2D modeling in urban areas: A case study comparing side by side PCSWMM 2D and HEC-RAS 2D

Simon Deslauriers and Gilles Rivard,


Two-dimensional (2D) hydraulic models are now frequently used by engineers to analyze more complex flow configurations in urban areas where 1D models are deemed inadequate or insufficient to represent the flow paths and resulting depths or velocities. With the growing computational capacity, the popularity of GIS based data and the availability of digital information on elevation, 2D modelling now seems to be a more accessible solution to assess overland flood problems. The number of numerical models which are now capable of undergoing such task is also growing and the modelers can now choose different models that have different capabilities. However, it is not always easy to evaluate the various approaches used in the different software packages which could produce variable results with the same data set.

In that context, a comparison was completed using three different approaches and software to evaluate and compare the results with an actual case of documented urban flooding. Two 2D models based on PCSWMM and a 2D model based on HEC-RAS were set up based on real project data and their results from a flood event were compared. PCSWMM allows for a dynamic interaction between the urban network (pipes, culverts and channel) and the 2D flooded area. PCSWMM also provides complete 2D approach to represent the channel network and flooded area. The two PCSWMM models are built with these two different methodologies. One uses a 1D-2D approach (quasi-2D mesh connected to a 1D conveyance network) and the other uses a complete quasi-2D methodology to represent both the creek and the flooded area.

The third model was built within HEC-RAS with the 2D routing capabilities. HEC-RAS can perform 2D unsteady flow modeling through a structured mesh of computational cells with detailed hydraulics table properties for each cell faces.

For the 3 different 2D modeling approaches, flow inputs used for the 2D models where obtained by a 1D PCSWWM models calibrated on estimated flood extend of the event.

After a description of the 3 approaches, a real flood event is used to compare the results obtained with these different models. The methodologies are also compared to highlight the advantages and disadvantages of each model or method in terms of model construction, computational time and results as compared with actual observations for the flooding event.


Permanent link: