[Vishnu_bharathi]

I am doing transient heat transfer simulation, where I have Fluid domain(high temp fluid is given at inlet) and solid domain(as insulation). Fluid pass through porous domain which has insulation(solid domain) over it. The fluid tries to increase the temperature of porous domain which is insulated.

I have attached residual monitor and imbalance that I observed. The residual shoots up over 350 Iterations approximately. Initially I had used outlet bdary condition but since I received 'A wall has been placed at the outlet and try using opening boundary condition', I changed the outlet to opening and additionally I moved the outlet distance by a factor of 3 compared with previous simulation run. Regarding the mesh quality - It is very good since I am using simple geometry with Multizone option. I did not receive any other warnings or messages while solver completed its run. I have attached images of the velocity vectors near outlet region. I use SST model, Total energy option(fluid is compressible). What I observed from the velocity vector contour is, when the temperature (from ambient) starts increasing near outlet the iteration number is about 350 and the residuals starts to osciallte or shoot up as seen from the image.

What might be the problem here? If anyother information is missing here I would like to add to have better understanding.

[ghorrocks]

The flow at the outlet appears to be flapping about. This is likely to be the cause of the "A wall has been placed..." and the poor convergence.

Is this flapping real? You can get flows like this, so it is possible it is real. But your outlet boundary is certainly too close to the action, it will need to be further downstream. That would be the first thing I would try.

[Vishnu_bharathi]

Originally Posted by ghorrocks

The flow at the outlet appears to be flapping about. This is likely to be the cause of the "A wall has been placed..." and the poor convergence.

Is this flapping real? You can get flows like this, so it is possible it is real. But your outlet boundary is certainly too close to the action, it will need to be further downstream. That would be the first thing I would try.

Thank you for the response

1. About the flapping, I do not know. I am really interested in the temperature distribution in the porous region(with its circumference insulated by solid domain) caused by the hot fluid passing through. So, I have started a simulation by moving the oulet by a factor 3.5 from the previous run.

Apart from extending the outlet far away, I am using 'opening' condition with 0 Pa relative pressure, with constant ambient temperature as 25°C at outlet boundary because of the % of wall has been placed message. For my case, I see the pressure just after the porous domain is very low(less than the ambient) and being set 'opening condition' at outlet, I observe inflows to certain extent.(outside its ambient so obviously this inflow make sense). How shall I view this?Extending the outlet further away from porous end, Will it have an effect on the message "A wall has been placed"?

2. I am using Interruption criteria to stop my simulation(to perform Charging and discharging cycles):

i) for this, in my previous run I used 'massFlowavg(Temperature)@outlet' to interrupt charging cycle simulation and also defined the monitor point using the same expression as above. When I observed the monitor points in the solver run, I see it osciallates like saw tooth. Why is that? Is it because of the backflow and irregular fluid flow near or at the monitor surface(ie., outlet)?

3. I made blocks of the porous domain in Design modeller using share topology. I did this in order to create monitor surfaces at regular interval from start of porous domain. I used expression 'massFlowavg(T)@porous start face,intermidate face 1, intermediate face2, intermediate face3,....,porous end face' to monitor average temperature. While monitoring I observed temperature at porous start face and at end face was correct as expected but temperatures at inter.face1,inter.face2,....inter.facen to be varying exponentally from positive to negative. What is wrong here? (porous start face , porous end face are the ones where porous medium has interface with fluid domain and other intermedia faces are created within porous domain for monitor surfaces)

Ultimately, I want to create an Interrupt condition where avg temp. will be used at specified distance from the domain start face.

P.S: Ref attached pic for question 2 & 3

[ghorrocks]

I would stop work on the charging and discharging cycles for now. You appear to have some fundamental problems with the simulation and I would sort them out before adding the additional complexity of a duty cycle.

The flapping is fundamental. You need to work out whether this is real (and then you need to model it properly) or false (and then you need to change the numerics to stop it).

Also the massive temperature results show something is wrong there too.

Some suggestions:
* Extent the outlet boundary further, much further.
* You appear to be having similar flapping on the inlet as well, so try extending the inlet as well.
* Does experimental results suggest this flapping is real?
* Are you sure your boundary conditions are appropriate?

[Vishnu_bharathi]

Hi Ghorrocks,

Thanks for the suggestion. I will look into it and see how it spans out.

Do you have any suggestion for why "massFlowAvg(T)@plane" does not work when used to find avg temperature at intermediate surfaces created from DM. It works for when I use "massFlowAvg" expression at interface surfaces but not in between faces in a domain.(eg., it gives +/- e+17 values for Temperature)
(Please check No. 3 above)

[ghorrocks]

No, I cannot explain the out of control temperature. Check the results file to see whether the temperature really is those sort of values in that region.

[Vishnu_bharathi]

Originally Posted by ghorrocks

The flow at the outlet appears to be flapping about. This is likely to be the cause of the "A wall has been placed..." and the poor convergence.

Is this flapping real? You can get flows like this, so it is possible it is real. But your outlet boundary is certainly too close to the action, it will need to be further downstream. That would be the first thing I would try.

I tried with extending the outlet further downstream. by approx 8 times the initial length. But unfortunately it did not change anything.I tried few more other basic things which I say it new thread.