Modeling a threaded part 1

[dem]

I am using COSMOS to look into the max. stresses at the shoulder fillet of a threaded part due to it's installation torque (using 2D axisymmetric elements, linear analysis). Simply applying the resulting axial force at the threads and reacting them at the "head" of the part results in stresses that are way too high - the model predicts that the part should start yielding way before the actual installation torque used is reached. Does anyone have any tips that might be useful?

Thanks

 

[prex]

If you mean that you see very high stresses at the bottom of the thread, that's quite normal: in fact any bolt will suffer some yielding when properly tightened. What you should look at is the average axial stress across the bolt and discard the rest.
In fact FEM is not a good method for analyzing threads, as local (elastic, hence theoretical) stresses critically depend on the thread bottom radius and other parameters, including the surface finish (a rolled thread behaves very differently from a machined one).

prex
motori@xcalcsREMOVE.com

http://www.xcalcs.com

Online tools for structural design

 

[mechdesign]

linear analysis is not the correct way to analyse the stress concentration at the root of thrd. it gives u stresses at the higher end.increase in no of elements may be a way to get what u want.

 

[dem]

I'm not looking at the threaded region of the part. If you were to think of the threaded part as a bolt, I'm looking at the max. stress in the fillet between the bolt shank and head.

Thanks for your responses.

 

[austim]

Hi, dem.

When you say that you are analysing for the effects of applying the installation torque, how have you calculated the axial tension that will be developed?

(That is my "diplomatic" way of asking if you have remembered to allow for the full effects of friction between threads and between the head/washer/clamped item).

 

[dem]

austim,

I'm using the following equation to come up with the tension load (solving for P):

Torque = 0.2 x dia. x P

 

[austim]

Hi again, dem

That looks good enough to me, and clearly does include proper allowance for friction. In which case my query was not particularly relevant to your problem - sorry about that.

(I only asked since I was once faced with some serious calcs which did ignore friction, and gave a tremendously over-estimated tensile load).