
[Sponsors] 
November 1, 2010, 20:47 
pressure calculationjunction box

#1 
Member
hh
Join Date: Sep 2009
Posts: 38
Rep Power: 9 
hello
in journal bearing, half sommerfeld condition must be applied. i.e we must set the negative pressure to zero. i read a little of junction box routines help but i don t know what should i do? how we can call pressure on all cell and return it? 

November 1, 2010, 21:37 

#2 
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 12,969
Rep Power: 100 
Can you explain what boundary condition you want to apply? I do not know what the Sommerfield condition is.


November 2, 2010, 03:02 

#3 
Member
hh
Join Date: Sep 2009
Posts: 38
Rep Power: 9 
in the journal bearing pressure is positive in converging portion and negative where it diverges.but Such pressures are rarely encountered in real bearings.
Mineral oils contain between 8 and 12% dissolved air. This air will start to come out of solution whenever the pressure falls below the saturation pressure. In many situations the saturation pressure is similar to the ambient pressure surrounding the bearing, and in these cases gas liberation will maintain the pressure in the divergent clearance space at close to the ambient level.to predicted this we use half sommerfeld condition. we setting all values of negative gauge pressures to the value of the operating pressure. Thus, the gauge pressures for the top of journal are set to zero when negative values occur. it is done before with fluent udf with this macro by "K.P. Gertzos" : DEFINE_ON_DEMAND macro thread_loop_c, begin_c_loop thread_loop_f, begin_f_loop i think this condition can be applied just for cell on the faces that has negative pressure and it is not necessary to add the all cell in the volume because pressure gradient in r direction is zero. 

November 2, 2010, 06:20 

#4 
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 12,969
Rep Power: 100 
This boundary is illdefined and is not physical. Either you specify a wall and let the solver calculate the pressure for you, or you specify the pressure and let the solver calculate the flow velocity and direction. The second does not make sense in case as the walls are, well... walls. So use a wall boundary condition.
To account for the air coming out of solution you will need to put in a physical model for that. You might be able to write a custom density function or maybe use the cavitation model, adapted to dissolved air. What are you trying to get out of this simulation? 

November 2, 2010, 10:49 

#5 
Member
hh
Join Date: Sep 2009
Posts: 38
Rep Power: 9 
Thank you
These are from some journals: “Cavitation is modeled by setting all calculated negative pressure and their gradients equal to zero throughout the iteration. The iteration is repeated until the pressure satisfies the convergence criterion” “When the film pressure drops down to the atmospheric pressure due to heavy external load or high operating rotational speed, rupture of the film or cavitation occurs. Taking into account this situation, the following boundary condition is used which applies the nonsubatmospheric pressure constraint: p>=p_ambient at pi<theta<2*pi When cavitation is included, the vaporization pressure, the surface tension coefficient and the noncondensable gas mass fraction are also needed.” “As a constraint, the hydrodynamic pressure needs to be greater than the liquid cavitation pressure everywhere in the flow domain, i.e. p>=p_cav at pi<theta<2*pi In all journals that simulated “journal bearing” with cfd, this condition is applied. Cfd simulation of journal bearing is part of my thesis. It is done befor with fluent and cfxTascflow. the half Sommerfeld solution results in more realistic predictions of load components, but this simple approach leads to a violation of the continuity of mass flow at the outlet end of the pressure curve. A better boundary condition is the Reynolds boundary condition but in one journal says that Reynolds BC cannot be use in fluent. simulation based on the assumption of incompressible fluid and laminar flow in the bearing. What is cfxTascflow? can junction box routines do it(like udf)? 

November 2, 2010, 17:51 

#6 
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 12,969
Rep Power: 100 
You have to be careful about these sort of models. The ones which describe simply limiting the pressure to a certain value which are not physically correct and are never going to converge anyway. CFX has a cavitation model buitin which you should use instead.
I don't understand the film pressure boundary condition. Where is it applied? At the edge of the film wedge? CFXTascflow is an old CFD solver which CFX purchased years ago. The rotating frames of reference, and basic solver approach in CFX5+ was adapted from CFXTascflow. Tascflow has been superseded by CFX5+. 

November 2, 2010, 18:31 

#7 
Member
hh
Join Date: Sep 2009
Posts: 38
Rep Power: 9 
the right bearing has "half sommerfeld condition"
http://www.4shared.com/photo/tAGtC5/jb_online.html it applied to all cell in domain with negative pressure. negative pressure occur in divergent portion. as your suggestion i use cavitation model. it has not bad result. i don t know value of oil saturation pressure. 

May 5, 2015, 12:40 

#8 
New Member
Join Date: Oct 2014
Posts: 11
Rep Power: 4 
Dear hosseinhgf,
can you explain how did you solve the problem? txh 

Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
OpenMP in Junction Box Routine  Hannes_Kiel  CFX  10  September 21, 2010 13:51 
Pressure Rise Error  emueller  CFX  0  May 5, 2009 11:08 
Pressure calculation in NavierStokes equations  Hadian  Main CFD Forum  1  November 10, 2008 03:47 
calculation of pressure difference wrong?  Ralf Schmidt  FLUENT  0  January 5, 2006 11:06 
Gas pressure question  Dan Moskal  Main CFD Forum  0  October 24, 2002 22:02 