[boffin5]

Piston engine powered aircraft have a coolant radiator, and the airflow through the radiator gets heated as it passes through. With the energy added from the heat, the exhaust flow exits faster, and gives a speed increase to the airplane. This is called the Meredith Effect, and the P-51 Mustang of WW2 is known for this. But the validity of the effect has been controversial.


So I have undertaken a CFD study to see if I can show this effect. I have half of an airplane fuselage, and a radiator in a ventral fairing with inlet and exhaust ducts. It is a chtMultiRegion case, with the regions being fluid (the atmosphere) and solid (a porous zone comprising the radiator). Attached are images showing the geometry, and some zip files.


Now the problem. It won't run. No matter how I adjust the BC files, I sill get this error message in the output:

Region: fluid Courant Number mean: 6052.08 max: 56515.2
Region: solid Courant Number mean: 0.396462 max: 1.77651
Time = 2


Solving for fluid region fluid
DILUPBiCGStab:  Solving for Ux, Initial residual = 0.107372, Final residual = 4.73594e-05, No Iterations 1
DILUPBiCGStab:  Solving for Uy, Initial residual = 0.058436, Final residual = 0.000132783, No Iterations 1
DILUPBiCGStab:  Solving for Uz, Initial residual = 0.0509489, Final residual = 0.000297498, No Iterations 1
DILUPBiCGStab:  Solving for h, Initial residual = 0.999996, Final residual = 0.00156465, No Iterations 1

    From function Foam::scalar Foam::species::thermo<Thermo, Type>::T(Foam::scalar, Foam::scalar, Foam::scalar, Foam::scalar (Foam::species::thermo<Thermo, Type>::*)(Foam::scalar, Foam::scalar) const, Foam::scalar (Foam::species::thermo<Thermo, Type>::*)(Foam::scalar, Foam::scalar) const, Foam::scalar (Foam::species::thermo<Thermo, Type>::*)(Foam::scalar) const, bool) const [with Thermo = Foam::hPolynomialThermo<Foam::icoPolynomial<Foam::specie> >; Type = Foam::sensibleEnthalpy; Foam::scalar = double; Foam::species::thermo<Thermo, Type> = Foam::species::thermo<Foam::hPolynomialThermo<Foam::icoPolynomial<Foam::specie> >, Foam::sensibleEnthalpy>]
    in file /home/ubuntu/OpenFOAM/OpenFOAM-8/src/thermophysicalModels/specie/lnInclude/thermoI.H at line 70
    Energy -> temperature conversion failed to converge:
          [0] iter          Test           e/h          Cv/p          Tnew
          [0] 0       486.244        193799       1037.72       2539.29
          [1] iter          Test           e/h          Cv/p          Tnew
          [1] 0        551.55        261164       1049.09      -58898.4
          [1] 1      -58898.4    9.2188e+13  -7.75261e+09      -47007.2
          [1] 2      -47007.2   3.02078e+13   -3.1755e+09      -37494.4
          [1] 3      -37494.4   9.89833e+12   -1.3007e+09      -29884.3
          [1] 4      -29884.3   3.24338e+12  -5.32777e+08      -23796.5
          [1] 5      -23796.5   1.06273e+12  -2.18229e+08      -18926.5
          [1] 6      -18926.5   3.48188e+11  -8.93869e+07      -15030.5
          [1] 7      -15030.5   1.14058e+11  -3.66101e+07      -11913.3
          [1] 8      -11913.3   3.73427e+10  -1.49899e+07      -9417.98
          [1] 9      -9417.98   1.22074e+10  -6.13161e+06      -7416.96
          [1] 10      -7416.96   3.97245e+09  -2.50029e+06      -5803.33
          [1] 11      -5803.33   1.27494e+09  -1.00948e+06      -4478.83
          [1] 12      -4478.83   3.91858e+08       -394822      -3329.03
          [1] 13      -3329.03    1.0402e+08       -138572      -2130.17
          [1] 14      -2130.17   1.42783e+07      -30478.3        376.06
          [1] 15        376.06       79481.1       1021.98      -60473.3
          [1] 16      -60473.3   1.05062e+14  -8.60722e+09      -48267.1
          [1] 17      -48267.1   3.44262e+13  -3.52555e+09      -38502.3
          [1] 18      -38502.3   1.12806e+13  -1.44408e+09      -30690.6
          [1] 19      -30690.6   3.69632e+12  -5.91506e+08      -24441.5
          [1] 20      -24441.5   1.21114e+12  -2.42286e+08      -19442.4
          [1] 21      -19442.4   3.96819e+11  -9.92408e+07      -15443.3
          [1] 22      -15443.3   1.29992e+11  -4.06465e+07      -12243.6
          [1] 23      -12243.6   4.25637e+10  -1.66436e+07      -9682.53
          [1] 24      -9682.53   1.39179e+10  -6.80927e+06      -7629.44
          [1] 25      -7629.44   4.53285e+09  -2.77822e+06      -5975.52
          [1] 26      -5975.52   1.45848e+09  -1.12375e+06      -4622.39
          [1] 27      -4622.39   4.51887e+08       -442130      -3459.85
          [1] 28      -3459.85   1.23428e+08       -158475      -2289.09
          [1] 29      -2289.09   1.97577e+07      -38718.2      -174.712
          [1] 30      -174.712       -451965       843.014      -73311.6
          [1] 31      -73311.6   2.72905e+14  -1.84719e+10      -58537.6
          [1] 32      -58537.6   8.94248e+13  -7.56615e+09      -46718.6
          [1] 33      -46718.6   2.93024e+13  -3.09912e+09      -37263.5
          [1] 34      -37263.5   9.60164e+12  -1.26942e+09      -29699.6
          [1] 35      -29699.6   3.14616e+12  -5.19963e+08      -23648.8
          [1] 36      -23648.8   1.03087e+12  -2.12981e+08      -18808.3
          [1] 37      -18808.3    3.3775e+11   -8.7237e+07      -14935.9
          [1] 38      -14935.9   1.10637e+11  -3.57294e+07      -11837.6
          [1] 39      -11837.6    3.6222e+10  -1.46291e+07      -9357.37
          [1] 40      -9357.37   1.18402e+10  -5.98375e+06      -7368.26
          [1] 41      -7368.26   3.85217e+09  -2.43964e+06      -5763.81
          [1] 42      -5763.81   1.23555e+09       -984533      -4445.77

In past posts regarding this problem, one solution was to increase the fineness of the mesh, but I already have made it quite fine in the important areas. I think I have some kind of problem in the boundary conditions, but have been unable to tweak them for the case to run.



I tried to attach a zip file of the entire case, but at 381 kb, it's too big. The constant folder itself is 274 kb, so that's too big as well. So I just attaching the '0' and system folders, in hopes that someone will find the problem. Or is there a way to further compress the file sizes?



Question: I have split the radiator into patches, so that I can tailor the inlet and outlet patches in the BCs. In the 0/solid BCs, are the radiator inlet and outlet patches to be treated in the same way as 'inlet' and 'outlet' in the domain (fluid) patches?



In my BCs, "ground" is just one of the patches around the domain. I have it so named, because I also run cases with wheeled vehicles, and in those cases I change its details as required.

[boffin5]

I got my case to run, but only if I set the U boundary condition for both air and solid, to 0.01.


Obviously, this isn't realistic; I would like the air/U value to be 56 meters per second. But if I make that change, it fails.


I hope that someone far smarter than myself can make use of that clue.

[Bloerb]

Initilize both pressure fields with the same value, instead of 0 use 84559 for all values as in the other file.

Can you solve just the air region and get a reasonable solution? Please post your constant folder, i.e your thermoPhysicalProperties. Just remove the polyMesh folders and upload them.


And please add the output from checkMesh. From what you have uploaded it should run just fine. Maybe your turbulence quantities are not well initialized. Difficult to tell from what i can see

[boffin5]

Thank you Bloerb, for responding to my problem. Actually, I created the thread 'Meredith Effect' and posted it, but to simplify things, I posted basically the same question under this link:


chtMultiRegionFoam diverges in second iteration


Files I attached in this link are for the same case, albeit with a much simpler body. And it fails in the same way. If I can get this one to run, it will be my template for the Meredith case.


In the 'diverges-second-iteration' thread you can find a zip file for my constant folder, and the checkMesh logs for both fluid and solid.


I appreciate your taking the time to help!


Alan w