FEA Analysis

[gtmspyder]

Anyone got any FEA software who can easilly tell me what wall thickness I need with 120mm dia steel tube (for a one-off chassis backbone) that will give me a tortional stiffness of about 5000nm per degree over 2 metre?

This will be supplemented with a spaceframe/cage so 5000nm/deg is not all I'm aiming at if anyone is worried!

[egomeister]

You don't need FEA for that - its just a simple calc assuming you are just basing it on the tube. Hold on...

[egomeister]

I make it 6.2mm - anyone care to verify that I don't have dyslexic calculator fingers? I can check it again more thoroughly after work if nobody else replies and confirms it.

Hopefully that edit was so stealthy no-one spotted it...

>> Edited by egomeister on Wednesday 13th July 14:21

[MEMSDesign]

I'm afraid I make it 6.2mm.

G = 80 GPa
T = 5000 N.m
L = 2 m
theta = 1 deg
ro = 60 mm
ri = ro - t

theta = T.L / [ 0.5*pi*(ro^4-ri^4)*G ]

For t = 6.2 mm
theta = 0.995 deg
or 5025 N.m per deg

For t = 5.7mm
theta = 1.069 deg
or 4679 N.m per deg

I'm a bit slow (but accurate first time)

>> Edited by MEMSDesign on Wednesday 13th July 14:41

[gtmspyder]

Thanks guys....looks like its 6.2mm then

As a rule is round tube tortionally (if thats the correct word?)stronger than square or rectangular section?

>> Edited by gtmspyder on Wednesday 13th July 23:06

[MEMSDesign]

gtmspyder said:

Thanks guys....looks like its 6.2mm then As a rule is round tube tortionally (if thats the correct word?)stronger than square or rectangular section? >> Edited by gtmspyder on Wednesday 13th July 23:06

Tortional stiffness of hollow rectangular section, thin walled follows similar forumla to before.

theta = T.L / K.G

Where
theta = twist angle
T = torque
L = Length
K = Geometric factor
G = Shear modulus of material (around 80GPa for steel)

for concentric hollow circular section (as shown above)

K = 0.5 * pi * ( ro^4 - ri^4 )

for hollow rectangular section

K = [ 2*t*t1*(a-t)^2*(b-t1)^2 ] / [ a*t + b*t - t^2 -t1^2]

Where
a = long section side length (outer dimension)
b = short section side length (outer dimension)
t = short wall thickness
t1 = long wall thickness

I'll leave it to you to play with the numbers. Without running the numbers, I'd say that circular is probably the best shape in terms of weight vs stiffness tradeoff, but don't take my word for it - I haven't checked and my job involves designing things which generally aren't circular. Are you not interested in bending stiffness too?

>> Edited by MEMSDesign on Thursday 14th July 09:26

[gtmspyder]

Thanks again for the help.

I am concerned by bending stiffness, but generally less so than tortional stiffness - my logic being that a little but of bending longitudinally is less harmful to the handling than tortional stiffness.