
[Sponsors] 
Alletto  2022  Comparison of overset mesh with morphing mesh 

LinkBack  Thread Tools  Search this Thread  Display Modes 
January 21, 2022, 13:20 
Alletto  2022  Comparison of overset mesh with morphing mesh

#1 
Senior Member
Håkan Nilsson
Join Date: Mar 2009
Location: Gothenburg, Sweden
Posts: 202
Rep Power: 16 
Comments and questions are welcomed on the article Alletto, Comparison of overset mesh with morphing mesh: Flow over a forced oscillating and freely oscillating 2D cylinder, OpenFOAM Journal, January 2022, https://doi.org/10.51560/ofj.v2.47.
Last edited by bigphil; April 11, 2022 at 08:02. Reason: Update date format 

January 21, 2022, 13:31 

#2 
Senior Member
Michael Alletto
Join Date: Jun 2018
Location: Bremen
Posts: 601
Rep Power: 14 
Thanks Håkan for opening the thread. I'll try to answer any questions which arise on the paper to my best.
Michael 

February 4, 2022, 02:15 

#3  
New Member
zink
Join Date: Oct 2015
Posts: 28
Rep Power: 8 
Quote:
How to calculate the work done as you have shown in the paper? How to run the case of morphing mesh by considering both fluid and structural forces? I tried to run your case (https://wiki.openfoam.com/1Dof_sprin...ichael_Alletto) I will really appreciate if you can help me. 

February 4, 2022, 07:52 

#4 
Senior Member
Michael Alletto
Join Date: Jun 2018
Location: Bremen
Posts: 601
Rep Power: 14 
Regarding the work done there is the formula in the paper and the python script in the case file


February 5, 2022, 22:44 
Question?

#5 
New Member
zink
Join Date: Oct 2015
Posts: 28
Rep Power: 8 
I am learning Vortex Induced Vibrations and want to seek guidance from you.
I run the tutorial for Morphing mesh case to validate the results of Mittal (2 DoF). I calculated the values for the case of U*=4 , m*=10,Re=100 The values for the simulation parameters come out to be: nu= 1.0496E06 K=148.28 RhoInf=148171.24 The results of the X and Z centre of mass comes to be way lower than the published in your paper. Please guide me. In the below text, I have mentioned my dynamicMeshDict file. sixDoFRigidBodyMotionCoeffs { accelerationRelaxation 1.0; // 1 accelerationDamping 1.0; //1.0 patches (cylinder); innerDistance 0.0008; //1.0 // equal to cylinder Radius outerDistance 0.032; // 0.032 // near the main domain boundary mass 0.03575; centreOfMass (0.0 0.0 0.0); // initially the cylinder is at the origin momentOfInertia (4.3472E07 1.144E08 4.3472E07); g (0 9.81 0); rho rhoInf; rhoInf 14817.12; //1.0e15 // this value should be as per the nu and rho value in transport properties report on; solver { type Newmark; } constraints { zxLine { sixDoFRigidBodyMotionConstraint line; // cylinder can displace only along the zaxis and c axis centreOfRotation (0.0016 0.0 0.0016); // i think initially at z=d location 0.0016 direction (1 0 1); } yAxis { sixDoFRigidBodyMotionConstraint orientation; // blocked the rotation around the yaxes centreOfRotation (0 1 0); } } restraints { verticalHorizontalSpring { sixDoFRigidBodyMotionRestraint linearSpring; anchor (0.0 0.0 0.00); refAttachmentPt (0.0016 0.0 0.0016); // It is attached to the boundary of cylinder (radius) stiffness 148.281; // as the case of U* 4 damping 0.00; restLength 0; } } 

February 6, 2022, 02:09 

#6 
Senior Member
Michael Alletto
Join Date: Jun 2018
Location: Bremen
Posts: 601
Rep Power: 14 
For U* = 4 the oscillations are in general quite low. The cylinder starts to oscillate for U* = 4.7. But in general questions like wy are my results not good are difficult to answer. Better if you ask more specific questions


February 6, 2022, 10:10 

#7  
New Member
zink
Join Date: Oct 2015
Posts: 28
Rep Power: 8 
Quote:
Should the value of inner distance be equal to diameter or radius? What’s the recommended value of outerDistance? The boundary of whole domain or somewhat in between the cylinder and main boundary ? Should the constraint of line be applied for 2DoF? When I removed the line constraint condition, the simulation diverges… do I need to refine the mesh ? Should the value of refAttachmentPt be equal to cylinder diameter or zero? Will really appreciate your guidance... 

February 6, 2022, 15:09 

#8 
Senior Member
Michael Alletto
Join Date: Jun 2018
Location: Bremen
Posts: 601
Rep Power: 14 
It is really difficult for my to understand what you mean. Can you post a question by the time and be more precise. Do you have specific questions to the paper. The topic of the thread is the paper and if you want some help for your work you should open a new thread I think


February 6, 2022, 15:21 

#9  
New Member
zink
Join Date: Oct 2015
Posts: 28
Rep Power: 8 
Quote:
I intend to seek guidance from you. In the paper you have specified linear spring restraint used for the morphing mesh case. As per the OpenFOAM code, linear sping is first order motion equation but the VIV problem is second order motion equation. How come you can get similar results? Please correct me if I am wrong. See below the code for the linear spring. https://www.openfoam.com/documentati...8C_source.html 

February 6, 2022, 15:32 

#10 
Senior Member
Michael Alletto
Join Date: Jun 2018
Location: Bremen
Posts: 601
Rep Power: 14 
What do you mean by first or second order motion. I've never heard this terms. What kind of morphing case are you referring. I analyzed three cases. A case without fluid forces, a forced oscillation and a Viv.
If someone needs 15 minutes just to understand your question you will likely not get any answer. 

February 6, 2022, 15:43 

#11  
New Member
zink
Join Date: Oct 2015
Posts: 28
Rep Power: 8 
Quote:
Linear Spring has equation F=Kx +c While for VIV, the motion equation is Mx¨ + Cx˙ + Kx = Fx Can you clarify how come linear spring constraint can solve the VIV equation? 

February 6, 2022, 15:53 

#12 
Senior Member
Michael Alletto
Join Date: Jun 2018
Location: Bremen
Posts: 601
Rep Power: 14 
The solid body motion solver solves the rigid body motion equation. The spring force and the fluid forces act on the body leading to the motion of the cylinder. This is written in the paper as well. Please read the paper carefully.


February 9, 2022, 06:16 

#13 
New Member
zink
Join Date: Oct 2015
Posts: 28
Rep Power: 8 
Thanks for clarifying. I was confused with the file you shared on your Github and the openfoam journal. Nice work and well documented.


February 14, 2022, 15:33 

#14 
New Member
Marcelo Italo Adotti
Join Date: Jul 2011
Posts: 7
Rep Power: 13 
Hi Michael,
Nice job! I am currently working with OF overset and some validation tests. I have a doubt about the "inline oscillating cylinder" condition, the motion applied in your paper is on the xaxis, co linear to the incident flow. Any reason why it is not with respect to the "z" axis?. Regards 

February 15, 2022, 00:57 

#15 
Senior Member
Michael Alletto
Join Date: Jun 2018
Location: Bremen
Posts: 601
Rep Power: 14 
Hello Marcelo,
No no specific reasons. You can also enforce the motion in zdirection or do a combination of the two 

February 15, 2022, 01:15 

#16 
New Member
zink
Join Date: Oct 2015
Posts: 28
Rep Power: 8 
Hi Alletto,
can you recommend me reading material that explains how the 6 DOF motion is solved using dynamic mesh. I wish to learn the mathematical aspect. Many thanks in advance. Regards. Ansab 

February 15, 2022, 04:52 

#17 
Senior Member
Michael Alletto
Join Date: Jun 2018
Location: Bremen
Posts: 601
Rep Power: 14 
Regarding the translation motion, the equation are in the paper. Regarding the rotation, I started but never ended a tutorial where I summerized the equation solved. Find the pdf attached.


February 16, 2022, 14:34 

#18 
Senior Member
Michael Alletto
Join Date: Jun 2018
Location: Bremen
Posts: 601
Rep Power: 14 
Is someone is interested the case files of the paper can be found here https://gitlab.com/mAlletto/openfoam...ases4PaperSend


February 18, 2022, 15:21 

#19 
New Member
Alvi Ahmmed
Join Date: Nov 2020
Posts: 7
Rep Power: 3 
Can you please explain how can I calculate spring constant(k) for a square cylinder having Re=100 , U*=6 with the equations ? Kind regards.


February 18, 2022, 16:10 

#20 
Senior Member
Michael Alletto
Join Date: Jun 2018
Location: Bremen
Posts: 601
Rep Power: 14 
U* = Uoo/(fn D) , fn = ω/(2 π) = (k / m)0.5/(2 π)
So if you want to achieve a given U* for a given diameter of the cylinder with a given mass and inflow velocity you have to set k accordingly. Hope this helps 

Thread Tools  Search this Thread 
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
objects oscillate at high speed causing collapse in Dynamic mesh and overset mesh ca  insane  OpenFOAM Running, Solving & CFD  0  May 17, 2020 22:35 
[snappyHexMesh] SnappyHexMesh/splitMeshRegion : region1 in zone "1"  GuiMagyar  OpenFOAM Meshing & Mesh Conversion  1  September 10, 2019 09:59 
Overset: refine background mesh without resetting the mesh  JohnAB  STARCCM+  6  May 19, 2014 13:48 
Help in setting up an overset mesh  RKE  STARCCM+  1  February 21, 2014 21:35 
Icemcfd 11: Loss of mesh from surface mesh option?  Joe  CFX  2  March 26, 2007 18:10 