Fatigue analysis of a welded joint 1

[Ameolive]

Dear all

I really appreciate if some of you can provide with some guidelines regarding knock-down factors to apply to material fatigue allowables for the fatigue analysis of a welded joint.

I'm asking this because I been reviewing a fatigue analysis of a welded joint connecting two pipe shaped components.

The fatigue allowables used for the analysis do not account for any reduction due to the welding on the basis that :

1) TIG welding is used.
2) A radiographic inspection of the welded area is requested after welding.
3) An heat treatement is requested after welding to restore static stenght.
4) A repetitive 100 hrs detailed visual inspection of the welding is requested in service
5) A reduction factor of 3 is anyway used on the average fatigue material allowables ( this because the analsys results will not be validated by any dedicated fatigue test on the part).

Despite the above considerations, I'm still reluctant to accept a fatigue analysis of a welded joint that does not taken into account any knock-down factor for welding.

On the other hand if we apply the standard 50% reduction we come up with fatigue allowables that are too low ( do not forget the reduction factor of 3 already applied )and I believe unrealistic considering all the provisions taken.

Thanking in advance for any help provided

Ameolive

 

[alexit]

I have done analysis of welded rod rotating while under bending load (~10M cycles in 2000 hours operation typical.) Because this was a primary safety concern (human injury can result if double failure) we chose to apply 50% reduction on top of x-ray analysis, anneal, heat-treat, post-process cryo, and 7x applied safety factor...never hear of one failure and all are out of service by now (replacement part so is trash every service cycle.)

Failure Effect Analysis should help you decide if you need to be more stringent in your analysis.

 

[Carburize]

I believe BS 5400 contains codified allowables for fatigue strength of various types of welded joints.

 

[Ameolive]

Many thanks to all for the reply as a matter of fact the pipes in question are part of a SAE 4130 frame supporting a rescue hoist used for human external cargo operation with helicopters.

So from this point of view is a critical installation and all the relevant precautions/ conservativism has to be taken for the fatigue analysis.

I surfed the web and I have got some interesting information : in particular report NACA TN 1261 provides a lot of useful information on fatigue strenght of welded SAE 4130 joints.

From the report in question it can be seen that at maximum a 40% reduction of fatigue allowable can be expected so I'm going to take this figure for may calculations together with the factor of 3 reduction used in any case.

Thanking again

Ameolive

 

[Tmoose]

NACA 1262 is one of my favorites. They stuck 4130 tubes to plates and subjected them to rotating bending. The stress concentration at the "toe" of the weld from geometry or undercut was the overPOWering factor. Crappy welds with nice blended toes far outlasted gorgeous welds with convex toe geometry

 

[Ameolive]

Thanks Tmoose,

can you tell me where I can find NACA 1262, I searched the NACA database but I could not find it.

I was quite interested in this report as I understood that there are some information also on the effects of post welding heat treatments.

Regards

Ameolive

 

[amorrison]

May be it is 1261.

http://naca.larc.nasa.gov/reports/1947/naca-tn-1261/

 

[Ameolive]

AMORISSON4

I confirm the report in question is NACA 1262.

After a search with google I managed eventually to download it on my computer.

TMOOSE is abosolutely right ( thanks again ) the information provided in NACA 1262 are unvaluable.

Quite interesting is the conclusion that the fatigue degradation of welded joint as opposed to machined parts seems to be almost exclusively dependent from the geometry of the weld fillet regardless of the quality of the welding process itself ( eg. there is no point of using TIG welding and post heat treatment if the quality of our weld fillet is poor ).

Another point to mention is that the fatigue degradation appears to be more sensible at high number of cycles ( in the range of 1 million ) with a reduction of the fatigue endurance down to 80% of the fatigue limit of the machined part.

As for my design I pretty confident that it meets the safety criteria considering that I have applied for fatigue calculation a reduction factor of 4 on the nominal fatigue endurance and that the maximum alternate stress estimated from a dedicated load survey is below 80 MPa ( 11 ksi ).
Further to that a detailed inspection for cracks at 100 hrs interval is established in the maintenance procedure.

Best regards
Ameolive

 

[arto]

http://naca.larc.nasa.gov/reports/1948/naca-tn-1262/naca-tn-1262.pdf

 

[Tmoose]

It's kind of unfortunate that the fillet geometry is SO important especially in this special case of bending and fatigue. It makes sense to me that the old NACA test liked concave joints for the better geometrical stress concentration, but convex fillets are far superior in resisting weld cracking during solidification.

A few weeks ago in a pretty high end fab shop I saw TIG used to blend the weld toe of a big MIG weld. All of a sudden it looks pretty valid.