Flow through a City Building Layout
This benchmark consists in a single-long period wave propagating up a piecewise linear slope and onto a small-scale model of the town of Seaside, Oregon. For numerically generating this wave we chose the second option of the two proposed approaches: Imposing the time series of incident wave elevation at x=5 cm to force the model at x=5 cm. Several model configurations have been considered. Two numerical resolutions were set: 1cm and 2cm. Three numerical schemes used: first order, second order and WAF. The numerical results presented here are for the second order scheme. Several values for the friction parameter were considered, ranging from 0.01 to 0.035. We also questioned ourselves about the spatial variability of the simulated variables slightly moving the sampling location in the y direction (by 4 cm).
The first remark concerns the simulated signal at the control point compared with the measured one. The exact arrival time is not well captured nor the train of waves between t=35 and 40 sec. This fit cannot be arranged by simply changing the shape for the initial condition, but need of the used of a dispersive model. For the initial phase of the propagation of the generated wave dispersion is important. Nevertheless, we observed that, in this case, including dispersion in the model is not mandatory in order to get good agreement with measured data. In the numerical results we found out that for location B1 flow depth and momentum flux where very sensitivity to small variations in the y-direction of the position of the sampling point, getting a much better fit for points moved slightly “downwards”. This analysis of the spatial variability of the simulated variables was performed motivated by the fact that, while the flow depth at locations B4, B6 and B9 fit well, it was clearly underestimated for the first location, B1. We checked out that slightly moving the sampling point downwards resulted in a much better fit with measured data (not shown, but can be found in the presentation slides). In Figure 6, the comparison between measured data (flow depth, velocity and momentum flux) and numerical model for different values of the friction parameter at locations B1 and B4 is shown. Figure 7 depicts the same comparison for locations B6 and B9. At location B9, for simulations performed with low values for the friction (around 0.012) reflected waves can be observed both in flow depth and velocity direction.
Figure 6. Comparison of measured flow depth, cross-shore velocity and cross-shore momentum flux with numerical results for varying friction coefficients for B1 and B4 locations.
Figure 7. Comparison of measured flow depth, cross-shore velocity and cross-shore momentum flux with numerical results for varying friction coefficients for B6 and B9 locations.
