Structural Steel - Residual Stresses

[abusementpark]

Does anyone have some good information on the general effect of residual stresses on structural steel? Is there a typical range that one can expect the residual stresses to fall within? And is the effect of residual stresses ever a consideration in design?

 

[JStephen]

1) No, at least I don't.
2) I would expect residual stresses in the vicinity of welds to range right up to the yield stress of the material. I've never seen any measurements.
3) Only in an indirect way- relating to weld distortion, weld cracking of restrained joints, embrittlement due to caustic or hydrogen exposure, weld procedure issues, etc.

 

[Rich2001]

I agree with JStephen, adding on to his point 2

During the weld solidification state the internal stress in the weld approach the ultimate yield strength of the filler material. Note is this dependent upon the dilution of previous weld deposits, base materials and new deposit materials, any metaloids, carbides etc that precipitated along the grain boundaries and he cooling rate.
The internal (residual) stress distribution centered along the centerline of the weld deposit is more or less a normal distribution reaching reach near zero (0) just outside the HAZ,

If the weld has been stress relieved the same distribution will exist but the maximum stress value will be around 30% of the ultimate tensile strength.

There has been some research and published documents on residuals stress of structural steel. Here are a few links FEMA 350 Welding Code,

http://www.fema.gov/plan/prevent/earthquake/pdf/fema-353.pdf

,

http://www.fema.gov/plan/prevent/earthquake/pdf/fema-351.pdf

,

http://www.fema.gov/plan/prevent/earthquake/pdf/fema-352.pdf

,

http://www.fema.gov/plan/prevent/earthquake/pdf/fema-353.pdf

Commentary an recommendations on FEMA 350

http://www.seaoc.org/seismpdfs/seismofaqs/SEAOC_F350_v16-0.pdf

Richard

http://www.6sproductivitycom.com

 

[corus]

There are residual stresses in structural steel generally produced during the rolling provess, and probably caused by differential thermal expansion as the product cools. You'll probably see it more where there is a larger mass in the section shape, such as at the juncture of the web and flange in an I section, or in a rail head or foot. In some cases this can cause cracks to form in an unloaded section, so the residual stresses must be up to UTS. This effect is more significant when assessing fatigue life of an unwelded structure. The effect on the overall structural integrity of the section for static loads is negligible.

Tata but not yet tara

 

[Ron]

Residual stress are not usually associated with static design considerations, as noted. Further, residual stresses induced after manufacturing of the product (welding, bending, heating, etc.) may be considered, depending on their influence on the final structural use and fatigue. Usually not a big deal, but can be in fracture critical structures.

Residual stress can be measured with a special type of strain gage (I have done this only once or twice out of hundreds of strain gage applications).

 

[JoshPlumSE]

I'm not sure there is a whole lot of practical field measurements of residual stresses. However, this is a fairly well understood and codified effect. There are code required assumptions that you make regarding residual stresses that vary depending on whether you are using welded or rolled members.

I believe that the Salmon and Johnson book on steel design has some good discussion on residual stresses. At one point, I think the AISC design manual also has some good discussion in the commentary. Though I haven't looked through the latest version to verify that it's still there.

 

[B16A2]

My behavior of steel course went over this in detail in grad school. Highly recommend taking this course. Basically you get reduced toughness in the K region of the beam. Big problem for bridge girders, not so much for buildings.

If you're asking yourself this, then you should also be questioning the residual stresses imposed from welding.

 

[bridgebuster]

Ditto on Salmon & Johnson. I used the 2nd Edition in graduate school. We spent a good deal of time on the subject as well.

 

[271828]

Several pieces of misinformation in this thread. Sorry.

Residual stresses are codified all over the place in the AISC Specification and they are described in lots of good steel textbooks such as Salmon, Johnson, and Malhas that has already been mentioned.

Most of the residual stress stuff in the Specification have to do with compressive residual stresses at the I-shaped beam flange tips. For example, Section E3 has Fcr for inelastic or elastic buckling. If the column is very skinny, then it will buckle at a stress that's so small that yielding won't have occurred at the flange tips. That' show we separate inelastic and elastic buckling. Same goes for flexural lateral-torsional buckling and flange local buckling in Ch. F. Residual stresses are accounted for in the Table B4.1 lambda values.

Residual stresses are everywhere in structural steel design. They're definitely there in the static strength checks and they have a lot of effects beyond welding and fracture toughness.

 

[B16A2]

271828,
Are we talking about the same thing? Are you saying that the code accounts for the residual stresses that occur in a beam from when it comes out of the caster?...caused by differential cooling and segregation?

 

[frv]

271828..

Thanks for posting that..

As I was reading all the posts it shocked me that so many fellow engineers weren't aware that these stresses are codified, albeit not explicitly. And the column stability equations are the best example. I think we should all do a better job reading the commentary (myself included).

 

[RFreund]

Doesn't AISC basically use 0.3Fy for residual stresses (off the top of my head..), in that Mr=0.7Fy*Sx, the 0.7Fy, being the maximum stress before yielding?

EIT

 

[WillisV]

0.3Fy is the newer value used for residual stresses. As JoshPlum alluded to, it used to be different values for rolled vs. built up shapes. From the 2010 Commentary:

"The only difference between the 1999 LRFD Specification (AISC, 2000b) and this
Specification is that the stress at the interface between inelastic and elastic buckling
has been changed from Fy - Fr in the 1999 edition to 0.7Fy. In the specifications
prior to the 2005 Specification the residual stress, Fr, for rolled and welded shapes
was different, namely 10 ksi (69 MPa) and 16.5 ksi (114 MPa), respectively, while
in the 2005 Specification and in this Specification the residual stress is taken as 0.3Fy
so that the value of Fy - Fr = 0.7Fy is adopted"

As others have said, the effect of residual stresses in this context is definitely codified, both directly in the column equations but also indirectly in the stiffness reduction factors of the direct analysis method.

 

[271828]

"Are you saying that the code accounts for the residual stresses that occur in a beam from when it comes out of the caster?...caused by differential cooling and segregation?"

Yes. When we're talking "residual stresses" in structural steel, this is usually what we're talking about. It's about as hard to find an equation in the AISC Spec. that DOESN'T have something about residual stresses as the other way around. Very slight exaggeration.

Not to pick on you personally B16A2, but the thread is surprising. If the structural nerds in a forum like this don't have "residual stresses" down cold, then non-nerds probably never even heard of them LOL.

I blame the 1989 ASD and previous specifications for this kind of thing. Those older equations masked the behavior more often than not and we have entire generations of engineers who don't know what's in the equations. [Ducks and assembles flame retardant suit!]