welded inconel 625 sheet fatigue/static strength 1

[madmonky]

Long time lurker, first time poster here. Hoping someone can shed some light on my problem...
I have (2) thin 0.063" thick inconel 625 sheet materials welded together using AMS5837 filler material with a complete penetration corner joint. The weld will be thermally loaded as well as put into low magnitude bending.
I am looking for a way to analyze this joint for both fatigue and static strength. I am having extreme difficulty finding published data on the allowable static stresses (or design strengths) in inconel welds. I am also having difficulty finding any kind of s-N curves for welded 625 inconel sheet. I am more worried about the fatigue analysis. What would you recommend i do? Essentially i need the data found in the AWS D1.1_ Structural Welding Code, only for inconel instead of steeel.

 

[TorsionalStress]

Maybe this link could help you!

http://www.alloywire.com/inconel_alloy_625.html

 

[madmonky]

Thank you very much for the link TorsionalStress. Even though i have a differenct AMS spec filler material, i guess the properties listed would be a good approximation of the static strength of the weld. Do you have any ideas about the fatigue portion of the problem (lack of s-N curve)?

 

[TorsionalStress]

Try these links,

http://www.hightempmetals.com/techdata/hitempInconel625data.php

http://www.ascgenoa.com/main/newsletter/9/

[2]6-07_AIAA_2007_2381-metal-Fatigue.pdf

Hope they help!

 

[madmonky]

I thank you again for the information TorsionalStress. The pdf that you provided has s-N information for simple inconel 625 plate/sheet. I was looking for something that addressed the fatigue life of the actual weld. My understanding is that a weld is essentially a casting with an inherent dendritic structure, as opposed to the relatively uniform grain size of the unwelded part. I dont think i can use the s-N curve for the unwelded sheet, as the welded area will certainly have lower ductility than the parent material. Or maybe that is just how inconel welds are analyzed, by assuming the weld will conform with the fatigue properties of the unwelded sheet material...i dont know? Can anyone provide with certainty how fatigue is addressed in inconel welds?

 

[EdStainless]

Do the stress analysis first. The stress in the welds should be well below the stress in the sheet. If so then you don't need to worry about the fatigue properties of the welds.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[madmonky]

Hi Edstainless. thank you for your reply. I am wondering why you say the stress will be lower in the welded area... The weld is in bending, and is at the area of highest moment. Lingaiah's Machine Design Handbook recommends 6M/lh^2 (l=weld length, h=weld width) for the stress in this type of joint (pg 12.12). This compares to the traditional 6M/bt^2 for traditional metal square section bending stress. The weld is complete penetration, meaning h = part thickness. This means that the stress equations are equivelent, meaning that the higher stress will be in the weld, not the parent material because it has a higher moment. Regardless, i am curious how traditional inconel welding fatigue checks are done. It doesnt seem that there is data out there to do it.

 

[TorsionalStress]

Check this link:

http://www.specialmetals.com/documents/Inconel alloy 625.pdf

 

[madmonky]

This is great reference info Torsion stress. It gives me a very good idea of what the weld is capable of. Unfortunatly I do not think it would be accepted as certifyable data. It states at the beginning of the document:

"The properties given in this bulletin, results of extensive
testing, are typical of the alloy but should NOT be
used for specification purposes."

 

[dhengr]

Madmonkey:

TorsionalStress.... has done an excellent job of pointed you in the right direction and given you considerable info., in considerable detail. Undoubtedly, you could find some more basic material properties info. from other material suppliers; and possibly some S-N curves for polished virgin material. Check with your filler metal supplier for info. on that exact material, and the welds produced by it. Because that info. is just as important in addressing your problem. Otherwise, the only thing he hasn’t done is run the tests, for you, to produce the S-N curves for your exact detail, loading conditions, welding procedures, etc., all of which you have not done a very good job of describing in enough detail.

You can infer all kinds of things (maybe good approximations) from existing S-N curves for various materials, and similar details or detail dimensions, etc., we do it quite often, but these are just educated engineering judgements unless you test your own detail and conditions; even down to such details as the orientation of the rolling grain in your sheet stock. You have to make a determination of the number of cycles you expect to see, then run a bunch of samples to some higher number of cycles, or to failure, so you start developing the S-N curve for your condition. I suspect you would be very lucky, indeed, to find curves someone else had developed which fit your conditions exactly. Also, we use average, actually minimum, mechanical properties for materials all the time in design.

 

[madmonky]

You are absolutely correct dhengr. I did not do as good a job as i should have done in describing the problem. I am immensly grateful to TorsionalStress for his input. Thank you both.

I guess I was incorrectly thinking that there were established fatigue s-n curves for inconel just as there are for Steel in the AWS D1.1 Structural Welding Document. From your reply it appears that this is an incorrect assumption.

thank you for your time. I apologize for the lack of clarity in my problem statement.

 

[EdStainless]

If the weld stress is higher than in the plain material then your welds are in the wrong place. Why would you design something with the highest stress in the location most likely to have a defect?

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[TorsionalStress]

Can you use a backing at the welded joint?

 

[madmonky]

I agree with you EdStainless. The weld placement is terrible and unfortunatly was locked in by manufacturing long before i got ahold of this. I only wish i were able to change it. The only positive thing about the configuration is that the max stress is relatively low in the part. At the weld location fmax~8ksi. average stress=4ksi. varying stress = 4ksi.

Torsional stress, weld backing may be possible on this joint, but could not be permanent. Does weld backing provide any benefits to the material properites of the welds? What would be the benefit of the backing?

 

[dhengr]

And, we still don’t know what the detail looks like, or how it is loaded, or all the stresses and their orientations, etc. The back-up bar left in place is often a new stress raiser, not a good thing. I suspect the point was that you want to be sure you have a complete weld without any defects at the root, which often happens. Grind or gouge the root side of the weld and reweld it so you are certain you have a complete, defect free, full penetration weld. If this is a corner, a reentrant corner, a fillet reinforcement on the inside would improve the geometry w.r.t. some stress orientations. Grinding the welds smooth will improve the fatigue resistance. Your problem is highly geometry, stress raiser, and defect sensitive, and anything you can do to minimize these is an improvement. If you could get in their to remove a back-up bar, you can probably make that clean-up and reinforcement fillet.

 

[TorsionalStress]

BACKING STRIP: A piece of material used to retain molten metal at the root of the weld and/or increase the thermal capacity of the joint so as to prevent excessive warping of the base metal.

 

[madmonky]

Thank you Torsion Stress. That is a very good idea, especially with the thin sheet and high temps required for welding nickel.

Here is a cross section of the weld

http://postimage.org/image/2afkz24h0/

I believe i may have found a solution to my problem of fatigue analysis. Norton's Machine Design has a very good section on estimating fatigue strengths of materials not covered by testing or published data (see pages 324-337). The only thing lacking in the analysis is a formula for estimating the starting theoretical fatigue strength of inconel welds. Steel is shown as having Se=.5Sut, and iron/aluminum/copper=.4Sut. If I am conservative and just assume 0.4Sut for the theoretical starting fatigue strength of inconel, use Sut for the inconel filler weld material (and not the parent material), use "as forged" for the surface factor, and then factor in all of the welding correction factors listed, I think I should have a good conservative approximation (that can be documented and referenced back to published methods). Does anyone see any flaws in this approach?

Interesting that I found various sources saying that inconel DID and DID NOT have an endurance limit :

those that did: (

http://books.google.com/books?id=Y0...mits for monel metal" inconel nickel&f=false)

&

http://www.egr.msu.edu/classes/me471/thompson/handout/class13_2005SFatigue.pdf

those that did not: (see Machine design, Norton, pg 312).