Combined 1D/2D - Flow errors

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Combined 1D/2D - Flow errors

SimR
I'm building a combined 1D/2D model using lateral structures to link the 1D channel to the 2D floodplains.
The model covers over 50 km of river and all lateral structures have a length of 1-2 kilometers. There are no gaps between them. The flow values used in the model are constant over the simulation period.
The model runs without any computation errors and the water profiles don't seem to show weird behaviors.

However I'm seeing values that I cannot explain when I look at the flow values in the output tables.
The model stabilized with a flow value in my downstream cross section that is about 170% what I expected and the flow values going from upstream to downstream seem to oscillate instead of going steadily up. The flow transiting in the 2D areas don't seem to explain this behavior.

When I enable iterations between 1D and 2D in the 1D/2D options, a lot of flow errors are popping up in the computation window. I'm guessing the flow errors are due to the lateral structures linking 1D and 2D but I'm not sure how to fix them. The lateral structures are defined using the weir equation (width 2m, weir coefficient 0.2).

I'm guessing the extra flow is created over the length of my model at every lateral structure but I'm not sure how to go about fixing this. I tried modifying the weir coefficients, computation interval and tolerances but not much luck so far.

I'm trying to avoid switching to the normal 2D Equation Domain for the lateral structures as I would have to rework the entire model to create a gap between each LS (more than 50 of them) to prevent 2D cells from spanning two different lateral structures but would it be a possible solution?

Reply | Threaded
Open this post in threaded view
|

Re: Combined 1D/2D - Flow errors

jarvus
You can try smaller and smaller time steps and see if that makes any improvements.

It is hard to know if the using the 2D equation domain would solve your problems or not.  I have some data sets where it definitely worked better.
Reply | Threaded
Open this post in threaded view
|

Re: Combined 1D/2D - Flow errors

cameron
You can try minimizing the number of points in the weir for the lateral structures to see if that helps, but you definitely need to have the 1D/2D iterations turned on. You may need to go smaller on the weir coefficient than 0.2, 0.05 might be appropriate depending on if it is natural ground or not. Timestep may also help.  

As for fixing the model for the 2D equation, you would just need to add breaklines at all splits between lateral structures and it will fix your problem.  

I would identify which lateral structures are having the 1D/2D iterations and see if you need to split them up or adjust them.
Reply | Threaded
Open this post in threaded view
|

Re: Combined 1D/2D - Flow errors

SimR
This post was updated on .
Thank you both for the answers!

Reducing the timestep from 5 to 1 second and adding 1D/2D iterations does reduce the flow errors significantly.

Some lateral structures are reaching the maximum numbers of iterations and showing flow errors for the first hour of the simulation but they are relatively small and as I'm applying a constant flow it shouldn't be an issue for the last time steps when the model is stabilized.

My boundary from 1D to 2D is mainly natural ground so should I favor 2D equations as a baseline?
I will try with lower weir coefficients and see how it works out.

Good point on the break lines! Still have a lot to learn.