3d air flow pressure driven flotran

[screetch82]

hey,
I am an intern and have to to a research for a printing press manufacturer. I have to compute the air flow
in a 3d duct. the duct ends with an small outlet which leads to a rotating cylinder. The cylinder has small chambers on the circumference. The outlet
of the duct covers 2 of the chambers. The air flow is created by sucking out the air of the small chambers through the duct.
All I know is the negative pressure which is produced by the vacuum pump. I know flotran computes an airflow using a press difference
But all I have is the pressure on one end and the volume of air on the other end? Can someone give me some hints how to do this calculation?

I have already modelled the duct and have done a few sample calc. But in order to even start the solution I set the pressure on the outlet (the vac-pump) to 0
and the pressure at the inlet (the opening where the suction ring is) to 80 kpa. So the pressure difference gave me and airflow.
The problem is that in reality the pressure drop is the other way around and I know just the outlet pressure. The inlet pressure decreases as more air is
sucked out of the chamber. Right? How do I do this? how do I work with neg pressures in flotran?
can someone give me parts of a simple input file relating to my problem? I'd be happy if some emails me..for fast communication.to screetch@gmx.net

 

[vortexman]

screetch82,

It doesn't matter what the pressures are at the inlet and outlet, just the difference. Flotran may not accomodate negative pressures, but except for that, any inlet and outlet pressures that give the correct pressure difference will be fine. Can you assume that the inlet pressure is always the same as the pressure in the room? This would be most convenient, and is probably a fair assumption as long as the air doesn't get drawn through a small orifice to get to the inlet of your duct. If this assumption is OK, then simply use the difference between the room pressure and the outlet pressure that you know, and you'll be fine. One thing to be aware of is that the pump that gives you low pressure at the outlet probably does not give you the same pressure at all flow rates. Depending on the goal of the simulation, the nominal figure that you have may be close enough. If you are trying to predict a very accurate flow rate, for example, then you need accurate inlet and outlet pressures, and may have a tough time anyway. If you are looking for the overall flow pattern, to find dead zones, for example, then the exact pressure difference wouldn't matter.

Good luck,

vortexman

 

[screetch82]

thanks alot first of all
do you use flotran?
since the fluid is air and the pressure difference is 0.7 bar or 70 kPa would I need to do a compressible analysis? The average pipe diameter is about 14 mm (it varies alot)
The flow rate has been meassure experimentaly and is about 6 m^3 per hour.

Can you give me a sample inputfile for a nozzle with all the important fluid property options and solution options?
Flotran never solves any of my compressible analyses. Seems like I am missing something.

and all my incompresibble turbulent analys seem to be wrong sind I get extremly high velocities.

I's really need a simple sample to compare.

 

[vortexman]

screetch82,

No, I have not used Flotran, so I can't help you with a sample input. If you know the volume flow rate from measurement, then you are probably better off specifying that volume flow rate at the inlet, and using some kind of pressure boundary condition at the outlet. I don't know exactly how Flotran works, but there may be an option to specify volume flow rate directly at the inlet, or else you could specify a constant velocity at the inlet, calculated to give the correct volume flow rate.

The problem with pressure boundary conditions is that you won't get the correct volume flow rate from the simulation unless you have high enough resolution to correctly account for the resistance to flow. This would require a very fine mesh, and properly chosen turbulence model, etc. I assume that this is why you are seeing velocities that are too high. If you specify the volume flow rate, most of the features of the simulated flow can be pretty accurate, without the requirement that your model accurately predict the flow resistance. Even so, your mesh needs to be fine enough, and high enough quality, to capture the details of the flow that you are after.

What are the questions you are trying to answer using this simulation? This is what really allows you to determine what you need in your model.

vortexman

 

[screetch82]

I have a log duct mostly a lot of cylinders with different diameters. At one end there is a compressor which has a capacity of about -0.7 bar. In front of the other there is at after certain time a chamber with a volume of air. After that time point the air is sucked out. This happens periodically. Now I am trying to simulate this. The problem is that the volume flow rate varies (increases once the chamber of air is in front of the opening).
Any suggestions?

 

[vortexman]

screetch82,

I don't completely understand the physical arrangement, but I understand the difficulty with time varying flow rate. The best way to handle this problem depends quite a bit on Flotran features. I think that the most natural way to deal with this is to use some kind of time varying boundary conditions. I don't know whether Flotran allows this, or how to do it if it is possible. Since it is difficult to get the flow rate correct by specifying pressure difference, I would probably impose a time varying velocity boundary condition on one end. If the time variation is caused by an inlet or outlet opening and closing, while a compressor is working at the other end, then you might indeed need to include compressibility effects.

vortexman