摘要Urban road networks function as surface passage for floodwater transport during extreme storm events to reduce potential risks in the city.However,precise estimation of these flow rates presents a significant challenge.This difficulty primarily stems from the intricate three-dimensional flow fields at road intersections,which the traditional one-dimensional models,such as Storm Water Management Model(SWMM),fail to precisely capture.The two-dimensional and three-dimensional hydraulic models are overly complex and computationally intensive and thus not particularly efficient.This study addresses these issues by integrating a semiempirical flow diversion formula into the SWMM source code.The semiempirical formula,derived from hydraulic experiments and computational fluid dynamics simulations,captures the flow dynamics at T-shaped intersections.The modified SWMM's performance was evaluated against experimental data,and the original SWMM,the two-dimensional MIKE21,and the three-dimensional FLUENT models.The results indicate that the modified SWMM matches the precision of the two-dimensional MIKE21,while significantly reducing computational time.Compared to MIKE21,this study achieved a Nash-Sutcliffe efficiency of 0.9729 and a root mean square error of 0.042,with computational time reduced by 99％.The modified SWMM is suitable for real-sized urban road networks.It provides a high-precision tool for urban road drainage system computation that is crucial for effective stormwater management.