基于图形处理器的涪江中段流域降雨汇流模拟方法研究

Simulation of Rainfall Runoff in the Middle Reaches of Fujiang River based on GPU

为增加径流汇流模型的时效性,提出一种基于图形处理器通用计算的汇流模型。首先,在稳定水源供给的条件下,通过对模型修改和完善生成流域前期水环境。最后,将小时降水数据叠加到前期水环境进行径流汇流模拟得到模拟结果。通过验证发现,模型模拟结果具有更小的水位变化误差,模拟精度进一步提高,模拟速度更快,满足实时计算要求。精度和性能的同时提升,证实汇流模型在流域暴雨洪涝灾害实时评估中具有重要的应用价值。

In order to increase the timeliness of runoff confluence model, a confluence model based on general computing of graphics processor is proposed in this paper. Firstly, under the condition of stabilizing the water supply, the water environment in the early stage of basin formation is improved by modifying the model. Finally, the hourly precipitation data are superimposed on the previous water environment to simulate the runoff confluence and get the simulation result. Through verification, it is found that the simulation result of the model have smaller water level variation error, the simulation accuracy is further improved and the simulation speed is faster. All of those meet the requirements of real-time calculation. The accuracy and performance are improved at the same time, which proves that the confluence model has important application value in real-time assessment of rainstorm and flood disasters in river basins.