技术在水动力学模型实时洪水预报中的应用

Application of Kalman filter technique in real-time flood forecasting

将卡尔曼滤波技术与水动力学模型实时洪水预报相结合,提出了"交替校正"方法;分别建立水位与流量的状态方程,进行交替滤波计算,较好地解决了滤波计算中不同类型状态量之间共同校正的问题.对长江干流局部河段的实例验证表明,卡尔曼滤波技术应用于单一河道水动力学模型没有破坏模型的稳定性,局部的校正对全河道具有较好的带动作用,说明卡尔曼滤波技术应用到水动力学模型的实时校正中是可行而且有效的.

Based on the Kalman filter, a technique applied in the real-time updating of flood forecasting, an alternate updating method was developed. Two Kalman filter equations of state, one for water level and one for discharge, were established and calculated alternately, and the different state variables in the filter calculation were updated jointly. A case study in part of the main stream of the Yangtze River indicates that the proposed method would help maintain the stability of hydrodynamic models for single rivers through local updating, and it is practical and effective in improving the accuracy of real-time forecasting.