Stresses at Junction of two pipes of different material welded together 1

[davemechuk]

I have two pipe ends of the same diameter that are welded together. The pipes being made from different materials having different Young's Modulus (E) and different Coefficients of thermal expansion (a). There is an internal fluid pressure operating at a high temperature.

I want to assess the principal stresses of the pipe ends at the weld junction using thick wall cylinder theory. For simplicity I'm ignoring the weld (for now) and assuming the pipe ends are perfectly bonded together. I am assuming there are no axial loads, and ignoring pipe loads for now.

How would I start to go about calculating the change in principle stresses due to the discontinuity of the junction under pressure under ambient conditions?

At the junction I know the following:
1. At the junction the strains will be equal for both materials
2. Displacements will be larger for the stiffer material
3. Displacements will be smaller for the less stiff material
4. There will be a bending stress in the meridonial direction

I suspect that I can assess the radial and hoop stresses using thick walled cylinder theory for each pipe under pressure and then superimpose resultant stresses due to the junction on the result?

I also need to consider the effects of stresses due to differences in thermal expansion.

Note I require principle stresses in all three directions, varying through the wall thickness.



Dave

 

[davemechuk]

2. Displacements will be larger for the stiffer material
3. Displacements will be smaller for the less stiff material

Apologies; the above points are in relation to the displacements for each pipe at the junction if I assessed each pipe separately.

Dave

 

[davemechuk]

I think I can calculate the resultant inner/outer radii and then calculate the resultant stresses based on the known displacement (see attachement). Right I'm going to shut up now... thanks in advance.

Dave

 

[TGS4]

You're heading in the right direction. Your calculations will (most likely) give you hoop and axial stresses and in-plane shears. You can use a 3D Mohr's circle to back out the principal stresses.

 

[dhengr]

Dave:
What kinds of specific (special) materials are you dealing with? What dimensions, temp. range, etc. etc.? For most steels E and coefficients of thermal expansion don’t vary that much. Maybe you would want to weld this pipe joint at some elevated preheat temp. which is the mid range of its operating cycle. Maybe you would want the two different pipes at different preheat temps. to try to balance the expansion +&-. And, maybe you are kinda over thinking this problem. What is the range of expansion movement btwn. the two pipes?

 

[davemechuk]

TGS4
Thanks for the sanity check. Though not sure how to work out shears of a pipe due to meridinal bending stresses... I'll have a look.

dhenger
Just for my benefit lets assume differences in structural and thermal expansion is not neglible. The other option would be to assume they are neglible, however I am still interested in solving such a problem. Though if it helps to consider the magnitudes of deflections; I'm looking at NiCu and Steel (I think differences in E and thermal expansion is roughly 10% and 20% respectfully).

Dave

 

[davefitz]

This is a common problem on high pressure/high temperature boilers. The radial shear stress is very high, and the cyclic fatigue and creep-fatigue damage can be problematic if the part is exposed to daily thermal cycles. In addition to the issue of stress and fatigue is the problem of high temperaure metallurgical trasnformations that occurs at the interface of 2 different materials- carbon may migrate from the high carbon material to the hi chromium material, for example,eventually leaving the interface brittle.

Another issue is the need to avoid placing this dissimilar metal weld in a section of pipe that is exposed to an external bending moment.

"Whom the gods would destroy, they first make mad "

 

[MJCronin]

AWS has done some research on this issue...

This may help:

http://www.aws.org/wj/supplement/WJ_1982_02_s58.pdf

MJCronin
Sr. Process Engineer
Venture Engineering & Construction

http://www.VentureEngr.com