
[Sponsors] 
August 4, 2014, 10:33 
non reflecting boundary conditions

#1 
New Member
rabea matouk
Join Date: Aug 2014
Posts: 7
Rep Power: 8 
Hello,
I have some hydrodynamic waves reflections on the outlet of computational domain using outlet pressure boundary condition so I decided to implement the non reflecting boundary conditions in our inhouse code which is incompressible navierstockes finite elements/spectral code because we will use the aerodynamic sources to compute the radiated noise in the aeroacoustics study so it is so sensitive to have accurate aerodynamics results near the outlet. I'd like your help to start and to provide me with references please. Thanks in advance 

August 7, 2014, 18:32 

#2 
Senior Member
Join Date: Aug 2011
Posts: 271
Rep Power: 12 
For incompressible flows you can use Orlanski boundary condition.


August 7, 2014, 20:12 
non reflective BC for incompressible flows

#3 
New Member
rabea matouk
Join Date: Aug 2014
Posts: 7
Rep Power: 8 
Thank you leflix for your advice and help.
You mean the convective BC: dU/dt + Un*dU/dn = 0 could you please give me more details and advice , this is my first try to implement BC? U is the longitudinal and vertical velocities or I have to apply it for the three velocity components? 

August 16, 2014, 15:59 

#4 
Senior Member
Join Date: Aug 2011
Posts: 271
Rep Power: 12 
yes indeed it is this boundary condition.
You need to use this boundary condition for the three velocity components indeed. dU/dt + Un*dU/dn = 0 dV/dt + Un*dV/dn = 0 dW/dt + Un*dW/dn = 0 the choice of the convective velocity Un is tricky check the paper of Orlansky, many choices are possible. 

August 17, 2014, 12:36 

#5 
New Member
rabea matouk
Join Date: Aug 2014
Posts: 7
Rep Power: 8 
Thank you very much leflix. I'll try to implement and test it. the convective velocity Un is really tricky but I'll try several suggestions and see the results.


August 19, 2014, 02:28 

#6 
New Member
Parth Thaker
Join Date: Mar 2014
Location: Surat , Gujarat , INDIA
Posts: 26
Rep Power: 9 
thnx lefflix


August 19, 2014, 03:09 

#7 
New Member
Parth Thaker
Join Date: Mar 2014
Location: Surat , Gujarat , INDIA
Posts: 26
Rep Power: 9 
for almost constant velocity flow at outlet can i use C = U , u = 1.0; , remains almost constant at outlet section , , , thnx in advance


August 19, 2014, 06:40 

#8 
New Member
rabea matouk
Join Date: Aug 2014
Posts: 7
Rep Power: 8 
I think we can use Un = U (free stream) for the u equation but what about v and w since the phase velocity is different.


January 16, 2019, 09:42 

#9  
Senior Member
Jianrui Zeng
Join Date: May 2018
Location: China
Posts: 157
Rep Power: 4 
Quote:
Do you have some experience in the implement of the convective boundary condition in openfoam? I am quite new in this area. Thank you. 

January 16, 2019, 10:37 

#10  
Senior Member
Lucky Tran
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 4,054
Rep Power: 49 
Quote:
The openfoam code is freely available on the internet. Do you have a specific question about the way it is implemented? 

January 16, 2019, 10:45 

#11 
Senior Member
Jianrui Zeng
Join Date: May 2018
Location: China
Posts: 157
Rep Power: 4 
Which BC code should be used to modify so that I can apply convective boundary BC in my simulation? Or are there any guide to introduce this boundary? I am new to this BC. Thank you.


January 16, 2019, 11:43 

#12  
Senior Member
Lucky Tran
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 4,054
Rep Power: 49 
Quote:
For incompressible flows the advective BC should already be sufficient. For compressible flows additional treatment is needed to prevent unwanted reflection of acoustic waves. E.g. Pointsot & Lelef. Because of this, for incompressible flows you hear advective or outflow BC (and you don't normally call this nonreflecting, although it does prevent reflection of hydrodynamic waves). Nonreflecting BC usually implies acoustics. 

January 17, 2019, 22:04 

#13  
Senior Member
Jianrui Zeng
Join Date: May 2018
Location: China
Posts: 157
Rep Power: 4 
Quote:
In my case, the governing quations exclude acoustic interactions and compressibility effects from the full compressible Navier–Stokes equations. It is called LowMach Number equation. I think it is still a compressible flow but a littile special. So I am confused if additional treatment of avective BC is needed to prevent unwanted reflection of acoustic waves? If so, which aspect should I modify? I will read the paper you gave later. Thank you. 

January 18, 2019, 00:55 

#14 
Senior Member
Lucky Tran
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 4,054
Rep Power: 49 
Since I can't see the actual governing equation, I can't say. If it truly remove acoustic waves then sure, then probably/maybe regular advective BC should work. But I don't know what types of Mach waves might be allowable in this model and if there are any, how the reflections need to be handled.


January 18, 2019, 04:12 

#15  
Senior Member
Jianrui Zeng
Join Date: May 2018
Location: China
Posts: 157
Rep Power: 4 
Quote:
∂(ρh)/∂t + ∂(ρUh)/∂x = ∂(k∂T/∂x)/∂x I am also confused about LowMach assumption. It is said that pressure fluctuations can be eliminated due to it. But I am still not sure what influences can the assumption give. Thank you. 

January 18, 2019, 04:22 

#16  
Senior Member
Filippo Maria Denaro
Join Date: Jul 2010
Posts: 5,620
Rep Power: 60 
Quote:
I suppose that you are considering a Taylor series for the pressure and density field around a base state that satisfies the gas law. This way you get an elliptic field for the pressure fluctuations having as source also the density fluctuations. The pressure BCs. are of Neumann type and you can use convective conditions at an outflow 

January 18, 2019, 08:50 

#17  
Senior Member
Jianrui Zeng
Join Date: May 2018
Location: China
Posts: 157
Rep Power: 4 
Quote:


January 18, 2019, 09:02 

#18  
Senior Member
Filippo Maria Denaro
Join Date: Jul 2010
Posts: 5,620
Rep Power: 60 
Quote:
If you have the elliptic equation for the pressure, you have to set the proper Neumann BCs. that follow from the convective condition. That is if you set something like dU/dn = 0 the normal pressure gradient must be congruently prescribed. 

January 19, 2019, 12:17 

#20  
Senior Member
Lucky Tran
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 4,054
Rep Power: 49 
Quote:
A hydrodynamic wave (i.e. vortex shedding) with a velocity (and pressure) fluctuation would be reflected backwards if you simply apply the du/dn = 0 constraint or pressure = some number. However, the advective/convective/outflow BC is so widely implemented in most codes that we don't think of it as anything special. So it is common misbelief that nonreflecting is only for compressible flows. 

Thread Tools  Search this Thread 
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
mesh file for flow over a circular cylinder  Ardalan  Main CFD Forum  7  December 15, 2020 14:06 
Domain Imbalance  HMR  CFX  5  October 10, 2016 06:57 
Low Mixing time Problem  Mavier  CFX  5  April 29, 2013 01:00 
[Commercial meshers] Trimmed cell and embedded refinement mesh conversion issues  michele  OpenFOAM Meshing & Mesh Conversion  2  July 15, 2005 05:15 
A problem about setting boundary conditions  lyang  Main CFD Forum  0  September 19, 1999 19:29 