No warnings, huge percent error

[jimardon]

I ran a 1D/2D model with no warnings in the runtime message. But the percent error is huge (27.66 for 1D and 35.46 for 2D). Any insight on what I could be doing wrong?

Basically, the model has a flow hydrograph BC in the upstream XS, a uniform lateral inflow on another XS, and downstream normal depth. The 1D – 2D connection is done through lateral structures.

What bothers me the most is that I don’t get any warnings after computing.

[cameron]

What is your time step? With the lateral structures, is it set to weir or 2D? Are you using the full momentum equation or the diffusion wave for the 2D solver?

[jimardon]

– 5 sec time step
– lateral structures set to weir (0.5 Cd)
– Diffusion wave

I don’t think any of these setting will help improve the percent error though.

Update: Decreasing the number of XSs significantly reduces the error, but still high (~19%).

[cameron]

Try changing to the full momentum.

How many lateral structures do you have? Do any cross near bridges in the 1D?

Try switching your lateral weir to solve with the 2D solver.

How many cross-sections do you have?

[jimardon]

Solved it. Issue was that lateral structures connecting to 2D tributaries need a higher Cd for the weir.

If anyone else runs on this problem: Use ~0.5 Cd for lateral structures connecting to floodplains, and between 2-5 when connecting to tributaries (calibrate accordingly to minimize percent error in mass balance).

Hope this helps.

Thanks!

[Author]

Posts