OpenFOAM : flow around a sphere at low Reynold number

maxou1993 · July 2, 2015, 18:51

Hi,
I meet several problems solving horizontal laminar flow (I'm not considering eddies and turbulence) around a sphere with OpenFOAM, so I have several questions :
1) for both icoFoam and simpleFoam solvers, the term (v.grad)(v) is included in the Navier-Stokes equation. Is there a not too dificult (I'm not experienced in OpenFOAM and i'm new on this forum

) means to definitely delete this convective term from Navier-Stokes by modifying these solvers ? Indeed, I would like to get the symmetrical (to the median vertical plane) steady state u solution of the horizontal Stokes flow but I only get a slightly assymetric solution (due to the (v.grad)(v) term I think). I have already decreased the Reynolds number as much as possible (so as to neglect the convective term relatively to the diffusive viscous term) but I would like to fix the problem more properly.
2) Then, I read on this website that this "Stokes steady-state flow around a sphere" can be solved by icoFoam . Unfortunately it is not the case for me (i get an unrealistic solution after a while, even if some time steps after the beginning of the simulation, u and p space distributions seem relevant) ... Do you have any idea of the origin of the problem ? Could the problem come from the term $\rho$ *dv/dt taken into account in the NS of the icoFoam solver (the theoretical Stokes equation neglects this temporal term) ?
Here are the main parameters I used for my simulation :
Re=0.002; Cubic mesh edge size : 2; Sphere diameter : 0.2; Viscosity : 10
In 0 files :
Inlet : Velocity=0.1m/s Pressure=zeroGradient
Outlet : Fixed pressure=0 Velocity=zeroGradient
Top, Bottom, Back, Front : cyclic
Sphere : automatically set as a wall by snappyHexMesh, Velocity=0, pressure: zeroGradient
internalfield : 0.1m/s for the velocity and a value for p/\rho calculated by using the pressure loss for a Poiseuille flow.
4) Are there best boundaries and initial conditions (for instance, setting my inlet velocity equal to 0 makes my model wrong since) that i could use (in addition to increase the cubicmeshedgesize/spherediameter ratio for example, for the inlet velocity condition) ?
5) A question more general about patches and walls which are generated in OpenFOAM in order to model our fluid boundaries (if i well understood) : when we fix a conditon for the pressure (respectively the velocity) on a fluid boundary, is it possible not to have to fix a condition for the velocity (respectively the pressure) for this fluid boundary ? (i would have liked to fix only velocity inlet and pressure outlet for my Stokes problem; fortunately, for the sphere boundary it can be shown theoretically that the zeroGradient condition for the pressure is true)
Thank you in advance.

maxou1993 · July 7, 2015, 05:13

I recognize that my previous post was long but it would be great if i had an answer to the main following question :
About patches and walls which are generated in OpenFOAM in order to model our fluid boundaries (if i well understood) : when we fix a condition for the pressure (respectively the velocity) on a fluid boundary, is it possible not to have to fix a condition for the velocity (respectively the pressure) for this fluid boundary ? (i would have liked to fix only velocity inlet and pressure outlet for my Stokes problem: are there options to have this type of boundary conditions if we can't fix only pressure on a patch or onyl velocity ?). Thank you in advance.

July 2, 2015, 18:51	OpenFOAM : flow around a sphere at low Reynold number	#1
maxou1993 New Member maxime Join Date: Jul 2015 Posts: 8 Rep Power: 10	Hi, I meet several problems solving horizontal laminar flow (I'm not considering eddies and turbulence) around a sphere with OpenFOAM, so I have several questions : 1) for both icoFoam and simpleFoam solvers, the term (v.grad)(v) is included in the Navier-Stokes equation. Is there a not too dificult (I'm not experienced in OpenFOAM and i'm new on this forum ) means to definitely delete this convective term from Navier-Stokes by modifying these solvers ? Indeed, I would like to get the symmetrical (to the median vertical plane) steady state u solution of the horizontal Stokes flow but I only get a slightly assymetric solution (due to the (v.grad)(v) term I think). I have already decreased the Reynolds number as much as possible (so as to neglect the convective term relatively to the diffusive viscous term) but I would like to fix the problem more properly. 2) Then, I read on this website that this "Stokes steady-state flow around a sphere" can be solved by icoFoam . Unfortunately it is not the case for me (i get an unrealistic solution after a while, even if some time steps after the beginning of the simulation, u and p space distributions seem relevant) ... Do you have any idea of the origin of the problem ? Could the problem come from the term $\rho$ *dv/dt taken into account in the NS of the icoFoam solver (the theoretical Stokes equation neglects this temporal term) ? Here are the main parameters I used for my simulation : Re=0.002; Cubic mesh edge size : 2; Sphere diameter : 0.2; Viscosity : 10 In 0 files : Inlet : Velocity=0.1m/s Pressure=zeroGradient Outlet : Fixed pressure=0 Velocity=zeroGradient Top, Bottom, Back, Front : cyclic Sphere : automatically set as a wall by snappyHexMesh, Velocity=0, pressure: zeroGradient internalfield : 0.1m/s for the velocity and a value for p/\rho calculated by using the pressure loss for a Poiseuille flow. 4) Are there best boundaries and initial conditions (for instance, setting my inlet velocity equal to 0 makes my model wrong since) that i could use (in addition to increase the cubicmeshedgesize/spherediameter ratio for example, for the inlet velocity condition) ? 5) A question more general about patches and walls which are generated in OpenFOAM in order to model our fluid boundaries (if i well understood) : when we fix a conditon for the pressure (respectively the velocity) on a fluid boundary, is it possible not to have to fix a condition for the velocity (respectively the pressure) for this fluid boundary ? (i would have liked to fix only velocity inlet and pressure outlet for my Stokes problem; fortunately, for the sphere boundary it can be shown theoretically that the zeroGradient condition for the pressure is true) Thank you in advance.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[mesh manipulation] Mesh Refinement	Luiz Eduardo Bittencourt Sampaio (Sampaio)	OpenFOAM Meshing & Mesh Conversion	42	January 8, 2017 12:55
[OpenFOAM.org] OF2.3.1 + OS13.2 - Trying to use the dummy Pstream library	aylalisa	OpenFOAM Installation	23	June 15, 2015 14:49
Flow over sphere low reynolds number	kadirggg	Main CFD Forum	2	November 2, 2010 14:06
LOW REYNOLD NUMBER	Limonghi	Main CFD Forum	1	March 13, 2007 11:37
[Commercial meshers] Trimmed cell and embedded refinement mesh conversion issues	michele	OpenFOAM Meshing & Mesh Conversion	2	July 15, 2005 04:15

July 7, 2015, 05:13		#2
maxou1993 New Member maxime Join Date: Jul 2015 Posts: 8 Rep Power: 10	I recognize that my previous post was long but it would be great if i had an answer to the main following question : About patches and walls which are generated in OpenFOAM in order to model our fluid boundaries (if i well understood) : when we fix a condition for the pressure (respectively the velocity) on a fluid boundary, is it possible not to have to fix a condition for the velocity (respectively the pressure) for this fluid boundary ? (i would have liked to fix only velocity inlet and pressure outlet for my Stokes problem: are there options to have this type of boundary conditions if we can't fix only pressure on a patch or onyl velocity ?). Thank you in advance.