
[Sponsors] 
August 31, 2012, 09:26 
Momentum Source (loss coefficient) and velocity decrease

#1 
Member
Join Date: Apr 2010
Posts: 61
Rep Power: 7 
Hello,
I'm trying to simulate the pressure drop of a flow obstacle with a momentum source applying a directional loss model, so the term applied results: S = −K_loss * (ρ/2) * U * U_z I'm wondering why the velocity decreases when I apply it (see figure), I can't see what's happening theoretically. Even in a real application, the obstacle reduces the flow area and so velocity is higher. I'm obtaining reasonable results of the pressure drop but the velocity is not what I expect (shown in the figure attached). Thank you very much for your hints. P.S: I post it here because I guess it's related with general CFD theory not with the particular code, by the way ANSYS CFX 

August 31, 2012, 09:51 

#2 
Senior Member
Jonas T. Holdeman, Jr.
Join Date: Mar 2009
Location: Knoxville, Tennessee
Posts: 101
Rep Power: 9 
I am interested in the idea, but don't understand the model. Could you explain the geometry and context please? Compressible/incompressible? Wouldn't a momentum source (force) be expected to change the velocity or somehow balance dissipative forces?


August 31, 2012, 10:20 

#3 
Member
Join Date: Apr 2010
Posts: 61
Rep Power: 7 
Thanks for your reply Jonas Holdeman.
I'm modelling an incompressible flow along a channel so there's some kind of flow straightener inside (I can't provide publicly exact characteristic of the device). I'm trying to model the pressure drop of this device without modelling it with that loss model. It have sense that applying a negative force in the flow affects the velocity in that way, but maybe you can provide some ideas. Regards. 

August 31, 2012, 18:04 

#4 
Senior Member
Jonas T. Holdeman, Jr.
Join Date: Mar 2009
Location: Knoxville, Tennessee
Posts: 101
Rep Power: 9 
I still don't have enough information about your problem so I will make some assumptions. The width or direction of the channel must be changing and you want to quickly restore some flow profile by inserting this "flow straightner". You want to model the effect of the physical straightner with some equivalent ficticous force or momentum source. Certainly the flow must be constant along the channel (average velocity inversely proportional to area) because of conservation of mass. Upstream and downstream of the obstruction the pressure will be decreasing in the direction of the flow, but there will be an additional decrease around the obstruction. I don't understand what velocity you are plotting in your figure.
With regard to your model, I think the functional form should be divergencefree (solenoidal) as it is a source term added to the divergencefree velocity, as can be seen from a pressurefree formulation of the momentum equation such as the streamfunctionvelocity FEM. Have others used this approach of using an equivalent momentum source? I still have to think about the idea and am a little skeptical from a consistency point of view. But the idea is interesting. Last edited by Jonas Holdeman; September 1, 2012 at 08:09. Reason: correction 

September 2, 2012, 05:07 

#5 
Senior Member
Rami BenZvi
Join Date: Mar 2009
Posts: 148
Rep Power: 8 
Your idea is quite common when porous media is simulated. The specific form you use is the Darcy law in zdirection. Some models also add another term (Forschheimer). Without specifying your case, I cannot see what is the expected solution, and what is wrong in the one you received.


Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Momentum source coefficient, cylindrical coordinates, circumferential component?  polakse  CFX  15  January 18, 2016 03:40 
Difficulty in calculating angular velocity of Savonius turbine simulation  alfaruk  CFX  8  December 3, 2013 17:51 
Velocity profile disturbance due to loss coefficient  rks171  Main CFD Forum  3  May 25, 2012 17:30 
liddriven cavity in matlab using BiCGStab  Don456  Main CFD Forum  1  January 19, 2012 16:00 
Momentum source coefficient and convergence  meh  CFX  2  November 19, 2010 17:57 