I want to express my question more clearly

[ping]

JHardy, Fhashemi,

Thanks your all words,I'll try do something like you said for a while

And I want to express my question more clearly once again

So first discussing pure torque in real life.

Actually, Pure torque divides into two sorts, one is free torsion which means there isn’t any stress along the axis, so there only is shear stress in the shaft. This is the most simplest pure torque situation such as someone twists a shaft on two end by the two hands with the same force,only twist no press or other force.

I want to simulate this case in Pro/M using 3D model not beam..

If only applying moment force(torsion) at two end of a shaft, no constraints, just like real life, Pro/M warns me ‘no constraint sets have been defined’

If applying moment force(torsion) at one end and constrain at another end,I think the most important thing is how to define this constraint. I tried several methods, but all effort was failed. Sometimes the result wasn’t the situation of real life like my thought or sometimes Pro/M didn’t let me go on. I’ve never got a result it look like in real life.

And other sort of pure torque is ‘constraint torsion’ which means there is stress along the axis even only applying torque, such as one end has been fixed somewhere, then a torque is applied to other end. In this case stress along the axis will be very big if the section of shaft is other shape except circle, square or ellipse. And just because poission’s effect the fixed end will be expand or contract at boundary. I want to simulate this situation too, and I faced more difficult and more fail.

So I really want to know how can I do.

Thanks you all again

 

[JHardy]

Ping,



I can't figure out how to attach files here, so I have sent you one by separate e-mail. I hope you can open Pro/E-SE files.



Getting your constraints right is always one of the big challenges with FEA. You have to apply sufficient constraint to allow the model to solve, but not too much so as to over-constrain the model, otherwise your results may be meaningless. Never forget that FEA is only a numerical solution to a theoretical approximation to a real-world problem – some level of approximation and simplification is inevitable. The art of FEA is to manage the simplifications and approximations in such a way that the FEA solution is useful to the solution of the real-world engineering problem.



For a simple problem like a shaft in torsion, I would suggest that you try either applying a boundary condition of symmetry on the mid-plane of the model, and only create a half-model of the real problem. Alternatively, you can try fixing one end, as has been previously suggested, or alternatively, apply equal and opposite loads at both ends of the model, and apply some very soft springs to constrain the whole model, while generating minimal (i.e. theoretically zero) restraint forces.



All of these approaches are valid modelling techniques. If none of the above makes any sense, I can only apologize for my lack of clarity, and suggest that you try to speak directly to an experienced FEA mentor in your local area, or consult one of the many excellent reference books on FEA modelling techniques. I have already suggested Adams and Askenazi as being one of the best books I have come across on the subject – well worth trying to find a copy!



Hope this helps.