|
[Sponsors] |
rhoPorousSimpleFoam LTNE segregated or coupled? OF6 |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
November 6, 2018, 09:34 |
rhoPorousSimpleFoam LTNE segregated or coupled? OF6
|
#1 |
New Member
Pamela Dahl
Join Date: Jan 2014
Posts: 1
Rep Power: 0 |
Hello, everybody.
I am trying to solve for compressible, turbulent flow, steady state, porous media, without assuming local thermal equilibrium. Thus, for LTNE I added an energy equation for the solid matrix with an energy source, linked to the fluid energy equation by the convective heat transfer term. The enthalpies seem not to converge. I've tried making an internal loop (the residuals tell me that actually no more than 2 or 3 iterations are necessary), and limiting the minimum solid enthalpy. To no avail: the relaxation factors I need just to make the solver run are terrible (0.00001, for both enthalpies). (I tried everything in the book: starting robustly/finishing accurately, the mesh is extremely orthogonal, probes to monitor enthalpies values, very low residuals to stop Simple: 10⁻8, etc.) Has anybody tried anything similar? Should I stop trying to make Simple work, and migrate asap to rhoPimpleFoam (adding porosity explicitly via fvOptions)? Also, I have just found a coupledFvScalarMatrix class that Klas Jareteg uses to couple regions. I'd like to try that to couple the fields he and hs, in order to solve the PDEs at once, instead of segregatedly. But OpenFOAM 6 does not have it, it seems. Has anybody tried it? The implementation seems kind of cumbersome... Many, many thanks in advance for any clue... Here's my EEqn.H file: { int valori = 5;//number of calculations of both EEqns in one internal iteration. int counteri = 1;//counter initialization. volScalarField& he = thermo.he(); fvScalarMatrix EgEqn ( fvm::div(phi, he) + ( he.name() == "e" ? fvc::div(phi, volScalarField("Ekp", 0.5*magSqr(U) + p/rho)) : fvc::div(phi, volScalarField("K", 0.5*magSqr(U))) ) - fvm::laplacian(turbulence->alphaEff(), he) - (hConv*areaEsp)*((hs/Cps)-(he/Cp)) == fvOptions(rho, he) ); fvScalarMatrix EsEqn ( fvm::laplacian(alphas, hs) + Q - hConv*areaEsp*((hs/Cps)-(he/Cp)) ); if (hs < minHs) hs == minHs;//limiting the solid enthalpy at its minimum value. while (counteri <= valori) { EgEqn.relax(); fvOptions.constrain(EgEqn); EgEqn.solve(); fvOptions.correct(he); EsEqn.relax(); EsEqn.solve(); if (hs < minHs) hs == minHs;//limiting the solid enthalpy at its minimum value. thermo.correct(); counteri = counteri + 1; } } Last edited by pamela; November 6, 2018 at 09:36. Reason: Forgot to say it was turbulent regime! |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Reasons for decoupling the pressure and velocity computations | highlando | Main CFD Forum | 36 | April 10, 2017 11:41 |
Segregated or Coupled flow? | hamzamotiwala | STAR-CCM+ | 7 | October 25, 2011 20:35 |
enthalpy in segregated and coupled | naomi | FLUENT | 4 | August 2, 2007 07:47 |
switching from coupled solver to segregated | Oz | FLUENT | 2 | November 8, 2006 17:02 |
Segregated or coupled solver? | Nori | FLUENT | 0 | September 29, 2005 15:24 |