I'm having trouble getting a 1D/2D model to run. The river meanders and has an expansive floodplain with skewed embankments so I chose to model it in 2D with a short 1D reach at a bridge location (I have not yet entered bridge geometries in the 1D area). I am trying to model the overflow bridges as culverts in a 2D Area Connection since they have simple geometries and so they don't have to be "wet" during entire model runs like 1D XSs.
The model runs until flows reach out-of-bank areas near the 2D Area Connection. I was however able to run the entire event with no culverts (overflow bridges) in the 2D Area Connection. It was not until I added culverts to the 2D Area Connection that the model became unstable. I made sure the 1D/2D Connection matches the terrain and the culvert inverts are at or above the connecting cell inverts.
Things I've tried:
-1 second to 2 minute time steps with and without time slicing
-Using the max amount of points/curves in the HTab along with a max headwater elevation
-Weir flow decay exponent and flow stability to 3
-Theta is at 1
-Full momentum and diffusion wave equations
Below are some screen shots of the model. Does anyone have any further ideas?
Try making the cells that connect to the culverts as large as possible. Your culverts look large so they can probably convey more volume in one time step than what the cell can provide at the initial time flow begins to reach the culverts. This causes continuity difficulties in converging to a solution. A "cheat" I sometimes have to do for cases like this is to alter the terrain at the head and tail water to create more volume in the cells. This is usually on the order of tenths of a acft or less so the artificial added volume doesn't have a impact on the peak elevation but helps when you have flow just starting to enter the culvert. Another thing you might try is to set a higher n-value in the bottom say 6" to foot of the culvert to reduce the initial capacity of the culverts.
Thanks for the help and for the great explanation! I made the 2D cells connecting to the culverts larger (using one cell on each of the head and tail water ends). I also created more water storage around these culverts and gave the culvert bottoms a higher mannings n value. This has drastically reduced the flow errors.
Have you tried this using Full 2D? The way you are currently modeling the 1D bridge is just way more simple for the software to analyze in 2D. If you dont know how i can explain based on response, if you do then give it a try. I also see that your 2D areas are connected via a lateral structure? Can i ask what you set your weir coefficient to? What was the mass balance error %?