stress generated by a thermal gradient

[beyfadn]

Hi
I'm a future civil engineer and i'm working on a marine concrete structure (solide).
I'm looking for a good finite elements software that can give me stress du to the variation of température (gradient).
I found a good one but it gives me only température and i'm interested in stress.
If you have any idea, please saaaaaave me

 

[IRstuff]

Student posting is not allowed

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[chicopee]

Remember double posting is frown upon.

 

[EnergyMix]

In order to "saaaaaaaave" you, you need to provide some information:

First off, was this an assignment from your teacher to research different types of finite element programs?

Second, how much money are you willing to spend? If we recommend something that is a couple of thousands of dollars is that acceptable? (It's unlikely we're going to recommend anything free.)

Third, what does it need to do besides give you stress due to temperature?

Want to know the do's and don'ts of Eng-Tips? Read FAQ731-376.
English not your native language? Looking for some help in getting your question across to others or understanding their answers? Go to forum1529.

 

[beyfadn]

Hi
well it's not a student post, i'm working on my first projet and the contractor is doing this job for the first time. it's difficult to calculate stress of a thermal gradient manually for a solid of concrete, so i'm looking for a finite elements program to do it in order to determine the steel section needed.

 

[davefitz]

Both conductive heat transfer and linear-elastic stresses are solved by the same numerical procedure, although the differntial equations are different. So, if you have a canned finite element program to solve for the temperature distribution , it should be feasible to rig the program to also provide linear elastic stress.

"Finite elements" method uses a numberical technique that requires the differential equation to be written in a "self adjoint" form, and the conductive heat transfer equation ( harmonic) and the stress distribution equation ( bi-harmonic) are both easily re-written in a self-adjoint form. Most of the effort then is related to properly imposing the boundary conditions, choosing proper grid size , and implementing correct integrating method ( for time dependent problems).

For reinforced concrete, though , one may consider that material as a composite of rebar, aggregate, and binder, so it does not striclty meet the assumption of homogenous properties. I would assume that certain local stress maximums may not be exceeded, but that is beyond my expertise.

"Nobody expects the Spanish Inquisition! "

 

[corus]

In a lot of cases you can calculate stresses from a thermal gradient using hand calculations. Refer to Roarke's forumlas for stress and strain for examples. Remember that you'll get no stress for thermal problems unless there's a restraint somewhere, or there's differential thermal expansion within the structure.

Otherwise you can go the whole hog and use one of many FE packages. Calculix is free and has a good reputation if you want to pay nothing.