cylinder subjected to internal pressure

[american500]

I have created a thin walled cylinder in PATRAN as following:
Skin Thickness: 2.5 mm
Cylinder Height: 40000 mm
Cylinder Radius: 3280 mm
Material: Aluminum, Young's modulus 70000, v=0.3
Load: Uniform Distributed internal pressure=0.059 N/mm^2

After running the model, I noticed that the cylinder skin becomes wobbly!! What is the cause, and how can i fix this issue.

 

[40818]

Can you give me info on elements used, contraint system, what you want out of the model,and what you mean by wobbly??

From a quick check of some numbers you should get:
Hoop Stress = 77408N/mm^2
Longitudinal Stress = 38704N/mm^2

If you dont get these numbers your model is flawed

 

[american500]

The cylinder is meshed using Quad4 elements. 112 elements on the circumference and 80 along its length. One end is constraint in translation only <0,0,0> in the global coordinate system (coord 0).

The Hoop stress obtained matches closely the calculated value 77.408 N/mm^2.

I expected the cylinder to expand uniformly around its circumference but instead it expanded in an irregular manner (variable radius).

 

[40818]

Off the top of my head:
If you have constrained 1 side, then when your applying your pressure load it has to be reacted by the contraint system, so you will have less radial deflection close to the contraints. Though if you simply support both sides and have them far enough away from the point of interest then you would get an even deflecetion in the middle, with it starting to get quirky the closer you get to the contraints.
By variable radius do you mean the effect i described above (small taper at end - almost like a cone) or is it a local thing?
Did you create a cylindrical coordinate system?
Also, what are you trying to prove/solve by the model, it might help with the solution.

 

[40818]

Just re-read your post, did you mean you wern't getting a "round" expansion (i.e circular), but more of a irreglar pattern if you look down in 2D?

 

[american500]

I am interested the effect of introducing passenger door cutout in both metallic and composite fuselage. I am starting off with the uncut metallic case first (only skin, no frames or stringers). I am going to analyze two separate cases, one being the bending moment shear force load and the second is the pressure differential load.

I tried to constraint the fixed end in both cylindrical and Cartesian coordinates, i also tried a case with an end cap on but the cylinder kept inflating in an irregular manner.

Yes, i am not getting a round expansion, if looking down the cylinder in 2-D view i see a something similar to a sine wave around the circumference.

I wish i could post a picture to be more clear.

Thanks

 

[40818]

Ok, have you any experience about fuselage stress analysis??

Firstly, go to nasa tech records, download and read the following reports:

NACA 998, 1135, 1202, CR-4731, TM-999, TN-687, TN-847, TN-929, TN-1098, TN-1830, TN1831, TN-1974, TN-D400, TN-D401, TN-D402, 1239

Also, you can use some ESDU methods, and also propriety manuals.

Simply put, for a pressure case on a pure cylindrical member, their is no bending moment as the whole sysytem is self balanced as a memberane tension.
Things start to get more complicated when you start introducing fames, stringers, and discontinuites of skin curvature. These systems are all still however statically determinate.
Introduction of cut-outs causes problems for the analysis and capabilites of the airfram, think back to the torsional differences between a cylindrical shape and the shame shape with a minute cut the length, think the difference is about 482 or so.

Anyway, rambling on.

As soon as you introduce the floor beam you step into a statically indetermiante structure, and decent FEM is the way forwards. Note the words "decent", because as soon as you move away from pressure cases your constraint system will be far from reality, and give crap results.

If you have a floor beam in your model you will end up with "S" curves around the intersection.

Have a read of the NACA tech notes and this will hopefully give you your answer as to the effect of the cut-outs.
Let me know if you need any help.

 

[american500]

Thanks allot, i will investigate the reports. I never worked on fuselage structures before.
Regarding my model, it seems that adding the frames and stringers eliminated the problem of skin wrinkling. The next step will be adding floor beams to have a “realistic” model of a passenger door cutout. I will use guidelines provided in C.Y. Niu’s book to size the reinforcement structure around the cutout in the metallic version. Are you aware of any methodologies for reinforcing passenger door cutouts in composite fuselage?
Thanks

 

[40818]

Starting with Mikes book is as good a place to start as anywhere, though i would keep a copy of Bruhn and any of Paul Kuhn's work (NACA documents and his book) to get further insight.

Interestingly enough, with composites, a square cut-out is somewhat beneficial with regard to corner stress concentrations (against a metallic airframe) due to the angular plies (0, 45, 90 etc) redistributing the stresses such that singularities are reduced. You have to be carefull with delamination failure due to interlamina shear stresses and normal stresses. As a very good rule of thumb, keep your cut-out sides parallel and perpendicular to the principal stresses along with a high percentage of fibres aligning with the principal stress direction.

You have to be carefull with laminate composites anyway, as the modes of failure can be varied and complicated.

As regard to any set methodologies of composite fuselage cut-outs im not aware of any, though maybe somebody from Spirit Aerosystem might be able to help with that one.