CFD中文网

OpenFOAM版本：OpenFOAM v9
求解器 ：multiphaseEulerFoam
模拟对象：圆柱状鼓泡床反应器，多相流，液相反应，无液相进出，气相不反应，底部进气，顶部出气。
算例来源：在tutorials的bubbleColumnEvaporatingReacting算例基础上修改而来，各相初始分布由setFields控制。

问题：即使关掉反应，从数值迭代上看，多相流计算第一个时间步的PIMPLE循环的相分数计算发散，见下日志。存下第一个PIMPLE迭代结果看，气相和液相的基础相莫名其妙地在原本该是0的地方，变成了1，见下图。
确认过setFields正确地初始化了相分布，上半部气相，下半部液相。由于第一次做多相流的算例，无法定位发散的原因是边界设置？constant的多相参数？相分数求解？还是别的问题？求各位大神指点一二，不胜感激。

尝试过的操作：

加松弛 改矩阵迭代求解器，GAMG换PCG，smoothSolver换PBiCGStab 调整气相液相和压力的出口边界类型为zeroGradient 关掉反应 加压力限制limitPressure 将气相从multiComponentMixture改成pureMixture

/*------------------------------------------------------------------------------
算例：
运行meshGen脚本生成网格，allrun脚本运行。
f21a20e8-830b-41ca-8175-7667fab9ab9b-p2.tar.gz

/*------------------------------------------------------------------------------
计算日志，第一个时间步

PIMPLE: Iteration 1 MULES: Solving for alpha.gas MULES: Solving for alpha.liquid gas fraction, min, max = 0.4999984327 0.009998944 0.990001056 liquid fraction, min, max = 0.5000016007 0.01 0.990001056 Phase-sum volume fraction, min, max = 1.000000033 1 1.000001067 MULES: Solving for alpha.gas MULES: Solving for alpha.liquid gas fraction, min, max = 0.4999968323 0.009997888113 0.9900010559 liquid fraction, min, max = 0.500003201 0.01 0.9900021119 Phase-sum volume fraction, min, max = 1.000000033 1 1.000001067 MULES: Solving for alpha.gas MULES: Solving for alpha.liquid gas fraction, min, max = 0.499995232 0.009996832338 0.9900010559 liquid fraction, min, max = 0.5000048013 0.01 0.9900031677 Phase-sum volume fraction, min, max = 1.000000033 1 1.000001067 smoothSolver: Solving for NH4+.liquid, Initial residual = 1.484075337e-05, Final residual = 1.159761323e-16, No Iterations 2 smoothSolver: Solving for NO3-.liquid, Initial residual = 0, Final residual = 0, No Iterations 1 smoothSolver: Solving for O2.liquid, Initial residual = 0.2795617027, Final residual = 6.228131728e-16, No Iterations 2 smoothSolver: Solving for H+.liquid, Initial residual = 0, Final residual = 0, No Iterations 1 Constructing momentum equations Tf.gasAndLiquid: min = 299.9947355, mean = 300.0339633, max = 300.0732112 DILUPBiCGStab: Solving for e.gas, Initial residual = 0.9999999986, Final residual = 7.01286441e-16, No Iterations 1 DILUPBiCGStab: Solving for e.liquid, Initial residual = 0.1121489206, Final residual = 1.679510257e-16, No Iterations 1 gas min/max T 299.9857036 - 300.0249711 liquid min/max T 299.9714439 - 300.032703 DICPCG: Solving for p_rgh, Initial residual = 0.9950749712, Final residual = 0.0007476549794, No Iterations 58 DICPCG: Solving for p_rgh, Initial residual = 0.01716581647, Final residual = 1.54813425e-05, No Iterations 40 DICPCG: Solving for p_rgh, Initial residual = 0.977211124, Final residual = 0.0007350869486, No Iterations 58 DICPCG: Solving for p_rgh, Initial residual = 0.01014511789, Final residual = 6.818727967e-12, No Iterations 107 PIMPLE: Iteration 2 MULES: Solving for alpha.gas MULES: Solving for alpha.liquid DICPBiCGStab: Solving for alpha.gas, Initial residual = 5.821421844e-17, Final residual = 8.934357652e-34, No Iterations 1 DICPBiCGStab: Solving for alpha.liquid, Initial residual = 4.308481516e-17, Final residual = 7.155231175e-34, No Iterations 1 gas fraction, min, max = 0.4310004415 -8.01486001 2.514951181 liquid fraction, min, max = 0.571470277 -0.01414169716 9.01486001 Phase-sum volume fraction, min, max = 1.002470718 0.9165471615 2.502164998 MULES: Solving for alpha.gas MULES: Solving for alpha.liquid DICPBiCGStab: Solving for alpha.gas, Initial residual = 4.38709213e-17, Final residual = 6.868609435e-34, No Iterations 1 DICPBiCGStab: Solving for alpha.liquid, Initial residual = 4.105585202e-17, Final residual = 6.57320396e-34, No Iterations 1 gas fraction, min, max = 0.3970086899 -84.20635147 74.29030785 liquid fraction, min, max = 0.632777971 -73.29030785 85.20635147 Phase-sum volume fraction, min, max = 1.029786661 0.9005093894 99.99686523 MULES: Solving for alpha.gas MULES: Solving for alpha.liquid DICPBiCGStab: Solving for alpha.gas, Initial residual = 2.799140256e-17, Final residual = 4.373723282e-34, No Iterations 1 DICPBiCGStab: Solving for alpha.liquid, Initial residual = 3.729944455e-17, Final residual = 5.877805245e-34, No Iterations 1 gas fraction, min, max = -7.192486694 -84273.15366 35913.14461 liquid fraction, min, max = 0.7041286022 -35912.14461 84274.15366 Phase-sum volume fraction, min, max = -6.488358091 -37604.99677 580.9315927 smoothSolver: Solving for NH4+.liquid, Initial residual = 0.0112574308, Final residual = 4.168831596e-13, No Iterations 13 smoothSolver: Solving for NO3-.liquid, Initial residual = 0, Final residual = 0, No Iterations 1 smoothSolver: Solving for O2.liquid, Initial residual = 0.2270999707, Final residual = 5.714438306e-13, No Iterations 12 smoothSolver: Solving for H+.liquid, Initial residual = 0, Final residual = 0, No Iterations 1 Constructing momentum equations Tf.gasAndLiquid: min = 299.9711527, mean = 299.9985045, max = 300.0301728

强行存下的第一次PIMPLE循环的结果，下半部应该全都0。
5e042630-92c4-4ba5-ba00-ba9f28ba6b20-1653619214(1).png

替代文字

你说上面这个图。这个算力对应的不是这个图。所以不一样。不过如果你增加近气速度的话，应该也是可以的。你加大10倍试试。

@李东岳 在 求并行生成网格以及划分网格的方法 中说：

@number44

那scotch和metis可以并行decompose吗？

不能。

不过我想到一个曲线救国的方案就是用snappyHexMesh生成网格。这个可以并行。只不过仅仅进行refineMesh处理而已。应该也可以。

如果这样的话，分别在各processor去refineMesh，reconstructPar会报错吧

@number44 在 foam-extend的fsiFoam并行效率问题 中说：

foam-extend的icoFsiFoam

您好，请问下，下面两者有什么区别呢？
foam-extend中自带的icoFsiFoam
跟
单独的fsiFoam