[saihasil]

hi
I am simulating a centrifugal blower in ansys fluent.
Boundary conditions are
Inlet Condition=Pressure Inlet =0 Pa
Outlet Condition=Mass flow rate(outward normals)=5kg/s
RPM of rotor = 3000rpm
Operating Condition =101325 Pa
steady
Pressure based solver
k-epi, standard,

I have only one doubt i.e why negative pressure created inside the domain? or What is that negative pressure?

for clarification i attached the image..please look give some ideas....

thanks

Regards
Rajesh H

[LuckyTran]

You need a lower pressure than 0 Pa in order to suck the air into the fan. You only get a static pressure rise after the fan.

There's also centrifugal effects which necessitates a lower pressure in the center of rotation.


If you don't like negative numbers then set the inlet to a big positive pressure.

[saihasil]

what is that negative pressure indicate ? Is that a suction pressure so it is having negative value?

[Time4Tea]

I think what it means is probably that the physical model you are using is invalid.


The air can't actually be at a negative pressure of -6.8 x 10^5 Pa. Because, as the static pressure reduces, the continuum assumption will break down. I.e. you will reach a point where you have a partial vacuum and the assumption of continuous pressure throughout the flow domain (which the CFD method relies on) will be invalid.



What this means in reality is that you wouldn't actually be able to achieve that specified mass flow rate with an inlet at atmospheric pressure. The only way that you could achieve it would be to have a higher inlet pressure, as LuckyTran said (i.e. compress the air first to increase its density).


Also, it seems strange to me that your outlet pressure is so similar to your inlet pressure, for this type of device. It is a compressor, so I would expect the outlet pressure to be significantly higher. Are you using a compressible model (ideal gas)?

[LuckyTran]

Yes it is a suction pressure in the 1D sense. And no it's not a suction pressure because of centrifugal effects. If you have flow going in a circle, the pressure at the outer radius is higher than the inner radius (due to centrifugal forces). You are seeing both effects.

Pressures in Fluent are gauge pressures. You set an operating pressure of 101325 Pa and the inlet gauge pressure is 101325 Pa. Negative pressure means that it's less than 101325 Pa. That's all.

[Time4Tea]

Originally Posted by LuckyTran

Pressures in Fluent are gauge pressures. You set an operating pressure of 101325 Pa and the inlet gauge pressure is 101325 Pa. Negative pressure means that it's less than 101325 Pa. That's all.

He's getting -680000 Pa. That's not just negative gauge, that's negative absolute pressure, which I believe is physically impossible (in a gas).


Something seems very wrong here, because the inlet pressure should be showing 0 Pa (from the boundary condition) and the outlet pressure should be higher.

[saihasil]

I have seen some journal papers , they mentioned negative pressure as suction pressure...

1) Numerical Analysis of Centrifugal Blower Using CFD,V. Krishna, K. Naresh Kumar, International Journal of Engineering Research & Technology (IJERT), ISSN: 2278-0181

2) Numerical Study of Flow in a Centrifugal Blower, 1ISHAN VARMA,RANJITH MANIYERI, WSEAS TRANSACTIONS on FLUID MECHANICS

[saihasil]

For example i have value of static pressure of -680000 Pa. negative sign indicates the "suction" and 680000 Pa represents the static pressure value. Is that correct?

[Time4Tea]

I think the second paper you referenced (Varma, Maniyeri) is better than the first one. It contains a lot of nice details of the model setup; whereas in the first one, it doesn't seem very clear what the boundary conditions are.


Looking at the second paper and comparing it to your plots, I think there are several things worth noting:

• Yes, they are also seeing a negative pressure at the inlet (Figure 10). However, the lowest theirs goes is -7400 Pa. Yours seemed to be showing -680,000 Pa, which is a huge difference. Theirs is still positive absolute pressure, which is possible in reality - yours is not.
• You can see from the same pressure plot that they are getting a significant increase in pressure between the inlet and the inside of the casing past the impeller, which is what you would expect.
• The colour of their pressure plot looks correct for the outlet being at 0 psi (gauge), which is what they set as the boundary condition. On your plot, it didn't look like that was the case.

So, if it was me, I would be investigating why your pressure plot looks so different to theirs and why your minimum pressure is so much lower.

[LuckyTran]

You don't need to look up a paper.... negative pressures are negative pressures and you can call this due to suction effect. That's obvious.

Originally Posted by Time4Tea

He's getting -680000 Pa. That's not just negative gauge, that's negative absolute pressure, which I believe is physically impossible (in a gas).

Something seems very wrong here, because the inlet pressure should be showing 0 Pa (from the boundary condition) and the outlet pressure should be higher.

By default Fluent limits the the absolute static pressure to 1 Pa so you can't get a negative absolute pressure within Fluent. So I'm not worried that non-physical pressures are obtained because Fluent won't spit out such numbers. Obviously the plot is just plain wrong if that is the case. Is the plot made in CFD post or something? Why is there a "pressure" of -1.6 MPa? Or maybe.... we are observing pressure in the moving reference frame and not looking at pressure a stationary frame?

The inlet pressure BC is a stagnation pressure and the static pressure can and will be lower if there is any flow, which there is.

[Time4Tea]

Originally Posted by LuckyTran

By default Fluent limits the the absolute static pressure or 1 Pa so you can't get a negative absolute pressure without Fluent.

Ok, I wasn't aware of that. I agree with you then, that the OP still needs to figure out what is wrong with their pressure plot.

The inlet pressure BC is a stagnation pressure and the static pressure can and will be lower if there is any flow, which there is.

Perhaps it might be better to instead assign a static pressure of 0 (gauge) at the outlet? I think that is what they did in the paper, and it would provide a better control on the static pressure in the domain.

[sbaffini]

Doing this exercise outside the original solver that produced the numbers without being sure of what you're doing is quite futile. More than ever with a fancy volume rendering feature that might do anything to your data.

First of all, I suggest to go back to Fluent and check what real min and max are. As already stated, Fluent won't allow you to reach absolute pressures below the chosen minimum, so there is a clear discrepancy here.

Besides this, absolute pressure values for incompressible flows (that's what they mention in the second paper) means nothing. You could set up the reference pressure at 1 billion and see no difference in your results (if no actual hard limit is effectively hit).

However, truth is, if there are just one inlet and one outlet, the typical approach would be to assign pressure at outlet and mass flow rate at inlet. Indeed, and you should have been aware of this, a pressure inlet condition actually specifies the TOTAL pressure, which is the sum of static and dynamic pressures. Thus, you are probably overspecifying your problem setup. The negative pressure you see is there to counterbalance the dynamic pressure that tries to follow the fixed mass flow rate. Long story short, nowhere in the domain your static pressure seems to be fixed.