Learning Forum

[_antonis_]

Hello to everyone, 

I am running FSI for ascending thoracic aortic aortic aneurysm. The issue I am experiencing is the display of reverse flow in fluent until a large time step in 100% of the outlet areas. The simulation is running for 0.8 seconds. I use a time step size of 0.001 seconds. The analysis shows reverse flow until 0.13 seconds for supra aortic vessels and until 0.15 seconds for the outlet in the descending part. After that it runs normally. 
I'm using a hyperelastic model. When I'm running with stiffer models or even with elastic model I don't face this issue. When I'm running with a hyperelastic, however, it occurs.
First I would like to ask if this is a serious problem, and if so, how can I solve it?

Thank you,
Antonis

[Rob]

Reverse flow means you have flow entering through an outlet, or leaving through an inlet. It's not usually a good thing, but that's down to the specific model. In your case, does the artery wall expand too far relative to the inlet flowrate? 

[_antonis_]

Hello Rob,

Thank you for your reply.

I expanded the outlets and the inlet to a model so the flow can fully developed. To be more specific, I ran individual-only CFD, and the residuals were clear. I did not face this issue. Only happens in FSI.

 

[Rob]

Have you done a system coupling step, ie has the geometry deformed when you get the warning?

[_antonis_]

The warning is displayed from the beginning until 0.15 sec as I mentioned. But I exported graphs of mass flow rate in every time step size. The curves of backflow have a pick until 0.05 seconds and then they start going to negatives to satisfy continuity. Until that time of 0.05 sec, in mechanical there is no deformation.

The deformation starts from 0.05 seconds and after.  Does that mean something? 

[Rob]

Have a look at the flow field and geometry shape/volume too. Does the mass flow on the inlet size fill the volume change, or does the volume expansion rate exceed the ability of the inlet to fill the domain?

[_antonis_]

I am not sure if i understand fully, but i am thinking that you are talking about the flow field and if it covers all the volume. I am adding a picture of the velocity profile at the time of 0.15 Sec. The field is filled nice and the distribution of the velocity doesn't have any irregularities.

[Rob]

OK, bad wording on my part. Your system has a volume, and an inlet flow rate (volume). As the fluid volume expands if the inlet flow volume rate increases more than the fluid volume increase rate then no material will be sucked back into the fluid; otherwise the extra material will be sourced from the outlet. 

[_antonis_]

ok now i understand. But as you can see from the contour the velocity values decrease as the flow expands. Isn t this correct? I mean the velocity is always higher than the inside because the boundary is in the inlet, so it is increasing until the peak. 

i was wondering if the aorta, because of the hyperelastic material behaves at the bigging as a pump, duo to its deformation and then stabilizes. I don t know what else to think, i tried many set ups.

Please inform me if i answered corrrect. 

[Rob]

Possibly - monitor the inlet flow rate, zone volume and outflow rate over a cycle and see what the numbers tell you. You'll need Excell as you need the difference in value between time steps. Note, you can get a reverse flow warning but have a net outflow if only small parts of the outlet are affected. 

[_antonis_]

I ve monitored the flow rates. Here is a graph.

[Rob]

Now look at volume change in (surface report), fluid volume change (volume report) and fluid volume out (surface report). 

[_antonis_]

Ok. The values are presented below. 

 

[Rob]

Which implies the volume of the domain is increasing faster than the increase in flow volume from the inlet, so additonal volume is pulled into the fluid domain from the outlets. The solver is behaving as instructed, you need to figure out why you've not got enough flow to fill the expansion from the inlet. 

[_antonis_]

so the problem is on the set up of Fluent in FSI analysis? 
because as i mentioned in individual CFD there is no reverse flow

[Rob]

It means the supply (inflow) isn't sufficient to match the expansion calculated by the Mechanical FSI part of the simulation. The solvers are doing what they're told, so it's the boundary setting or structural material property that needs looking at. 

[_antonis_]

in mechanical the material properties, how can affect the flow? In the way i said before or i should see something else?
Also in Fluent, what is your suggestion to change? (parameters in dynamic mesh, solution methods etc)

 

[Rob]

I'm not suggesting changing anything: your model seems to be doing exactly what you've told it. 

If the fluid domain expands due to information from the Mechanical solver that extra volume needs to be filled with fluid. That fluid must enter the new larger volume from somewhere. If there isn't enough flow from the inlet where else can it come from? 

[_antonis_]

But the structural model takes the information of the fluid and expands it. In mechanical deformation and stresses are in a good values. There is no empty space between wall and fluid. The wall deforms according to fluid flow. 

[Rob]

Correct, and if the fluid expands more rapidly than the increased inflow can fill the space additional fluid must enter the domain from the outlet. Hence the reversed flow. 

[_antonis_]

Right. So the problem is in mechanical. I used experimetnal data and a hyperelastic model. I did not use prestress, is that maybe the reason?

[Rob]

Possibly, if the vessel is too small to start with you're going to need to fill it. Or run 5-6 cycles and see if the problem goes away. 

[_antonis_]

Hello Rob, 
A little update. Indeed it was the material properties of the structural model. The strains seemed to be in high values. 
Thank yo for your advice

One more general quastion, because you are the only experts i can contact with. How can i implement fiber reinforced model for the FSI of aorta?

[Rob]

Best to post a thread in the Mechanical section - you'll get a good response from someone in there.