rhoPorousMRFSimpleFoam/compressor simulation blows up

mpat · August 8, 2012, 09:42

Hi all,

I am trying to simulate a compressible flow in centrifugal compressor (for turbocharger) with rhoPorousMRFSimpleFoam (OF 2.1.1 Version). I am modeling the inlet duct, the impeller and the outlet diffusor. I have 3 domains: stator-rotor-stator. For Interfaces I use cyclicAMI. To save computing time I am using cyclic boundary conditions. The impeller is rotating with 160000 rpm.

I started using k-omega SST Modell with wall function. This simulation runs for ca. 600 Iterations but eventually blows up because time step continuity errors rise. Pressure and omega residuals diverge. Then I checked my y+ with yPlusRAS utility and it reported y+ < 1. When I use no wall functions for k and mut the simulation runs for ca. 5500 Iterations but then same happens. In ParaView I see one cell (on blade) where pressure and omega increase dramatically.

I think it is mesh problem. My mesh is generated with ICEM CFD. It is tetra mesh with prism boundary layer. ICEM reports no errors and the quality is good enough for CFX. In OpenFOAM checkMesh reports some non- orthogonal and few skew faces, but with checkMesh –allTopology –allGeometrie there are some under-determined cells additionally reported. Some of these faces/cells are located on blade where error occurs.

How do I proceed with these errors? How to avoid these under-determined cells? Could this be the reason for the simulation blow up? I would really appreciate your help. Thanks in advance ;-)
Maja

checkMesh.log
Case

kalyangoparaju · August 9, 2012, 14:54

Hi Maja,

If the checkMesh reports nonOrthogonal and skewed cells, then there is a very good chance that the blowup is because of the mesh. Try changing the nNonOrthogonalCorrectors and also initializing the simulation using potentialFoam. This should definitely help the case.

As an example - I was analyzing the flow through a very complex 2 substrate catalytic converter for which I built the mesh in snappy. The mesh looked good and checkMesh showed an error of only 3 skewed cells. But when I ran the case, I never got it to converge to good results if the nNonOrthogonalCorrectors were less than 10 !!!! I know that sounds absurd but it is true. I still am not able to figure out why I had to use so many correctors if according to foam I had no nonOrthogonal cells and only 3 skewed cells.

So, try fiddling with those correctors. Most of the time, your potentialFoam solution itself can be a clue to whether your solution will converge or not ( continuity and interpolated U error should be as close to 1 as possible if not below it ) You can use this criterion to fine tune the number of correctors u need.

Kalyan

mpat · August 10, 2012, 10:52

Huhu Kalyan,

Thank you for your advice. Can you please tell me how can I initialize my flow with potentialFoam if the flow velocity results from impeller rotational speed? I have Total Pressure, Total Temperature and zeroGradient for velocity on Inlet. I have tested it with flowRateInletVelocity and pressureInletOutletVelocity on outlet. I need ca. 15 nNonOrthogonalCorrectors but same error happens eventually:

Code:

Time = 519

smoothSolver:  Solving for Ux, Initial residual = 0.00689174020389124, Final residual = 0.0001811298084334255, No Iterations 2
smoothSolver:  Solving for Uy, Initial residual = 0.007125217146934263, Final residual = 0.0001859035926144115, No Iterations 2
smoothSolver:  Solving for Uz, Initial residual = 0.01007109609581601, Final residual = 0.0002720657574050938, No Iterations 2
DILUPBiCG:  Solving for h, Initial residual = 0.01189626781015002, Final residual = 0.0004430857512486936, No Iterations 1
GAMG:  Solving for p, Initial residual = 0.525246072638058, Final residual = 0.02234065612714114, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.1210393076383517, Final residual = 0.004059726915845455, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.03711323447992431, Final residual = 0.001207459503253744, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.01390244007788457, Final residual = 0.0004040200625956493, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.005329744823495349, Final residual = 0.0001524608155663665, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.002159844936659785, Final residual = 6.187981539139611e-05, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.0009000768414453766, Final residual = 2.614056448078217e-05, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.000387175087517713, Final residual = 1.142583341601287e-05, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.0001718465581034602, Final residual = 5.160508215150291e-06, No Iterations 2
GAMG:  Solving for p, Initial residual = 7.794569833811138e-05, Final residual = 2.266434677289963e-06, No Iterations 2
GAMG:  Solving for p, Initial residual = 3.477625348436517e-05, Final residual = 1.019722809542639e-06, No Iterations 2
GAMG:  Solving for p, Initial residual = 1.57441454955239e-05, Final residual = 4.360974326864854e-07, No Iterations 2
GAMG:  Solving for p, Initial residual = 6.92933500221939e-06, Final residual = 2.016293952604061e-07, No Iterations 2
GAMG:  Solving for p, Initial residual = 3.173510949418744e-06, Final residual = 8.7284304841048e-08, No Iterations 2
GAMG:  Solving for p, Initial residual = 1.421339504689786e-06, Final residual = 4.250775374814685e-08, No Iterations 2
GAMG:  Solving for p, Initial residual = 6.693670543287874e-07, Final residual = 1.860688605016314e-08, No Iterations 2
time step continuity errors : sum local = 0.0006145710207738517, global = 2.659732169674872e-05, cumulative = 0.0003956096006971732
rho max/min : 1.785699960099344 0.5971102771881479
smoothSolver:  Solving for omega, Initial residual = 0.002288630882997602, Final residual = 4.497457628363702e-05, No Iterations 1
smoothSolver:  Solving for k, Initial residual = 0.003047717993801081, Final residual = 6.695371931754372e-05, No Iterations 1
ExecutionTime = 21.46 s  ClockTime = 22 s

Time = 520

smoothSolver:  Solving for Ux, Initial residual = 0.00685892500157054, Final residual = 0.0001781598558160968, No Iterations 2
smoothSolver:  Solving for Uy, Initial residual = 0.007116418862045025, Final residual = 0.0001839081410649369, No Iterations 2
smoothSolver:  Solving for Uz, Initial residual = 0.009954523806062012, Final residual = 0.0002701273723458191, No Iterations 2
DILUPBiCG:  Solving for h, Initial residual = 0.01185654992594479, Final residual = 0.0004422577762352827, No Iterations 1
GAMG:  Solving for p, Initial residual = 0.5217401646589753, Final residual = 0.02240546892794447, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.1204252318653997, Final residual = 0.004028617590806268, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.03687051608255586, Final residual = 0.001180687618896112, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.01358671728843339, Final residual = 0.0003956252492256043, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.005212434153393288, Final residual = 0.0001467219739165364, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.002095649830267571, Final residual = 5.831582947256874e-05, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.0008761050175603406, Final residual = 2.583571916038806e-05, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.000383257128952095, Final residual = 1.131748182531108e-05, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.0001711316968720861, Final residual = 5.052897435634979e-06, No Iterations 2
GAMG:  Solving for p, Initial residual = 7.631093755718225e-05, Final residual = 2.260425128892119e-06, No Iterations 2
GAMG:  Solving for p, Initial residual = 3.449474051643401e-05, Final residual = 1.012405010496451e-06, No Iterations 2
GAMG:  Solving for p, Initial residual = 1.549985569778341e-05, Final residual = 4.635358865057642e-07, No Iterations 2
GAMG:  Solving for p, Initial residual = 7.111843979700321e-06, Final residual = 2.083560348490843e-07, No Iterations 2
GAMG:  Solving for p, Initial residual = 3.208706672168953e-06, Final residual = 9.71363574529989e-08, No Iterations 2
GAMG:  Solving for p, Initial residual = 1.492962361641866e-06, Final residual = 4.395685059333027e-08, No Iterations 2
GAMG:  Solving for p, Initial residual = 6.775063608301946e-07, Final residual = 2.080044109382489e-08, No Iterations 2
time step continuity errors : sum local = 0.0007113392680664244, global = -2.948635003943702e-06, cumulative = 0.0003926609656932295
rho max/min : 1.788032591204212 0.5965712276411931
smoothSolver:  Solving for omega, Initial residual = nan, Final residual = nan, No Iterations 1000
smoothSolver:  Solving for k, Initial residual = nan, Final residual = nan, No Iterations 1000
ExecutionTime = 26.1 s  ClockTime = 26 s

Time = 521

smoothSolver:  Solving for Ux, Initial residual = nan, Final residual = nan, No Iterations 1000
smoothSolver:  Solving for Uy, Initial residual = nan, Final residual = nan, No Iterations 1000
smoothSolver:  Solving for Uz, Initial residual = nan, Final residual = nan, No Iterations 1000
DILUPBiCG:  Solving for h:  solution singularity

I did this with two mesh (only tetra), one coarse, one fine, but it did not help.

Code:

Create time

Create polyMesh for time = 0

Time = 0

Mesh stats
    points:           14639
    faces:            149430
    internal faces:   139886
    cells:            72329
    boundary patches: 17
    point zones:      0
    face zones:       3
    cell zones:       3

Overall number of cells of each type:
    hexahedra:     0
    prisms:        0
    wedges:        0
    pyramids:      0
    tet wedges:    0
    tetrahedra:    72329
    polyhedra:     0

Checking topology...
    Boundary definition OK.
    Cell to face addressing OK.
    Point usage OK.
    Upper triangular ordering OK.
    Face vertices OK.
    Topological cell zip-up check OK.
    Face-face connectivity OK.
  <<Writing 93 cells with with two non-boundary faces to set twoInternalFacesCells
   *Number of regions: 3
    The mesh has multiple regions which are not connected by any face.
  <<Writing region information to "0/cellToRegion"

Checking patch topology for multiply connected surfaces ...
    Patch               Faces    Points   Surface topology                   Bounding box
    ROTOR_HUB           543      381      ok (non-closed singly connected)   (-0.005940277634 -0.02298990972 0.01023706532) (0.01757740074 -0.0006413035989 0.03043711472)
    ROTOR_BLADE         2101     1115     ok (non-closed singly connected)   (-0.01002530289 -0.02249982629 0.0102374897) (0.01391894245 -0.003520271167 0.02794919761)
    ROTOR_SHROUD        930      524      ok (non-closed singly connected)   (-0.01691088676 -0.02299648304 0.01732514381) (0.01749520845 -0.001966941138 0.03042087057)
    INLET_DUCT          562      326      ok (non-closed singly connected)   (-0.01691089439 -0.01562956807 0.03043420009) (-0.002200154341 -0.0006411794424 0.05291999435)
    INLET               145      90       ok (non-closed singly connected)   (-0.01691079521 -0.01562951296 0.05291694634) (-0.002200275083 -0.0006411794424 0.05291999435)
    OUTLET              463      268      ok (non-closed singly connected)   (-0.008105474472 -0.04616911035 0.01021563911) (0.03538647188 -0.02966462808 0.01584203791)
    OUTLET_DUCT         977      563      ok (non-closed singly connected)   (-0.008105474472 -0.04616911035 0.01021563911) (0.03538647188 -0.01484315558 0.01732754156)
    OUTLET_INTERFACE_SLAVE182      111      ok (non-closed singly connected)   (-0.004198823929 -0.02300056671 0.01023706436) (0.01757749445 -0.01484315558 0.01732754156)
    OUTLET_INTERFACE_MASTER434      252      ok (non-closed singly connected)   (-0.004199062824 -0.0229972754 0.01023706532) (0.01757740074 -0.01484326831 0.01733062408)
    INLET_INTERFACE_SLAVE147      90       ok (non-closed singly connected)   (-0.01691089439 -0.01562918925 0.03043420009) (-0.002200161517 -0.0006413034201 0.03043730953)
    INLET_INTERFACE_MASTER386      219      ok (non-closed singly connected)   (-0.01691088676 -0.0156291945 0.03041778481) (-0.002200161362 -0.0006413035989 0.03043711472)
    INLET_PER1          197      120      ok (non-closed singly connected)   (-0.01691089439 -0.001968802096 0.03043711472) (-0.005511092863 -0.0006411794424 0.05291999435)
    INLET_PER2          197      120      ok (non-closed singly connected)   (-0.006751238925 -0.01562956807 0.03043711472) (-0.002200154341 -0.005093347119 0.05291999435)
    OUTLET_PER1         236      138      ok (non-closed singly connected)   (-0.008105474472 -0.04547789764 0.01021563911) (-0.004065802574 -0.02262504959 0.01732515335)
    OUTLET_PER2         236      138      ok (non-closed singly connected)   (0.01749445574 -0.02975333815 0.01021563911) (0.03538647188 -0.01484315558 0.01732515335)
    ROTOR_PER1          904      499      ok (non-closed singly connected)   (-0.01691088676 -0.02264410973 0.01023706532) (-0.003978910923 -0.0006413035989 0.03043711472)
    ROTOR_PER2          904      499      ok (non-closed singly connected)   (-0.006751214565 -0.01706289718 0.01023706532) (0.01757740074 -0.005093398461 0.03043711472)

Checking geometry...
    Overall domain bounding box (-0.01691089439 -0.04616911035 0.01021563911) (0.03538647188 -0.0006411794424 0.05291999435)
    Mesh (non-empty, non-wedge) directions (1 1 1)
    Mesh (non-empty) directions (1 1 1)
    Boundary openness (-1.008508880210902e-16 6.737618413690084e-17 1.747059006460879e-17) OK.
    Max cell openness = 2.258267238391119e-16 OK.
    Max aspect ratio = 4.580087858669184 OK.
    Minumum face area = 2.13575888608433e-08. Maximum face area = 3.238582471401818e-06.  Face area magnitudes OK.
    Min volume = 2.152195729490874e-12. Max volume = 1.713436717452314e-09.  Total volume = 1.106455360888172e-05.  Cell volumes OK.
    Mesh non-orthogonality Max: 65.86404963883847 average: 18.07799753123574
    Non-orthogonality check OK.
    Face pyramids OK.
    Max skewness = 0.5164240204619981 OK.
    Coupled point location match (average 2.823201989552231e-12) OK.
    Face tets OK.
    Min/max edge length = 0.0001148631272861437 0.003225834417998265 OK.
    All angles in faces OK.
    All face flatness OK.
    Cell determinant (wellposedness) : minimum: 0 average: 1.449341833456967
 ***Cells with small determinant found, number of cells: 266
  <<Writing 266 under-determined cells to set underdeterminedCells
    Concave cell check OK.

Failed 1 mesh checks.

End

Code:

Create time

Create polyMesh for time = 0

Time = 0

Mesh stats
    points:           119506
    faces:            1285019
    internal faces:   1228117
    cells:            628284
    boundary patches: 17
    point zones:      0
    face zones:       3
    cell zones:       3

Overall number of cells of each type:
    hexahedra:     0
    prisms:        0
    wedges:        0
    pyramids:      0
    tet wedges:    0
    tetrahedra:    628284
    polyhedra:     0

Checking topology...
    Boundary definition OK.
    Cell to face addressing OK.
    Point usage OK.
    Upper triangular ordering OK.
    Face vertices OK.
    Topological cell zip-up check OK.
    Face-face connectivity OK.
  <<Writing 231 cells with with two non-boundary faces to set twoInternalFacesCells
   *Number of regions: 3
    The mesh has multiple regions which are not connected by any face.
  <<Writing region information to "0/cellToRegion"

Checking patch topology for multiply connected surfaces ...
    Patch               Faces    Points   Surface topology                   Bounding box
    ROTOR_HUB           3534     2059     ok (non-closed singly connected)   (-0.005945404573 -0.02299999161 0.01023706532) (0.01757740074 -0.0006413035989 0.03043711472)
    ROTOR_BLADE         15837    8090     ok (non-closed singly connected)   (-0.01006028831 -0.02249981902 0.0102374897) (0.01391894245 -0.003480050475 0.02794864382)
    ROTOR_SHROUD        6125     3202     ok (non-closed singly connected)   (-0.01691088676 -0.02299888183 0.01732514381) (0.01749520845 -0.001966371457 0.03042126661)
    INLET_DUCT          2550     1384     ok (non-closed singly connected)   (-0.01691089439 -0.01562956806 0.03043419193) (-0.002200154339 -0.0006411794424 0.05291999435)
    INLET               635      356      ok (non-closed singly connected)   (-0.01691079521 -0.01562951296 0.05291688432) (-0.002200275083 -0.0006411794424 0.05291999435)
    OUTLET              1274     714      ok (non-closed singly connected)   (-0.008105474472 -0.04616953992 0.01021563911) (0.03538647188 -0.02966462808 0.01584203305)
    OUTLET_DUCT         7277     3809     ok (non-closed singly connected)   (-0.008105474472 -0.04616953992 0.01021563911) (0.03538647188 -0.01484315558 0.01732754528)
    OUTLET_INTERFACE_SLAVE837      465      ok (non-closed singly connected)   (-0.004198823929 -0.023002 0.01023706436) (0.01757749445 -0.01484315558 0.01732754528)
    OUTLET_INTERFACE_MASTER2663     1412     ok (non-closed singly connected)   (-0.004199062824 -0.02299999161 0.01023706532) (0.01757740074 -0.01484326831 0.01733062769)
    INLET_INTERFACE_SLAVE667      372      ok (non-closed singly connected)   (-0.01691089439 -0.01562918925 0.03043419193) (-0.002200161517 -0.0006413034201 0.03043731022)
    INLET_INTERFACE_MASTER531      298      ok (non-closed singly connected)   (-0.01691088676 -0.0156291945 0.03041780696) (-0.002200161362 -0.0006413035989 0.03043711472)
    INLET_PER1          1230     671      ok (non-closed singly connected)   (-0.01691089439 -0.00196880207 0.03043711472) (-0.005511092863 -0.0006411794424 0.05291999435)
    INLET_PER2          1230     671      ok (non-closed singly connected)   (-0.006751238924 -0.01562956806 0.03043711472) (-0.002200154339 -0.005093347119 0.05291999435)
    OUTLET_PER1         667      376      ok (non-closed singly connected)   (-0.008105474472 -0.04547789764 0.01021563911) (-0.004065802574 -0.02262504959 0.01732515335)
    OUTLET_PER2         667      376      ok (non-closed singly connected)   (0.01749445574 -0.02975333815 0.01021563911) (0.03538647188 -0.01484315558 0.01732515335)
    ROTOR_PER1          5589     2908     ok (non-closed singly connected)   (-0.01691088676 -0.02264410973 0.01023706532) (-0.003978910923 -0.0006413035989 0.03043711472)
    ROTOR_PER2          5589     2908     ok (non-closed singly connected)   (-0.006751214566 -0.01706552291 0.01023706532) (0.01757740074 -0.005093398461 0.03043711472)

Checking geometry...
    Overall domain bounding box (-0.01691089439 -0.04616953992 0.01021563911) (0.03538647188 -0.0006411794424 0.05291999435)
    Mesh (non-empty, non-wedge) directions (1 1 1)
    Mesh (non-empty) directions (1 1 1)
    Boundary openness (-1.333273483219948e-16 2.025107292171666e-16 -1.062012456559783e-16) OK.
    Max cell openness = 2.372167083931545e-16 OK.
    Max aspect ratio = 4.349712291115716 OK.
    Minumum face area = 2.251662057643688e-09. Maximum face area = 6.317981495047761e-07.  Face area magnitudes OK.
    Min volume = 9.282675234161758e-14. Max volume = 1.647429053664735e-10.  Total volume = 1.107251948954049e-05.  Cell volumes OK.
    Mesh non-orthogonality Max: 50.17858023799637 average: 16.47284712748219
    Non-orthogonality check OK.
    Face pyramids OK.
    Max skewness = 0.6033462107301415 OK.
    Coupled point location match (average 2.327840432511681e-12) OK.
    Face tets OK.
    Min/max edge length = 4.764886039802482e-05 0.001414873923756246 OK.
    All angles in faces OK.
    All face flatness OK.
    Cell determinant (wellposedness) : minimum: 0 average: 1.574006420034666
 ***Cells with small determinant found, number of cells: 773
  <<Writing 773 under-determined cells to set underdeterminedCells
    Concave cell check OK.

Failed 1 mesh checks.

End

Do you perhaps know what these under- determined cells are or how to avoid them? Because i can find anything about them in forum.

Thanks

Maja

August 8, 2012, 09:42	rhoPorousMRFSimpleFoam/compressor simulation blows up	#1
mpat New Member Maja Patarcic Join Date: May 2012 Posts: 2 Rep Power: 0	Hi all, I am trying to simulate a compressible flow in centrifugal compressor (for turbocharger) with rhoPorousMRFSimpleFoam (OF 2.1.1 Version). I am modeling the inlet duct, the impeller and the outlet diffusor. I have 3 domains: stator-rotor-stator. For Interfaces I use cyclicAMI. To save computing time I am using cyclic boundary conditions. The impeller is rotating with 160000 rpm. I started using k-omega SST Modell with wall function. This simulation runs for ca. 600 Iterations but eventually blows up because time step continuity errors rise. Pressure and omega residuals diverge. Then I checked my y+ with yPlusRAS utility and it reported y+ < 1. When I use no wall functions for k and mut the simulation runs for ca. 5500 Iterations but then same happens. In ParaView I see one cell (on blade) where pressure and omega increase dramatically. I think it is mesh problem. My mesh is generated with ICEM CFD. It is tetra mesh with prism boundary layer. ICEM reports no errors and the quality is good enough for CFX. In OpenFOAM checkMesh reports some non- orthogonal and few skew faces, but with checkMesh –allTopology –allGeometrie there are some under-determined cells additionally reported. Some of these faces/cells are located on blade where error occurs. How do I proceed with these errors? How to avoid these under-determined cells? Could this be the reason for the simulation blow up? I would really appreciate your help. Thanks in advance ;-) Maja checkMesh.log Case

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Velocity blows up suddenly after 30,000+ iterations	lordvon	OpenFOAM Running, Solving & CFD	15	October 19, 2015 14:52
Interfoam blows on parallel run	danvica	OpenFOAM Running, Solving & CFD	16	December 22, 2012 03:09
SRFSimpleFoam tutorial case blows up on laptop (but runs on desktop?)	AHMeng	OpenFOAM Running, Solving & CFD	1	April 27, 2012 09:34
[GAMBIT] Split volume - when reflecting volume, the copy blows up	enfinity	ANSYS Meshing & Geometry	7	May 3, 2011 03:13
help, solution blows up!	noName	Main CFD Forum	3	May 10, 2005 02:26

August 9, 2012, 14:54		#2
kalyangoparaju Member Kalyan Join Date: Oct 2011 Location: Columbus, Ohio Posts: 53 Blog Entries: 1 Rep Power: 15	Hi Maja, If the checkMesh reports nonOrthogonal and skewed cells, then there is a very good chance that the blowup is because of the mesh. Try changing the nNonOrthogonalCorrectors and also initializing the simulation using potentialFoam. This should definitely help the case. As an example - I was analyzing the flow through a very complex 2 substrate catalytic converter for which I built the mesh in snappy. The mesh looked good and checkMesh showed an error of only 3 skewed cells. But when I ran the case, I never got it to converge to good results if the nNonOrthogonalCorrectors were less than 10 !!!! I know that sounds absurd but it is true. I still am not able to figure out why I had to use so many correctors if according to foam I had no nonOrthogonal cells and only 3 skewed cells. So, try fiddling with those correctors. Most of the time, your potentialFoam solution itself can be a clue to whether your solution will converge or not ( continuity and interpolated U error should be as close to 1 as possible if not below it ) You can use this criterion to fine tune the number of correctors u need. Kalyan