# Meaning of turbulence terms when flow direction is reversed at mass flow inlet

 Register Blogs Members List Search Today's Posts Mark Forums Read

 October 27, 2014, 12:55 Meaning of turbulence terms when flow direction is reversed at mass flow inlet #1 New Member   zack Join Date: Oct 2014 Posts: 13 Rep Power: 11 Background: It is known that we sometimes change mass flow inlet and velocity inlets to outlets by simply reversing velocity direction in FLUENT. However I have a question.In case of velocity inlet the turbulence kinetic energy and dissipation % are specified for outside air coming into system. And at pressure outlet the same two terms are only used when backflow occurs. Question: Then if I reverse mass flow inlet velocity direction such that it is now AN OUTLET what do the turbulence terms signify the terms only for use in backflow or infact it is rigidly specifying a boundary condition value of k and e for k-e equations inside the system at the boundary?

November 5, 2014, 22:15
#2
Senior Member

Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,735
Rep Power: 66
Quote:
 Originally Posted by zackufairu infact it is rigidly specifying a boundary condition value of k and e for k-e equations inside the system at the boundary?
That is indeed the case. You can use a velocity inlet with any specified parameters in order to impose those conditions at the inlet. Any property you can specify will be imposed, (temperature, k, epsilon, intensity, length-scale, volume fraction, etc).

This is not a bug, since that is how velocity inlets work. And this property of velocity inlets (but used as an outlet) is can be exploited in order to impose "outlet boundary conditions." It's has niche applications for situations in which the outlet boundary condition is known from say a velocity traverse experiment but only pressure at the inlet is known.

November 6, 2014, 12:45
#3
New Member

zack
Join Date: Oct 2014
Posts: 13
Rep Power: 11
Quote:
 Originally Posted by LuckyTran That is indeed the case. You can use a velocity inlet with any specified parameters in order to impose those conditions at the inlet. Any property you can specify will be imposed, (temperature, k, epsilon, intensity, length-scale, volume fraction, etc).
But what about temprature and turbulence specifications for other boundary conditions(eg pressure inlet/outlet)
Do they also fix the values (temprature,k,e)? Because I read somewhere that the value of temprature is subject to change at a pressure boundary condition as per the solution in the system

 November 6, 2014, 13:59 #4 Senior Member   Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,735 Rep Power: 66 At a true outlet boundary, there are many free variables and those variables are subject to change. For those boundaries you specify parameters that are used when there is reversed flow (so-called backflow properties). If the flow is outward as intended, then the temperature, k, epsilon at the outlet boundary is computed according to the interpolation and discretization scheme. But at an inlet, those variables are imposed, they are the boundary conditions. There is no "option" to not impose them. For a velocity inlet, I think pressure is the only variable that might be exempt, and only for the subsonic pressure-based solver case. Otherwise, at a velocity inlet, or any type of inlet, you impose many boundary conditions (temperature, k, epsilon, etc) and all those imposed boundary conditions are also imposed at the pseudo-outlet condition. Some remarks that may or may not apply depending on what you intend to do with a velocity inlet at outlet: For transported properties like temperature, the only means for a imposed temperature outlet condition to propagate back upstream is via diffusion. If the Peclet number is large, the influence of the upstream diffusion may be considered negligible. In this case you will be able to still get a converged solution and fairly accurate since the influence of the non-sense boundary condition does not affect the problem much. On the other hand, if you want to determine the temperature at the outlet of your domain using this method, it will not work since you specified the temperature at the outlet. zackufairu likes this.

November 6, 2014, 15:19
#5
New Member

zack
Join Date: Oct 2014
Posts: 13
Rep Power: 11
Quote:
 Originally Posted by LuckyTran For a velocity inlet, I think pressure is the only variable that might be exempt, and only for the subsonic pressure-based solver case. Otherwise, at a velocity inlet, or any type of inlet, you impose many boundary conditions (temperature, k, epsilon, etc) and all those imposed boundary conditions are also imposed at the pseudo-outlet condition. .
I hope I am catching on if by pseudo outlet you mean only velocity inlet tuned outlet and not other outlet boundary conditions.?
Quote:
 Originally Posted by LuckyTran For transported properties like temperature, the only means for a imposed temperature outlet condition to propagate back upstream is via diffusion. If the Peclet number is large, the influence of the upstream diffusion may be considered negligible. In this case you will be able to still get a converged solution and fairly accurate since the influence of the non-sense boundary condition does not affect the problem much. On the other hand, if you want to determine the temperature at the outlet of your domain using this method, it will not work since you specified the temperature at the outlet
So you mean I have to be extra careful in imposing a velocity outlet making sure the outlet temprature is not really my concern and the flow from upstream is very convective to prevent "diffusion"of the imposed temprature and turbulence at velocity outlet?

just correct my statements if you think i understood them wrong.
I really appreciate the help here,cfd is so exciting when explained right.

 November 6, 2014, 15:30 #6 Senior Member   Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,735 Rep Power: 66 You are correct on both points. zackufairu likes this.

November 6, 2014, 15:35
#7
New Member

zack
Join Date: Oct 2014
Posts: 13
Rep Power: 11
Quote:
 Originally Posted by LuckyTran For transported properties like temperature, the only means for a imposed temperature outlet condition to propagate back upstream is via diffusion. If the Peclet number is large, the influence of the upstream diffusion may be considered negligible. In this case you will be able to still get a converged solution and fairly accurate since the influence of the non-sense boundary condition does not affect the problem much.
I must say that is by far the best physical explanation I could receive which was absolutely non mathematical yet so intuitive.
I have learnt something today.
Thank you

 Tags dissipation, k-e model, mass flow inlet