User Forum of Software BASEMENT

dazzle

User

Registered: 2015-05-07

Posts: 12

Culvert/HQ-Relation unrealistic flow

Hi,

it is the first time for me setting up a BASEMENT-calculation with a culvert. I have the culvert implemented as a Inner Boundary/HQ-Relation. The related HQ-Table consists of water surface elevation (first column) and discharge (second column). It is ~2m wide culvert/string for up- and downstream strings, Q is limited in the table to 17.2 m³/s. So flow velocities should be below 10m/s for sure, but at the downstream culvert boundary BaseVIZ shows velocities of over 50 or 100m/s, highly fluctuating. Has anyone noticed a similar behaviour and maybe has a solution? I am kind of stuck

Best
David

Lukas Vonwiller

Developer

Registered: 2014-09-04

Posts: 141

Re: Culvert/HQ-Relation unrealistic flow

Hi David

I guess this is related to the downstream string. Try to treat this like an inflow boundary. My guess is that you should avoid high bed slopes in your mesh right after the downstream string. Try to change the topography and mesh accordingly. Furthermore you could try to use initial conditions to make sure that the cells below the downstream sting are wet.

I hope this helps!
Best, Lukas

dazzle

User

Registered: 2015-05-07

Posts: 12

Re: Culvert/HQ-Relation unrealistic flow

Hi Lukas,

I improved mesh quality regarding emelent angles and slopes at both strings with no success. See screnshot below with the mesh on the left (flow direction is left to right) and BaseVIZ to the right indicating extremely high velocities at the outlet.

Elevation at the nodes at the upstream stringdef = 513.2
Elevation at the nodes at the downstream stringdef = 513.08

HQ-Table:
513.00  0.00
513.25  0.1
513.75  3.54
514.25  5.40
514.75  6.85
515.25  8.17
515.75  9.38
516.25  10.43
516.75  11.37
517.25  12.27
517.75  13.16
518.25  14.01
518.75  14.81
519.25  15.58
519.75  16.33
520.00  16.7
520.45  17.2
... and extends in 1m increments up to elevation of 530 with Q fixed to 17.2

I am thankful for any other clue...
Best David

Last edited by dazzle (2025-05-23 06:23:30)

Lukas Vonwiller

Developer

Registered: 2014-09-04

Posts: 141

Re: Culvert/HQ-Relation unrealistic flow

Hi David

Just some thoughts: Have you tried with a purely triangular mesh? Does maybe a longer downstream string improve the velocities? Maybe try a finer mesh in general and especially around the downstream string (more than 4 cells)? Last but not least, try BASEHPC, where the inflow/outflow is averaged over the whole stringdef and hence the upstream and downstream stringdef can consist of a different number of nodes.

Hope this helps!
Br, Lukas

dazzle

User

Registered: 2015-05-07

Posts: 12

Re: Culvert/HQ-Relation unrealistic flow

I have successfully used meshes that partially consist of quadrilateral elements with BASEMENT before.
This is the first time I am implementing a culvert and obviously having troubles with it.
Is there a rule-of-thumb regarding the mesh refinement (element size, relationship Q to required nodes/width of stringdef) for inflow- and outflow boundaries?

Best
David

Lukas Vonwiller

Developer

Registered: 2014-09-04

Posts: 141

Re: Culvert/HQ-Relation unrealistic flow

Is there a rule-of-thumb regarding the mesh refinement (element size, relationship Q to required nodes/width of stringdef) for inflow- and outflow boundaries?

There is not really a rule for this. But if I had to suggest something, I would say at least 5 to 10 cells over the width of your inflow/outflow boundary. Of course this depends, among other things, on the width and lateral variation of your inflow/outflow cross section.