A Non-linear Optimization Model to Minimize Flood Risks on Urban
Roadways Due to Storm-Drain System Deficiencies
Abstract
Urban flooding is caused due to poor drainage design, extreme weather,
and excessive rain. Such flooding severely affects the road
infrastructure. While there are a number of hydrologic software (e.g.,
TR-55, HydroCAD, TR-20, HEC-RAS, StreamStats, L-THIA, SWMM, WMOST, MAST,
HY-8) available to examine extent of urban flooding, the softwares
primarily require walking through a series of manual steps and address
each study area individually preventing a collective view of an urban
area in an efficient manner for hydrologic analysis. Furthermore, the
softwares have no ability to recommend optimal culver pipe sizes to
minimize flooding. In this paper, we develop a non-linear optimization
formulation to minimize urban flooding using underdrain pipe size as a
decision variable. We propose a solution algorithm in an integrated GIS
and Python environment. Monte Carlo Simulation is used to simulate
rainfall intensity by using empirical data on extreme weather from the
National Oceanic and Atmospheric Administration. An example using the
storm-drain system for the Baltimore County is performed. The results
show that the model is effective in identifying storm-drain deficiencies
and correcting them by choosing appropriate storm-drain inlet types to
minimize flooding. The proposed method eliminates the need to examine
each study area manually using existing hydrologic tools. Future works
may include expanding the methodology for large datasets. They may also
include a more sophisticated modeling approach for estimating rainfall
intensity based on extreme weather patterns.