CFD中文网

cyberk

帮李老师补充一下OpenFOAM里面average的实现，对于vol场要先插值成surface场

template<class Type>
tmp<GeometricField<Type, fvPatchField, volMesh>>
average
(
    const GeometricField<Type, fvPatchField, volMesh>& vtf
)
{
    return fvc::average(linearInterpolate(vtf));
}

所以|3|0|2|到average函数里面以后要先插值成

|  3  |  0  |  2  | 
     1.5    1

再基于

$v_{celli}=\frac{\sum (|\bfS_f| v_f)}{\sum |\bfS_f|}$
得到中间网格的值为1.25

cyberk

对于openfoam7中的算例
OpenFOAM/OpenFOAM-7/tutorials/multiphase/twoPhaseEulerFoam/RAS/bubbleColumn

blocks
(
    hex (0 1 2 3 4 5 6 7) (25 75 1) simpleGrading (1 1 1)

将其blockMeshDict中的网格数从(25 75 1)修改为(50 150 1)和(100 300 1)得到bubbleColumnFine和bubbleColumnMorefine。然后运行./Allrun，这两个算例会分别在4.82s和1.785s发散
报错信息如下：


Courant Number mean: 0.193424 max: 2.38733
Max Ur Courant Number = 1.70125
Time = 1.785

PIMPLE: Iteration 1
MULES: Solving for alpha.air
MULES: Solving for alpha.air
alpha.air volume fraction = 0.300065  Min(alpha.air) = -1.42077e-16  Max(alpha.air) = 1.01947
Constructing momentum equations
smoothSolver:  Solving for e.air, Initial residual = 0.00251771, Final residual = 1.55265e-08, No Iterations 2
smoothSolver:  Solving for e.water, Initial residual = 6.01592e-05, Final residual = 5.94292e-08, No Iterations 4
min T.air 300
min T.water 300.167
GAMG:  Solving for p_rgh, Initial residual = 0.000206863, Final residual = 2.12447e-09, No Iterations 6
PIMPLE: Iteration 2
MULES: Solving for alpha.air
MULES: Solving for alpha.air
alpha.air volume fraction = 0.30008  Min(alpha.air) = -0.000167232  Max(alpha.air) = 1.00044
Constructing momentum equations
smoothSolver:  Solving for e.air, Initial residual = 0.000588382, Final residual = 1.82242e-08, No Iterations 2
smoothSolver:  Solving for e.water, Initial residual = 4.07019e-05, Final residual = 5.33076e-08, No Iterations 4
min T.air 300
min T.water 300.167
GAMG:  Solving for p_rgh, Initial residual = 0.000299811, Final residual = 4.64565e-09, No Iterations 6
PIMPLE: Iteration 3
MULES: Solving for alpha.air
MULES: Solving for alpha.air
alpha.air volume fraction = 0.300078  Min(alpha.air) = -0.286261  Max(alpha.air) = 1.3119
Constructing momentum equations
smoothSolver:  Solving for e.air, Initial residual = 0.000643277, Final residual = 7.16972e-08, No Iterations 4
smoothSolver:  Solving for e.water, Initial residual = 3.51151e-05, Final residual = 6.91279e-08, No Iterations 3
min T.air -57.4121
min T.water 300.167
GAMG:  Solving for p_rgh, Initial residual = 0.00045527, Final residual = 8.56624e-09, No Iterations 13
bounding km, min: 2.22028e-16 max: 0.0271421 average: 0.000650568
smoothSolver:  Solving for epsilonm, Initial residual = 0.129979, Final residual = 9.93029e-12, No Iterations 1
bounding epsilonm, min: -0.100887 max: 55.5581 average: 0.246321
smoothSolver:  Solving for km, Initial residual = 6.55536e-12, Final residual = 5.87794e-14, No Iterations 1
bounding km, min: -0.000581692 max: 0.0170913 average: 0.000649872
#0  Foam::error::printStack(Foam::Ostream&) at ??:?
#1  Foam::sigFpe::sigHandler(int) at ??:?
#2  ? in "/lib/x86_64-linux-gnu/libc.so.6"
#3  ? in "/lib/x86_64-linux-gnu/libm.so.6"
#4  pow in "/lib/x86_64-linux-gnu/libm.so.6"
#5  Foam::pow(Foam::Field<double>&, Foam::UList<double> const&, double const&) at ??:?
#6  Foam::tmp<Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> > Foam::pow<Foam::fvPatchField, Foam::volMesh>(Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> const&, Foam::dimensioned<double> const&) at ??:?
#7  Foam::dragModels::SchillerNaumann::CdRe() const at ??:?
#8  Foam::dragModel::Ki() const at ??:?
#9  Foam::dragModel::K() const at ??:?
#10  Foam::BlendedInterfacialModel<Foam::dragModel>::K() const at ??:?
#11  Foam::twoPhaseSystem::Kd() const at ??:?
#12  Foam::RASModels::mixtureKEpsilon<Foam::EddyDiffusivity<Foam::ThermalDiffusivity<Foam::PhaseCompressibleTurbulenceModel<Foam::phaseModel> > > >::Ct2() const at ??:?
#13  Foam::RASModels::mixtureKEpsilon<Foam::EddyDiffusivity<Foam::ThermalDiffusivity<Foam::PhaseCompressibleTurbulenceModel<Foam::phaseModel> > > >::correct() at ??:?
#14  Foam::twoPhaseSystem::correctTurbulence() at ??:?

通过paraview查看发散前的结果，可以看到epsilonm的值在局部区域有震荡（epsilonm的量程为2e-16到384）
2f065a0d-4439-4cb3-8341-0fdfcb3b5ad0-图片.png

调整松弛因子也未能解决问题，想请教一下各位前辈这种发散是由什么造成的？又该如何解决？

cyberk

@shawJ
您好，我也觉得应该是驱动方面的问题，但我没有时间仔细去研究。
我目前的解决方案就是我上一条回复中提到的，采用官网的apt安装方式。

cyberk

@李东岳 VMware vSphere，在服务器上使用的虚拟机

cyberk

@李东岳李老师好，第一个7763是真系统跑的，第二个7763是虚拟机跑的。

cyberk

CPU型号：Intel(R)  i7 12700
系统：真系统 Ubuntu18.04
OpenFOAM版本：OpenFOAM-7
20 284.15
16 286.09
8 211.63
4 267.82
2 361.69
1 618.32

cyberk

CPU型号：AMD 霄龙 7763双路
系统：虚拟机 Ubuntu18.04
OpenFOAM版本：OpenFOAM-7
128 44.44
64 47.05
32 63.96
16 140.24
8 159.79
4 253.19
2 414.95
1 891.46

cyberk

CPU型号：AMD 霄龙 7763双路
系统：真系统 Ubuntu18.04
OpenFOAM版本：OpenFOAM-7
128 36.17
64 42.78
32 73.29
16 99.53
8 129.85
4 240.07
2 414.92
1 861.36

cyberk

更新一下
在没有独显的服务器上采用官网源码编译的方式[1]会黑屏，但是采用官网apt安装的方式[2]就正常，原因不明

[1]https://openfoam.org/download/7-source/
[2]https://openfoam.org/download/7-ubuntu/

cyberk

提供一下资料出处
https://zhuanlan.zhihu.com/p/566101670
这里面提到Ubuntu18.04和Ubuntu22.04关闭加速3D图形会导致paraview黑屏，而Ubuntu20.04开启加速3D图形则会报错，具体原因不明

cyberk

我在新服务器上安装了ubuntu18.04的真系统，然后按照官方教程安装了openfoam7以及paraview，但是发现打开paraFoam是黑屏，旧服务器是正常的，但是系统和软件不是我装的。通过查资料发现虚拟机上也有类似的问题。就是在这个界面的加速3D图形如果取消，就也会导致paraview黑屏，我猜测这是同一个原因导致的，所以想问一下真系统里面如何设置打开加速3D图形？
c3d5ea45-6669-4b25-9079-a76b8a4fa807-图片.png
补充一些信息，paraview只是黑屏，但是命令行不会报错，而且该有的按钮也能点击。

cyberk

@caijinjin 你好，请问你是用ICEM画的网格吗？我最近也在尝试用ICEM画搅拌槽的网格，可以看我之前的发帖记录。但是我画的网格在计算的时候搅拌桨面上的epsilon值存在一些问题，初步排查原因似乎是由于周向划分数量导致桨叶处第一层网格的y+过小或者过大？想问一下你遇到过这种情况吗？可以加微信交流一下吗？我的微信是wangyu1895

cyberk

@李东岳李老师您好，除了上面提到1.更换结构网格 2.采用新版本的openfoam 之外，您说的调整设置具体是指fvscheme里的离散格式和fvsolution中的残差、矩阵求解器、残差等吗？您能给一个大概的调整方向的建议吗？我看到您之前发的帖子中有提到twoPhaseEulerFoam等pimple类算法的的亚松弛因子就算给1也会有保证矩阵对角占优，我观察pimple系列的求解器的tut中亚松弛因子也基本都是给的1，我如果降低亚松弛因子会有助于提高twoPhaseEulerFoam计算的稳定性吗？