Learning Forum

[Kevin Gnanaraj]

 Hi,

I'm simulating a shell wing using imported pressure data for a static structural analysis. However, when I apply the pressure, it expands outwards. So I changed the pressures to negative, but the wing deflects down. Is this correct behaviour, or do I need to process the pressure data in a different way? Is there a way to see which side of the wing surface the pressure is being applied to?

[peteroznewman]

Airflow around a wing creates a low pressure on the top surface of the wing and high pressure on the bottom surface. The sum of these pressures over the surface area of the wing creates the lift force to hold the plane up. 

To apply pressure to a shell model, you need to apply the pressure difference between the absolute air pressure inside the wing and the absolute air pressure on the outside surface of the wing.

Suppose a plane is flying at sea level, and ambient air pressure is around 15 psi. The absolute pressure on the top surface of the wing may be around 14.4 psi, while the bottom surface of the wing may be around 15.3 psi.

Therefore, you would apply an outward pressue of 0.6 psi to the top surface of the wing and an inward pressure of 0.3 psi to the bottom surface of the wing.  These pressures will have the effect of deforming the upper skin outward, the lower skin inward and bending the entire wing upward.

Whether you put a + or a – sign in front of the pressure on each surface depends on the surface normal. In Mechanical, an arrow displays the direction of the pressure. If you type a positive value and the arrow points in the wrong direction, just add a – sign and the arrow will point the opposite direction. In SpaceClaim, on the Measure tab, you can select a surface and click the Normal button to see an outward normal arrow. Positive pressure is defined as an inward normal on any surface.

Is the imported pressure absolute pressure or relative (gauge) pressure?  You have to know this. If it is gauge pressure, you need to know the ambient or reference pressure.

[Kevin Gnanaraj]

 

I’m getting Cp values from OpenVSP, where I’m running my CFD. I’ve been turning it into absolute pressure using this formula: https://en.wikipedia.org/wiki/Pressure_coefficient#Definition

In SpaceClaim, the normals are pointing away from the surfaces of the wing, so it seems that pressure on the bottom surface would need to be turned negative, and on the top would need to stay positive, if I’m understanding you correctly?

 

[peteroznewman]

No Kevin, I don't think you have fully understood what to do.

If you multiply Cp by 0.5*rho*V^2 you will get the pressure difference.

The pressure difference will be positive on the lower surface and negative on the top surface of the wing.

The pressure difference is the applied pressure in Ansys.

Since the surface normals for the entire wing are outward normals, a positive applied pressure will point inward.

That means the positive applied pressure on the lower surface will point inward (upward) and the negative applied pressure on the top surface will point outward (upward).

These applied pressures will have the effect of deforming the upper skin outward, the lower skin inward and bending the entire wing upward.

You can see the arrows in Mechanical when you apply the pressure to confirm the direction is as expected.

Therefore you should not have to change sign on any of your pressure differences.

[Kevin Gnanaraj]

 

Hi Peter, I’m not sure I fully understand. I’m actually also adding the free stream static pressure to get p in that formula, since I know it from ISA tables, so I’m not using a pressure difference as you seem to be talking about. When I run ANSYS without changing any signs, the following deformation, which is unphysical:

As can be seen, the skin is not transferring the pressure load to the ribs and the spars, but treating those as boundaries to deform through. If I don't add the static freestream pressure, as you suggest I do, I get this result:

[peteroznewman]

Kevin,

On the Result tab is the Display Scale Factor. I think you have this turned up so high that the bottom skin is poking up through the top skin.  Try setting it to 1.0 (True Scale) value.

Plot directional deformation along the Z axis so you can see different colors when the panel moves up vs down. This is more informative than total deformation. 

[Kevin Gnanaraj]

Thanks Peter, it was set to autoscale. The deformation looks more realistic.

Can you clarify whether I should be adding the freestream static pressure or not? I imagine that term decides whether inner pressure of the wing is accounted for in ANSYS

[peteroznewman]

Here are two examples to think about pressure on a shell model.

Imagine a steel drum where there is a screw-in plug to seal pressure in the drum.  With the plug removed, there is equal pressure on the inside and outside of the drum so the applied pressure would be zero and there is no deformation.  Now boil some water in the drum and when it is filled with steam, screw in the plug and float the drum in a tub of ice cold water. As the steam condenses and the temperature goes down inside the drum, the pressure on the inside of the drum will go down, but the atmospheric pressue on the outside of the drum doesn’t change. At some point the drum collapses when the pressure difference is larger than the buckling strength of the drum.

If you take another drum and screw the plug in at sea level when the pressure on the inside of the drum is 101,000 Pa. Then take the drum up to an altitude of 9km, but keep it at the same temperature as it was at sea level the air pressure drops to 31,000 Pa.  If you have a shell model of the drum, you would apply an outward pressure of 70,000 Pa.  The sides of the drum would extend outward.

It is always the pressure difference.  If the wing is not pressurized, the static pressure on the inside of the wing is Pinfinity, the same as the air pressure at whatever altitude the plane is flying at.  The pressure in the air moving over the outside of the wing surface is either higher or lower than Pinfinity.  It is the difference that is applied to the shell model surface.

[Kevin Gnanaraj]

Thanks Peter, that explanation cleared up the issue for me.

My hang up was that I was using a solid model to get familiar with ANSYS, and I wasn't sure about how it handles pressure loads.