Torsion Constant - J

[AeroJDK]

Hey everyone...quick question:

I have a cross section of a solid that is, well, quite nasty looking, but I need to calculate the Torsion Constant J (not polar moment of inertia). Is there a software package that would allow me to do this? I have been using SolidWorks for my modeling and other structural properties, but I can't seem to make it spit out the Torsion Constant.

Any help would be greatly appreciated...thanks all.

Jon

 

[JAE]

RISA Section is a program that calculates J for generic shapes.

http://www.risatech.com

 

[corus]

There are free software that can calculate the section properties of a general section. There is a site mentioned in the spreadsheet forum where you can download one.

corus

 

[corus]

http://yakpol.net/SectProp.html

corus

 

[AeroJDK]

Thanks for the links. I was browsing through the SectProp spreadsheet, and correct me if I'm wrong, but it looks like that is calculating the Polar Moment of Inertia J, NOT the Torsional Constant J....right?

If anyone has any input on this or, again, knows where I could find something that calculates the Torsional Constant J, that would be most helpful. I should have mentioned that I also have access to CosmosWorks...

Jon

 

[AeroJDK]

I apologize - I see the various worksheets down at the bottom of the spreadsheet, in which the third one is calculating J. Only problem is that my cross sections are nothing like that.

Basically, I have a cross section of an aircraft with the skin and the stringers attached to the skin. I have a solid model of this, but I need to find the Torsional Constant at a give cross section...cute, eh? This is why I'm trying to find software, etc. that would assist me with this...

If I was able to find the Torsional J for all the stringers and skin seperately, is it a simply summation to get the torsional J constant?

Jon

 

[Drej]

> I was browsing through the SectProp spreadsheet, and correct me if I'm wrong, but it looks like that is calculating the Polar Moment of Inertia J, NOT the Torsional Constant J....right?

For the "torsion_properties" sheet you should refer to Roark (since this is the reference used). Roark uses a particular form for the torsional calcs which include this constant, so refer to this. The convention can be somewhat mixed for a "torsion constant". It's down to a description of the parameter J I suppose wherever it's used. Usually the torsion constant and the polar moment of inertia are the same thing - J - but it does differ. The torsional modulus is r/J (cf. section modulus in bending problems Z = y/I).

The "Section_by_nodes" sheet looks like it is indeed the "polar moment of inertia", but you'd need to go throught the calcs on the sheet to verify that it is truly the polar moment and not just the sum of the Ixx/Iyy second moment of areas.


Cheers,

-- drej --

 

[VoyageofDiscovery]

Hi AeroJDK,

For torsional resistance constant R and Polar Moment of Inertia J are approximately the same for closed sections with them identical for circular sections.

The general equation for closed tubular sections is:
R = 4(A)^2/(sum of (mean perimeter/associated thickness))

Where A is the area enclosed within the mean perimeter.

HTH

VOD

 

[GregLocock]

Does that apply to heavily re-entrant closed sections? I'm pretty sure it becomes an unworkably innacurate approximation quite quickly.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[Denial]

The FE package Strand7 (

http://www.strand7.com)

has an inbuilt "utility" for defining any arbitrary beam cross-section. It will then calculate all the section properties, including the torsion J. Furthermore it also allows you to use that defined cross-section directly in any subsequent structural modelling, and will display correctly any beam that uses that cross-section when you ask the program to show beams as true cross-sections rather than single lines.

 

[VoyageofDiscovery]

Hi Greg,

That general equation is for thin tube type structures that should apply to an aircraft cross-section. If the cross-section is not tube in nature, such as a thick cross-section relative to the thickness of the elements and with reentrant corners, I doubt this formula will be accurate.

Regards

VOD

 

[GregLocock]

Thanks, that's what I thought. The only way I'm happy with these days is to build a FE plate model of a straight beam of the section, and then test that in torsion, with appropriate end conditions.

At the very least designing appropriate end conditions is an eye opener, if you haven't done it before.

The advantage of building these simple models is that you can also investigate joint efficiencies very quickly - which given that a badly detailed joint can be only 50% efficient is well worth doing.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[JStephen]

"but you'd need to go throught the calcs on the sheet to verify that it is truly the polar moment and not just the sum of the Ixx/Iyy second moment of areas"

Aren't these the same things? IE, sum of R^2*dA = sum of (x^2 + y^2)*dA = sum of x^2*dA + sum of y^2*dA

 

[fleisher]

Try Sectionwizard from REI (STAAD.Pro guys)

http://www.reiworld.com.

It can calculate the warping constant as well if you need to calculate warping stresses on your x-section.

 

[GregLocock]

JStephen, no they aren't. Torsion violates the assumption that plane sections remains plane, amongst other things. That is, by using the polar moment of inertia, you will overestimate the rigidity of your structure, in many cases.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[rb1957]

aeroJDK,

see Roark, 7th ed, pg381
then there is a table of torsional constants depending on the shape and constraint conditions. remember this is for prizamatic and open sections. it would appear that there isn't a general calculation approach for determining this, other than for open, thin walled sections (being 1/3 bt^3).

closed sections, and particularly sections with stiffeners and multi-cells, need a different approach, because the shear flow is different in the various segments of the cross-section; refer bruhn.

 

[JStephen]

Greg, I meant that "polar moment of inertia" and "sum of the Ixx/Iyy second moment of areas" are the same thing- not that these are equal to the "torsional constant".

I notice in "Formulas For Stress and Strain", they use J for the polar moment of inertia and K for the torsional constant, which would avoid some of the confusion above.

 

[jameslord]

ShapeDesigner v5.0 is a very good software that can compute the properties ( torsion J , warping Cw including the stresses analysis)of any composite section (thin , solid, mixed etc..) at

http://www.mechatools.com