Thermal stress at circumferential weld, acceptance criteria 2

[wacee]

Hello all, I'm a newly graduated mechanical engineer trying to navigate all the different standards.
I have simulated the thermal stress at a circumferential weld between two different metals but I am not sure to what acceptance criteria I can compare these values. Looking into ASME B31.3, can I compare the simulated longitudinal stress sigmazz to the Basic allowable stress S ? or the shear stresses simulated to 0.8 S, the allowable shear stress ?

Thank you for your insight !

 

[TGS4]

wacee - good question and thank you for asking.

Before you start diving into comparing certain stresses to other certain allowables, I need you to rethink your premise and ask yourself - how will this stress cause a "failure" - and more specifically, what is the failure mode? Since it is a thermal stress (cause by satisfying compatibility, it is not related to burst (or plastic collapse), which is caused by stresses that satisfy equilibrium. Since it is not plastic collapse (burst), then you are left with buckling (only if some of the stresses are compressive), ratcheting (progressive incremental inelastic deformation), or fatigue. Start looking in that direction before deciding what is the appropriate criteria.

You mentioned ASME B31.3. Is this application a B31.3 Code situation?

 

[r6155]

I´m curious: temperature, materials, thickness. diameter.... ?

 

[davefitz]

I have witnessed prompt low cycle fatigue failure due to thermal stress on high energy piping at USA electric power plants, when welding thin wall P91 HP main steam pipe ( 1.5" thk) to 2.25Cr stop valves that have about 50% of the P91 allowable stress at 1050F, and therfore are over twice the thickness of the P91 pipe ( 3+ inch vs 1.5" ). The issue is that both ASME B31.1 and section I allow a rate of change of wall thickness at a slope of 0.5 near the weld line , while the European code recommends using a weld neck or using a slope of less than 0.33 in thermal cycling service. The problem seems to be that ASME section I does not yet recognize thermal fatigue damage while the US power plants now have been converted to thermal cycling service. Refer to EN 12952-3 table B.1 item 2 vs table B.3 item 1.The better practice ( as recommended by GE) is to use a P91 transition piece between the main steam pipe and the thicker 2.25Cr valve with a slowly thicknening slope.

"...when logic, and proportion, have fallen, sloppy dead..." Grace Slick

 

[wacee]

Hello, thank you for your answer !

The standard deals with thermal stress only from a fatigue stand point and it depends on the number of cycles.
I was trying to look at the stress profile when the cold boundary temperature is first applied so I believe there is no notion of fatigue in this case. Maybe ratcheting ? Since the stress at the welds could possibly cause yield if it was too great. I am not sure.

The use case is a liquid ammonia pipeline so if understood correctly B31.3 applies.

r6155, to answer your question, the temperature is -33°C. The idea is to compare different materials but it will be LTCS and stainless steel. Diameter is 6" and thickness is 10 mm.

 

[r6155]

Do you really consider that the difference between the operating temperature of the pipe -33°C and the atmospheric temperature produces thermal expansion?

Regards

 

[wacee]

Hello again,
I apologize for the late reply.

I do consider there is thermal expansion even if it is small, I have a model to quantify how small but I'm still wondering against what values to compare the resulting stress. Maybe SMYS ? Even after going through B31.3, I wasn't able to determine the displacement stress range.

I am interested in a comparison criteria because I would also like to verify that thermal stress for ammonia and colder applications like LN2.

 

[weldstan]

It's really thermal contraction and it is so small when operating at -33 C that it can be ignored at weld joints to dissimilar metals like low carbon steel to austenitic stainless steel. It should not be ignored in the pipeline design with regard to the length of the pipe and providing freedom of movement to allow for the shrinkage of the piping system when installed versus in operation.

 

[TGS4]

wacee - I am going to push back against your Jun 14 answer that B31.3 "deals with thermal stress on from a fatigue stand point". In fact, the limit on "expansion" stresses within B31.3 is for "elastic follow-up", which is the B31.3 term for ratcheting.

If you are looking at a circumferential weld at a dissimilar weld location, then your R*(Δα)*ΔT will likely be quite small at your temperatures.

 

[KevinNZ]

@davefitz
A bit OT, but when you weld 1.5" P91 to 2.25Cr stop valves you must transition in the material with the higher allowable stress. If you transition the lower strength material it will be too thin at the joint.

 

[davefitz]

KevinNs
Yes, that is true, except the ASME code does not require it. Unfortunately a design engineer straight out of college often does not have such insight and so such errors get into the field and cause failures.

"...when logic, and proportion, have fallen, sloppy dead..." Grace Slick