Cancel
Showing results for 
Search instead for 
Did you mean: 

Beam Cross Section - Torsional Constant Values

Experimenter
Experimenter

Hello,

I'm doing a stuctural optimization of an I-beam and I need to input the value of the torsional constant of the beam into other softwares. So in FEMAP I'm interested into inputting the beam dimensions and get the beam properties as an output. I would like to do it through .txt or .dat files.

 

The problem is the following: I'm able to do so using PBEAML on Nastran, and I'm able to get the equivalent  PBEAM properties ECHOed in a .f06 or .pch file. But the values of the torsional constant are quite different from the "Original" and "Alternative" section evaluation methods in FEMAP (Define Poperty - Beam Element Type ->Shape->Section Evaluation).

 

For example:

Choosing a NASTRAN I shape with the following dimensions:

DIM1=0.39

DIM2=DIM3=0.04
DIM4=0.02

DIM5=DIM6=0.045

 

One obtains the following "Torsional constant" values according to the different "Section evaluation":

- Original: 1.88276E-6
- Alternative: 1.86624E-6

- PBEAML: 3.35E-6

 

I wonder why in PBEAML the value is almost twice as much as the Alternative (using the same reference point). Does anyone know the methods that are used here? Is there any way that I can use a .txt or .dat file as an input and get the "Alternative" properties as an output?  

 

A colleague of mine calculated the same torsional constant using other software and obtained the same result as the "Alternative" type.

 

Thanks in advance.

 

PS: I've also calculated it using J=(2*b*t_f^3+h*t_w^3)/3, and I have obtained a different value from the previous 3.

2 REPLIES

Re: Beam Cross Section - Torsional Constant Values

Siemens Phenom Siemens Phenom
Siemens Phenom

I believe PBEAML's in NASTRAN use hard coded formulas just like yours, however, both the NASTRAN one and yours are based on thin-wall theory.  The cross section defined is not very thin-walled. FEMAP's calculation of the torsional constant is done with a FEA solve of the cross-section, it will work with both thin-wall cross sections, and with thick.  I would go with the FEMAP calculated values for this one.

Re: Beam Cross Section - Torsional Constant Values

Experimenter
Experimenter
Thanks for your answer!