河流预报模型参数多目标优化算法研究在沁水河流域应用

Research on the Multi-objective Optimization Algorithm for River Forecast Model Parameters in the Qinshui River Basin

水文预报的合理校准主要与模型及参数有关,为了提供更准确的模拟,在垂向混合产流模型的基础上结合水利工程影响,对流域降雨径流过程进行模拟。同时在多目标控制下覆盖整个权衡范围,并在其基础上应用MOPSO及MOSCEM-UA算法生成的Pareto前沿解对模型参数进行合理估计。经案例分析,对比两种算法模拟结果,MOSCEM算法在模拟过程中百分比偏差Bais基本在理想范围10%内且残差较小,相关关系78%在0.90以上。表明该算法在多目标模式下效率更高,为研究区水文预报模型多目标优化提供更准确的模拟。

The reasonable calibration of hydrological forecast is mainly related to the model and parameters.In order to provide more accurate simulation,the rainfall runoff in the basin is simulated based on the vertical mixed runoff model combined with the influence of hydraulic engineering.At the same time,under the multi-objective control,the whole trade-off range is covered,and the Pareto front-end solution generated by MOPSO and MOSCEM-UA algorithm is used to estimate the model parameters reasonably.Through a case analysis,the simulation results of the two algorithms are compared,the percentage deviation of the MOSCEM algorithm in the simulation is basically within 10% of the ideal range and the residual is small,and the correlation is over 78% above 0.90.It shows that the algorithm is more efficient in multi-objective mode and provides more accurate simulation for multi-objective optimization of hydrological forecasting model in the study area.