wood truss structure

[newmann]

I'm quite new in FEM and want to calculate wood truss structure (beam elemnts). But I have no iedea which values should I use for my structure to define properties, because wood is 3D orthotropic.

Thanks a lot.

 

[EdR]

Either use 3-D anisotropic values or come up with equivalent isotropic values that give the results you are looking for...

Ed.R.

 

[SWComposites]

For a beam model you don't need 3D properties, only the axial E and G. Use the same values as you would use for a hand analysis of the truss.

 

[RWF7437]

I was always taught that the members of any truss are always axially loaded, or are assumed to be. If so, they are NOT "beams". If they are not axially loaded then the structure is some kind of frame, rigid or semi rigid.

 

[newmann]

I tried to make assumtion that I have isotropic material. In a literature, I found E=10000 H/mm2 and G=800 H/mm2 (DIN norms). But for isotropic material, the following equation G=E/2(1+v) have to be valid. I have in resalt v=5.25 that looks totaly wrong and I have warning message from nastran: one or more mat1 entries have unreasonable or inconsistent values of E,G or NU.

 

[GBor]

Nu should never exceed about 0.5, and that value is for soft rubber. The harder a substance gets, the closer nu gets to 0.3(steel) or down to about 0.2 for some composite materials. For your wood truss structure, you should be using truss elements, not beams. "beams" are not always axially loaded...they take bending. "Truss" elements are axially loaded.

Beams and trusses, while they appear similar, are analytically different by the "degrees of freedom" for each element. Search this forum, I'm sure someone has probably written a dissertation on degrees of freedom.

Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.

http://www.borowskiengineering.com

 

[EdR]

Garland is absolutely correct...the max value for nu is 0.5 (and you don't want to use nu=0.5)...Suggest you check a table for typical values of E and G (or nu) for the specific wood you have....then make other values consistent with the relation between E,G, and Nu...

I think you can use either beams or truss elements and if modeled correctly they will give similar results for your structure...In any case don't believe any results using the values Nastran is complaining about....The results are garbage....

Ed.R.

 

[GBor]

I would agree with Ed.R. that beams and trusses SHOULD give similar results, if your structure is truly a truss, but for a truss structure, truss elements are generally faster in computation (less degrees of freedom) and they insure that you get only axial loads in the memebers.

 

[ThomasH]

Wood, interresting material.

Strictly speaking wood is not a isotropic material, therefore G = E / (2(1+v)) does not apply. Depending on the direction you study compared to the fiberdirection the E-module varies, a lot.

What you need is a 3d anisotropic material model, or some simplifications.

You say truss, that means that E-module parallel to the fibers (typically the "highest" value) is what you need.
G controls shear deformation, not a very significant parameter for a truss.
Finally v, also not that important for a truss. It is typically 0.025 or 0.4 depending on fiber direction. Say 0.025, calculate G as for the isotropic material using a correct value for E.

I don't think you will be too far from the truth. Also, Nastran (at least the MSC flavour) as far as I know uses the values you input. It's complains if isotropic relation is "violated" but nothing is changed in the numbers. "Garbage in, garbage out".

Good Luck

Thomas

 

[EricZhao]

Let me just deviate from what the first post asked a little bit. Does anybody have any successful experience to use FEA for wood material analysis?

I doubt the material property of wood would have so much stand deviation, which makes numerical computation meaningless. Am I right?

 

[GBor]

Good Comment, EricZhao. I have done some wood material analysis, and with some reasonable success, but it isn't as simple as a metal analysis. For this truss, however, I would think it should be pretty accurate.

If this is as simple as a house roof truss, I would suggest calculating it by hand. A complicated building truss, FEA could speed things up.

 

[newmann]

I have EL=10000 N/mm2,NuLR=0.292,NuLT=0.449,NuRT=0.39,NuTR=0.374,NuRL=0.036 and NuTL=0.029. As I undestand, for the analysis I can use EL and NuLT. Is it correct? Thank you everyone for the very useful answers.