[Tensian]

Hi Foamers,

I am having several problems in order to compute pressure coefficient in OpenFOAM & ParaView.

Last releases of OF shows something like "pressureTools", whic can easily compute pressure coefficient, but I don´t know how to use it.

I am simulationg an stationary, incompressible and laminar flow using simpleFoam. The flow is driven by a pressure difference between inlets and outlet. Temperature is 293 K over inlets and 306 K over the wall of the pipe (modified simpleFoam in order to compute heat trasnfer between air and wall). Pressure at inlets is given (0 Pa), also at the outlet (negative pressure). Velocity is computed from pressure (pIOV).

I am trying to compute it directly using the definition of Cp, but:

1. Which is my p_ref? I guess its atmospheric pressure (101325 Pa), but I am not sure.
2. The formula says:

C_p=(p_ref-p)/0.5*rho*U²

so I am using simpleFoam for simulating flow trough a complicated pipe. Then:
a) rho=1.225 (air density at ambient temperature)
b) p in incompressible solver is p/rho, so the true values to put on the formula is p*1.225?
c)What is U²? I guess it's the freestream velocity, but I am not sure. Does this value depend on the point where we are evaluating the C_p or is it a fixed number (like for example U in Reynolds number computations)? I don´t know how to compute it in a pressure driven flow.

3. Anybody knows how to extract a value at a given point? I mean, I need to obtain velocity at point (1.5, 6.32,-1.43). How can I do that in ParaView?

Thanks in advance for sharing knowledge.

[Djub]

H Tensian,
I don't know much about OF, but I think I can help in the global understanding of your problem.

The history of pressure coefficients comes from aerodynamics: when a flow (uniform) is coming on a structure, you can define a "dynamic pressure", calculating half density times square of flow speed: 1/2 rho U² (homogeneous to a pressure). Bernoulli's equation states that the in this kind of flow, the "total pressure" which is the sum of actual pressure and dynamic pressure, is constant (with a lot of hypothesis: laminar and constant flow, uncompressible, no height potential...).
So p + 1/2 rho U² is constant.
Far upstream, U is uniform constant named Uo (or Uinf); same for P usually named Pref. The total pressure is Pref+1/2.rho.Uo²=Ptot
Let imagine a stagnation point (on the nose of an aircraft, car, chimney,...): U is null, so the pressure is the total pressure: p=Ptot=Pref+1/2.rho.Uo² .
In other parts where maybe the flow is slowed by the object, we have Ptot=Pref+1/2.rho.Uo² = p +1/2.rho.U , so p-Pref= 1/2.rho.(Uo²-U²) < 1/2.rho.Uo² . The over pressure on the object cannot be more than the dynamic pressure.

So people have invented the concept of pressure coefficient: Cp=(p-Pref)/(1/2.rho.Uo²)
Generally speaking, this coefficient is less than one; it equals one on the upstream face of the object (stagnation point), decrease to 0 when the flow is parallel to the surface, and is negative in the wake of the object. There is no negative limit, but a coefficient exceeding -3 is very rare.

In your case, you said inlet pressure is 0; so your Pref is 0.
You need to evaluate the inlet velocity to calculate your dynamic pressure: I suggest the average speed within your whole inlet. Note it should be the same than in the outlet (the main different being that in the inlet, the velocity should be uniform, not in the outlet).

I don't know if these explanations are still correct with thermal effects.

To have a extract a value at a specific point, I suggest to use the "probeLocation" utility (User manual p-94). Or, within your controlDict, use a "probes" function.

I hope all this can help.

[Tensian]

Hi Djub,

Thanks for your fast reply.

I am not sure if I am understanding. In order to obtain U_inf I should:
-compute average velocity at the inlet with the values provided by the experiment

or

-compute it analytically using some kind of formula (Bernouilli?)

I am confussed. I have imposed in my (two) inlet a value of 101325/1.225 Pa (incompressible case, so I divide pressure by density) and T=293 K and at the outlet a value of 101315/1.225 Pa. This generates a pressure difference which drives the flow. Wall is assumed to be rigid, no-slip, with a temperature of 306 K. Initial condition for p is 101315/1.225 Pa, v=0 m/s.

I ran simulation and I obtain pressure values. So, for example, if hydraulic dyameter of inlets is Dh_in=0.01 m and hydraulic dyameter for outlet is Dh_out=0.0005 m and I obtain p/rho=82710, so p=101319.75, how do you will compute U_ inf and pressure coefficient?

Thanks!

[Djub]

Hi,

For Uinf, your first item is OK (compute average velocity at the inlet).

In this case, Pref is 101325/1.225 m²/s².

you should get inspiration from the scientific literature in your field. How do other people do?

Good luck

[Lisandro Maders]

Originally Posted by Djub

Hi,

For Uinf, your first item is OK (compute average velocity at the inlet).

In this case, Pref is 101325/1.225 m²/s².

you should get inspiration from the scientific literature in your field. How do other people do?

Good luck

Hi Djub,

I didn't understand what the Pref needs necessarily to be the inlet pressure. Couldn't it be any value he wanted? Or, even better, it should be equal to the Pref of some experiment or simulation he is trying to compare, shouldn't it?

I am asking it because I am struggling with this in a centrifugal pump problem.
I am trying to compare my results with an experimental data as well as another numerical results. It is not said which was the Pref used to figure out the Cp values in the experiment, and the numerical study made some assumptions (actually, try and error) and found a value of 700 Pa. The only thing is known is that this pref should be the static pressure in the suction pipe of the centrifugal pump, however, this part of the domain is not simulated.

In this case, if I am using gauge pressure, could I use the value of my inlet, for example, or even make a try and error approach as well? Because this Pref will depend on the absolute pressure value, not only on the difference between the inlet and outlet (gauge pressure values).


I hope that is clear, if it is not, ask me then I change the way of the question. I really would like to sort it..


Cheers!

[Djub]

Hi Lisandro,
Sorry for that, but I feel unable to answer you. In your field, I don't know what pressure to take as a reference. It seems to me really a strong mistake to talk about a Cp value, without defining both the Reference pressure (atmospheric pressure, inlet or outlet, etc...), which is a kind of offset, and the so called dynamic pressure, 1/2.rho.Uo², which is a kind scaling factor.
however, all this relies on the Bernoulli 's equation and its assumptions: it doesnot work throught a fan for instance. So I am quite sure it could not be applied within a pump... Before the pump, yes; after the pump yes; crossing the pump no.
Sincerely,

[Lisandro Maders]

Originally Posted by Djub

Hi Lisandro,
Sorry for that, but I feel unable to answer you. In your field, I don't know what pressure to take as a reference. It seems to me really a strong mistake to talk about a Cp value, without defining both the Reference pressure (atmospheric pressure, inlet or outlet, etc...), which is a kind of offset, and the so called dynamic pressure, 1/2.rho.Uo², which is a kind scaling factor.
however, all this relies on the Bernoulli 's equation and its assumptions: it doesnot work throught a fan for instance. So I am quite sure it could not be applied within a pump... Before the pump, yes; after the pump yes; crossing the pump no.
Sincerely,

Yes I understand.. It makes total sense. The interesting is that I am talking about a very "famous" validation case from ERCOFTAC of a centrifugal pump with vane diffusers. There are some experimental data, including Cp.

Thanks a lot for your reply! That was very useful!

Lisandro

[goksusoydan]

Hi everybody. I have a question about pressure coefficient Cp. We are modeling the flow around a cylinder. And we want to find the Cp values around the cylinder. In the fluid flow conditions and transient analysis, we couldn't get the value of 1 in Cp coefficient. Also the pressure values are really small. There is something wrong about the density. Do you know which density does transient cfd analysis use while calculating the pressure values. The pressure values change if we model the flow with VOF.

[Lisandro Maders]

Originally Posted by goksusoydan

Hi everybody. I have a question about pressure coefficient Cp. We are modeling the flow around a cylinder. And we want to find the Cp values around the cylinder. In the fluid flow conditions and transient analysis, we couldn't get the value of 1 in Cp coefficient. Also the pressure values are really small. There is something wrong about the density. Do you know which density does transient cfd analysis use while calculating the pressure values. The pressure values change if we model the flow with VOF.

Hi goksusoydan,

First of all, which solver are you using? And which thermophysical models are you choosing?

Regards,

Lisandro

[goksusoydan]

I use ANSYS Fluent, and doing transient analysis using Shear Stress Turbulence Closure model. Do you know how Ansys-Fluent calculates the static, dynamic and total pressure in transient analysis. Because if we use volume of fluid method, we get larger pressure values than if we dont use. So ı need to know which values in the pressure equation that Ansys uses while calculating the pressure in transient analysis.

Thank you very much for your interest.

[Lisandro Maders]

Originally Posted by goksusoydan

I use ANSYS Fluent, and doing transient analysis using Shear Stress Turbulence Closure model. Do you know how Ansys-Fluent calculates the static, dynamic and total pressure in transient analysis. Because if we use volume of fluid method, we get larger pressure values than if we dont use. So ı need to know which values in the pressure equation that Ansys uses while calculating the pressure in transient analysis.

Thank you very much for your interest.

Hi,

I think you are confused about somethings. VOF does not mean Volume of Fluid. Taking directly from Fluent's manual: "The VOF model can model two or more immiscible fluids by solving a single set of momentum equations and tracking the volume fraction of each of the fluids throughout the domain. Typical applications include the prediction of jet breakup, the motion of large bubbles in a liquid, the motion of liquid after a dam break, and the steady or transient tracking of any liquid-gas interface."

Revise the setup of your simulation. If you want to simulate the flow around a sphere and is doing a transient analysis to capture the vortex fluctuations in the wake region, which I suppose you want to do, you do not need the VOF model.

Regards,

Lisandro

[goksusoydan]

Hi,

I think I couldn't explain what ı actually want to ask? I know what VOF means. My question is about pressure and presure coefficient. Our problem is modeling the flow around a burial cylinder, and we try to get pressure and pressure coefficient values around the cylinder wall. Although the water height on the cylinder is 45 cm, the maximum static pressure values that ANSYS give around the cylinder wall is approximately 13 Pa for Re=7000 for transient analysis. Also we couldn't get the value of 1 for pressure coefficient.
(Cp=(P-Po)/(0.5*water density*U^2).

Thank you very much.

Sincerely.

[Lisandro Maders]

Originally Posted by goksusoydan

Hi,

I think I couldn't explain what ı actually want to ask? I know what VOF means. My question is about pressure and presure coefficient. Our problem is modeling the flow around a burial cylinder, and we try to get pressure and pressure coefficient values around the cylinder wall. Although the water height on the cylinder is 45 cm, the maximum static pressure values that ANSYS give around the cylinder wall is approximately 13 Pa for Re=7000 for transient analysis. Also we couldn't get the value of 1 for pressure coefficient.
(Cp=(P-Po)/(0.5*water density*U^2).

Thank you very much.

Sincerely.

I misunderstood your problem. First, it is missing the Area in the Cp calculation right? I suppose you only forgot to put it, since you probably are using the monitors of Fluent to calculate it, which uses the reference values (which contains the Reference Area as input parameter).

Also, which reference pressure are you using (there is an input parameter in Fluent for this)? The fact it is a transient analysis does not change anything in pressure calculation as far as I know.

Regards,

Lisandro

[goksusoydan]

We have tried a lot of calculations. We have experimental measurements and try to analyze it numerically in 2D. The cylinder lies horizantally on the channel bed. In the reference values part we choose "compute from inlet".

For the reference value of pressure we gave maximum static pressure value calculated in the inlet which is 4402 Pa (=rho*g*h=998.2*9.81*0.45).

For the area value in the "reference value" section, do we have to give the face area of the cylinder in the velocity direction ? (=d*b, d is the cylinder diameter, b is the channel width).

The problem we couldn't fix is the (P-Po) value in the Cp equation. The fact is we should get the value of 1 in the stagnation point. That means the (P-Po) must equal to the dynamic pressure (0.5*rho*u^2) in that point. I really don't understand what is wrong in our solution. Because we have a good agreement in the veloicty field.


I really thank you for your interest.

[Lisandro Maders]

So I think you are misunderstanding the Cp calculation. Are you performing a 2D or 3D analysis?

Regarding to Cp equation, it is this one: Fd/(0.5*rho*A*v^2)

Fd is obtained by integrating pressure values at cell's face areas, but Fluent does it for you. Instead of explicitly calculate the Cp, uses the surface monitor of Fluent.

The reference area needs to be consistent to the experimental results you are comparing to. If they used the sectioned are, use it so.

Are you using Fluent to calculate Cp or you are taking pressure values and calculating it explicitly instead?

Regards, Lisandro

[goksusoydan]

Hi,

We are doing 2D analysis. Fluent can be able to calculate the Cp. We can also export the pressure coefficient and pressure values. Fluent gave us a good distribution of pressure coefficient but we think the values are incorrect because we didnt't get the value of 1 as it is given in the attachment. So we also export static, dynamic and total pressure values from the Fluent (we have them in data file quantities of Fluent) and calculate the Cp explicitly, we couldn't fix it either.

Thanks..

[Lisandro Maders]

Are you able to share the mesh and fluent files? Then I can take a look for you and see what is going on..

[Shanis]

Hey Lisandro,

Did you ever get an answer for your ERCOFTAC study? I am using the same data for validation and have the same problem, not sure what reference pressure they used.

[n.phililipova]

Dear FLUENT Users,
I have to calculate pressure loss coefficient for tee junction by using FLUENT. How I can devide total pressure to dynamic pressure?
Is there anyone who can help me?
I am grateful for that help in advance!

[Lisandro Maders]

Originally Posted by Lisandro Maders

Hi,

I think you are confused about somethings. VOF does not mean Volume of Fluid. .....

I was taking a look at this right now and realized I made a mistake. VOF actually means Volume of Fluid, sorry about that.


Regards,

Lisandro Maders