[san.pirate]

Does average static pressure=0 at (outlet) signify Neumann boundary conditions.

If I require Streamwise gradient=0 at the outlet, will the above condition suffice it?

Thanks

Santhosh

[ghorrocks]

This is all discussed in the documentation.

But from memory it defines dirichlet BCs on pressure and neumann BCs on velocity and scalars. So yes, a static pressure BC will give you zero streamwise gradient in other variables providing the boundary is perpendicular to the flow.

[san.pirate]

Yes,

The Flow is perpendicular, and when the Static pressure is Zero, The only pressure acting is the Dynamic pressure, So, By Bernoulli's equation, the dynamic pressure provided by the velocity of the fluid is constant or the flow is fully developed.

Thank you for clearing my doubt.
San

[san.pirate]

Hello,

I am having another doubt in this regard, even if the total pressure in the system is equal to the dynamic pressure provided by the velocity need not be constant.

So how is it possible to apply the boundary condition of zero velocity gradient at the outlet.

[alac1407]

Hi everyone.
I have a problem, I want to set outlet's pressure is 0.8 atm. but I cannot distinguish between "Static pressure and Average Static pressure"
Please give me an advice.
Thank you very much !

[ghorrocks]

Static pressure means the static pressure at all locations is set to the defined value. Average static pressure means that the area average of static pressure over the patch is the defined value.

[Far]

What is the physical significance of both options. When to use static pressure (fixed every where) and when to use the Average static pressure(floating with average constrained)?

In my opinion static pressure condition may be the good for the boundary far away from the object to be studied, in other words flow is more or less uniform at outlet boundary.
On the other hand the average static pressure may be more physically correct as it allows the conditions to vary along the boundary. This can be the situation in nozzle (may be diffuser, compressor, turbine, combustion chamber) where the velocity is higher in centre and low in the wall region. In other words pressure may change along the boundary so as to capture the true picture of flow form the interior and at the same it also forces the mass flow to be some average value (also changes from cell to cell) corresponding to average value of static pressure on whole boundary.

Sounds logical?

[ghorrocks]

Yes, you are correct. But also consider the numerical implications - specified static pressure is more numerically stable than average static pressure. For some simulations this may be important.

[Far]

1. Take an example of compressor, we knew from theory that the static pressure increases from hub to casing in response to centrifugal force, so if we restrict the outlet condition to single static pressure condition everywhere, wouldn't it be non-physical. It is also implied that the mass flow rate is also same every where on the outlet plane.

2. More over I have simulated NASA rotor 37 with both conditions and found that the results are overlapping (performance (pressure, efficiency) curve overlapping) , but It is also important to mention they don't have same mass flow rate or location on map for the same pressure ratio or efficiency. This result is also surprising for me that if average pressure in both cases is same then why mass flow rate is different.

3. It is also hard to think how can be static pressure be same throughout the outlet plane, if outlet plane is not far away from the wall.

4. Also I have statistics from the CFX 13 tutorial guide that shows this:
Among 32 tutorials there 14 are cases which use the average static pressure and static pressure as boundary condition at outlet (and mostly with zero pressure)

1. eight cases used average static pressure
2. six cases used static pressure

(it is not mentioned in tutorials why static pressure or average static pressure condition is chosen)
What this shows?

5.
There is also one note in tut of coal combustion; reads as :
Average pressure boundary condition leaves pressure profile unspecified while constraining the average pressure to the specified value. In some situation, leaving the profile fully unspecified is too weak and covergence diffculties (not talking about accuracy!!!) may result. The pressure profile blend feature woks around this by blending between a unspeficied pressure profile and fully specified pressure profile (so this is not completly loose condition!!!). By default, the pressure profile blend is 5%. For swillring flow, however, any amount a uniform pressure profile is inconsistent with the radial pressure profile which should naturally develope in response to the fluid rotation and therefore pressure profile blend must be set to zero.
This statement is also supporting the use of average static pressure either with specified blend or with zero blend.

[Far]

Originally Posted by ghorrocks

Yes, you are correct. But also consider the numerical implications - specified static pressure is more numerically stable than average static pressure. For some simulations this may be important.

I am still not able to decide whether I should use average static pressure or static pressure.

[ghorrocks]

1. Take an example of compressor, we knew from theory that the static pressure increases from hub to casing in response to centrifugal force, so if we restrict the outlet condition to single static pressure condition everywhere, wouldn't it be non-physical. It is also implied that the mass flow rate is also same every where on the outlet plane.

This is an application where the averaged condition makes sense. Also: it does not imply the MFR is the same on the outlet plane. The velocity can vary.

This result is also surprising for me that if average pressure in both cases is same then why mass flow rate is different.

Then the boundary condition is having a subtle difference on the flow. Have a look around the outlet and try to find the difference.

3. It is also hard to think how can be static pressure be same throughout the outlet plane, if outlet plane is not far away from the wall.

It will be the same if you tell it to be the same.

The pressure is essentially constant normal to the flow through a boundary layer. That is why this BC is commonly used.

The tutorials teach you which buttons to press. They do not teach you CFD. You need a knowledge of CFD to decide whether averaging is what you want or not.

I am still not able to decide whether I should use average static pressure or static pressure.

Then do a sensitivity analysis on both options and you will find out, without a doubt, which option is best. And you will probably learn a wholy heap about the importance of outlet BCs in CFD. I reckon that is worth the effort - that is how I learnt this stuff, you try it and see what it does.

[Far]

I have similar results, didn't find in any difference in compressor map. 100% overlap

[Far]

The tutorials teach you which buttons to press. They do not teach you CFD. You need a knowledge of CFD to decide whether averaging is what you want or not.

I know that the static pressure is constant on the outlet plane in actual rotating test rigs, but outlet is far far away from the compressor as compared to CFD. Also they have diffuser and other equipment to make flow uniform with lower kinetic energy at outlet.

Should I put a large tank/reservoir on the outlet and see what type of flow I get on outlet, whether varying static pressure or uniform static pressure?

[ghorrocks]

If the experiment has downstream equipment to smooth the flow out then you should model it, one bit at a time, until your results start behaving themselves.

[Far]

Reference : J. Turbomach Volume 128 Jan 2006 page 126 NA Cumpsty and J.H. Horlock

If static pressure conditions are to be used at exit (total to static performance measurement) they should be defined where the static pressure is uniform, as in a large plenum or in the open atmosphere. remedy to this problem is that move the outlet where static pressure is uniform, otherwise this approach will introduced error.

Non uniform static pressure is clear evidence that the flow is being turned or accelerated and such regions do not lend themselves to providing a good reference condition downstream of a turbine or compressor.

While this may be true at the diffuser exit, whose function is reduce the velocity to level (mach 0.1) acceptable to combustion chamber with minimum distortion at exit plane

Now we can say if the flow is not uniform then why should we say it so to code?

[D.B]

Hi,
What I had seen in my past work was that flow eventually mixes out after a certain distance behind the rotor. So as per Cumpsty ( your reference paper) static pressure boundary condition should be stated there. Now as far as specifying average static pressure with pressure profile blend factor goes, I would say it would be farther from reality as compared to specifying only static pressure.

Also what I have seen from various experiments and literature was that the pressure profile evens out to a great extent 2.5-3 chords downstream of the rotor. So specifying pressure ( or average static pressure ) at this location would not make a huge difference and should be more appropriate.

-D.B

[Far]

Hi,
What I had seen in my past work was that flow eventually mixes out after a certain distance behind the rotor. So as per Cumpsty ( your reference paper) static pressure boundary condition should be stated there. Now as far as specifying average static pressure with pressure profile blend factor goes, I would say it would be farther from reality as compared to specifying only static pressure.

What shall be difference if we only define average static pressure with zero blend in contrast to average static pressure with specified blend of 5%

[D.B]

Originally Posted by Far

What shall be difference if we only define average static pressure with zero blend in contrast to average static pressure with specified blend of 5%

At a far away boundary where the profile actually is evened out, I would say the difference shouldn't be much.. at a nearer boundary where there actually is a variation in the pressure distribution at the exit plane I would say obviously average static pressure with blend factor is better. But let me point out that it would be far from reality, because of presence of swirling features like passage vortices and tip flow vortices and other non uniformitiesin the flow. If you want a better solution and have experimental data with you, I would say at a nearer boundary a mass flow outlet condition is better suited than the pressure boundary conditions.

-D.B.

[Far]

so you are ruling out the uniform static pressure boundary condition if it is placed too close?

And from your discussion I also infer that average pressure static with zero blend factor should work good regardless of location of outlet boundary?

Mass flow is not a option !!! because some times we have to work with new designs and it is very hard to make map near choke conditions.

[D.B]

Originally Posted by Far

so you are ruling out the uniform static pressure boundary condition if it is placed too close?

Definitely, since it is very far from reality.

Originally Posted by Far

And from your discussion I also infer that average pressure static with zero blend factor should work good regardless of location of outlet boundary?

Definitely not. What I am saying is that it is better that in condition far away from the rotor, it would work fiine but I am not sure about the zero blend factor part.

Originally Posted by Far

Mass flow is not a option !!! because some times we have to work with new designs and it is very hard to make map near choke conditions.

Yes, I know, this is for cases when you have experimental data and need to study some flow physics characteristic.

-D.B