GTAW 304SS to 410SS w/o preheat & w/o PWHT 1

[tc7]

I have the unfortunate job of field repairing equipment that involves 302/304 threaded fasteners that were originally GTAW welded to a 410 base plate (the base plate was originally heat treated to HRC35-40). The attached fasteners need to be removed and replaced.

All guidance I have seen so far indicates that I should preheat the 410 and then stress relieve after welding - neither will be practical or possible in my present circumstance.

I am tentatively selecting a 309L filler as my calculations show a resulting FN of ~10, not bad (I think). This is based on WRC-92 and Dr Kotecki calculations with the assumption of 15% dilution with each base material. Same calculation with 310 welding rod shows an FN off the chart so I assume this indicates a fully austenitic weld will result(not good, I think).

There is no governing welding code for this job, so a full blown procedure qualification is not required.

Any advice/comment/criticism from the board is invited please.

Thank you.

 

[EdStainless]

Can you get any pe-heat?
How much cracking can you tolerate?

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[metengr]

PWHT may not be necessary given the size of the fillet weld on the 410 base material. I would try to locally preheat to 500 deg F. No preheat could result in cracking or at most significant hardening in the weld region. How much carbon is in the 410 SS material you are using?

 

[tc7]

Hello Metengr & Mr. Stainless,

We might be able to get an acetylene torch up there and get a couple of hundred degrees of preheat, don't know if we can allow 500 though.

The carbon is quite low, our analyzer is indicating only .063%

But now my problem has taken a bit of a twist, and it turns out the fasteners are 303 stainless, not 302/304 as originally told to me. The analyzer reading is telling us we have >.300% sulfur in these fasteners.

I am aware that sometimes problems in welding high sulfur steel can be offset by using a high manganese filler, in this case maybe now I should consider using a 316LMn filler ? The FN calculation drops to about 2. What do you say about that ?

 

[metengr]

tc7;
Send out one of the fasteners for chemical analysis to be sure. I would use a propane or MAPP gas torch, not acetylene. The lower carbon version of 410 ss will help to reduce cracking under minimal to no preheat.

If the fasteners are 303 stainless steel, I would use 310 or 312 stainless steel, NOT 316LMn.

 

[tc7]

Gee whiz metengr, I thought you would have been proud of me for thinking of the 316LMn !! am I wrong about the benefit of Mn on high sulfur?

You seem very adamant abot NOT using 316LMn, but the 316LMn fits the PWHT exclusion of ASME B31.1 table 132 footnote for the P-No.6 for A numbers A-8 or A-9, does it not ?

 

[racookpe1978]

Why do you believe you cannot get to a 500 degree preheat?

A hand-held oxy-aceteleyne torch can get the area that hot (you need the weld point, plus at least 2-3 inches around the weld to get to full preheat, and the nearby area to be 200-300 deg F to prevent the heat from bleeding off too quickly.

 

[tc7]

good question racook, I should have explained that. The 410 base plate is embedded in yet another larger base plate that is bolted and sealed with some type of polysulfide material and so the engineers are placing a temperature restriction upon this repair.
Thanks.

 

[metengr]

tc7;
I will give you an A for effort. The 316LMn will have a tendency to cause problems with either hot cracking in the weld deposit or dilution with the martensitic stainless steel in the fusion zone. Normally, it is not used for dissimilar metal welding.

 

[gtaw]

This one should provide for an entertaining read. Please, let us know how this repair goes for you. All I see is one opportunity after another.



Best regards - Al

 

[tc7]

Hey Al! don't be bashful, offer your insight to this problem.


In thinking about this overnight, what would anyone think about this approach: welding a butter layer of Inconel 82 (ERNiCr-3)applied to the 410 at 350 deg preheat and hold the 350 for about 15 minutes and slow cool. Then apply a butter layer of the same Inconel 82 to the 303 fastener and fan cool. PT both butter layers and then fit, tack and weld both together with the same 350 deg preheat on the 410 side and slow cool. (Note: using the Inconel didn't pop out of my brilliant mind, I read the mention of using ERNiCr-3 on the web site of a 410 supplier).

Thoughts any one?

 

[gtaw]

I believe the problem is the HAZ of the 410 stainless. If it is welded without the benefit of preheat, how can one ensure slow cooling? Sounds like wishfull thinking. This material is air hardenable when cooled from elevated temperature. The recommended preheat is 500 degrees F. This sounds like a disaster in the making. I just hope no one is injured as a result of a half hearted attempt at making this repair. I hope someone takes a step back and really thinks about the proper approach. There have been some good suggestions, but there are constraints imposed that run counter to a successful repair.

Best regards - Al

 

[tc7]

Just what is the "problem with the HAZ" that you are alluding to Al ? HAZ will always occur, as you well know, so there are only a couple of means at our disposal to offset the effects of the HAZ: preheat and PWHT. ASME acknowledges, as did metengr, that some circumstances permit elimination of PWHT. On the other hand, I stated above that I can apply ~350 deg preheat and can likely let that creep up to 400, which by the way is what B31.1 guidance indicates. Sure, more preheat is better if one has the luxury of no constraints, but where is your recommended 500 deg printed? Nothing "half hearted" is going on here - a problem was posted and request for help was made, contribute with specifics if you want but no need for you to be so critical.

At this point the bigger concern I have is welding to the 303, not the 410.


Incidentally, here is a very interesting report that pertains to repair welding 410 shrouds on steam turbines and trials were done with ERNiCr-3. Oddly, if I'm reading this correctly they chose to use no preheat but then applied a very substantial PWHT. The report is from the "Gruppo Frattura" (?) who I never heard of before but seem to be dedicated to the study of fracture mechanics.

 

[gtaw]

My point is that the cards are stacked against a successful repair.

AISI 410 belongs to the family of Martensitic stainless steels. They are air hardenable if I remember correctly. The HAZ cooling rate will be accelerated if it is already joined to a heavy plate as you indicated and high preheat is out of the question. PWHT; not possible, so if there is a hardened HAZ there is little chance of mitigating the problem. We haven't broach the need to use low hydrogen welding techniques.

AISI 303 belongs to the family of free machining austenitic stainless steels. High sulfur is a problem which you seemed to be well aware of. What is the sulfur content, around 0.3% or so? That is about 5 times more than the upper limit when good weldability is needed. Manganese will help mitigate the affects of sulfur, but 10 times the sulfur content is recommended when sulfur is on the high side. This isn't on the high side, this is on the very high side.

I don't criticize because one is asking questions, but I think an alternative to welding is encouraged in this case. Reading the scenario painted in this case is like watching a car headed at high speed toward a cliff edge and discovering the brakes are disconnected and the doors are locked.

You've been handed a can of worms. There is very little working in your favor.

The best I can do is say, "Lots of luck, you have my sympathy."

Best regards - Al

 

[tc7]

Understood Al, at first you seemed to be belligerent for the sake of being belligerent. Indeed it's a tough situation but there is a solution and it will be attacked in a thoughtful manner so as not to cause death or destruction. I am holding this job up until I am ready to declare a reasonable course of action to my sponsors.

This is how my plan has evolved as of this morning:
1. butter the 303 side with 308L (instead of the original plan to butter with NiCr-3).
2. butter the 410 side with the NiCr-3 using 350-400F local preheat and hold the 350 for a while then taper down in temperature over the course of 15-20 minutes (this is the slow cooling process that you said I was thinking wishfully about - it works, done it before.)
3. PT everything.
4. complete the weld between the 308L buttered 303 and the NiCr-3 buttered 410 with NiCr-3, again holding 350-400F preheat and slow cool by the process described in step 2. We'll also do this in multiple bead fashion, not unlike a temper bead operation.
5. I'll do some trials of this on scrap beforehand and polish some cross sections to examine for cracks. Maybe I can even get a traverse done before the real work begins just to see what we end up with.

FYI - The process will be GTAW with high purity gas, so almost by definition this is a low hydrogen process.

The NiCr-3 filler I intend to use will have 2.5-3.5% Mn which is right at that 10x sulfur ratio we would like to have plus my 308L butter layer should dilute the 303 sulfur down to a considerably much better ratio.

 

[gtaw]

Never say never.
Low heat input, fast cool for the 303, it might work.

As I said, this should be entertaining as well as educational. I'm never one to say I never make mistakes. That's how I learn. Try it, see what happens, tweak the process and try again.

Make sure you let us know how it works.


Best regards - Al

 

[CoryPad]

Can you explain why you chose 308L over NiCr-3 for the 303 weld? I can see why you are concerned about welding the 303 fastener, not desirable.

 

[tc7]

Yes sir, glad you asked -
In my very unscientific way of thinking, the 308L has a ferrite number >10 and the NiCr-3 ferrite number doesn't exist and the 310 which I was originally thinking about has a ferrite number of about zero or less so I thought the 308L would add to crack prevention concerns plus some ferrite would add to greater solubility of sulfur. It also looks like 309L would serve my purposes as well.

 

[DekDee]

tc7,
In your OP you stated there was no need for a "full blown procedure" as you are not working to any code.
Based on the diversity of replies I would have thought this argument went out the window.
The only way to prove something works is weld it, test it and review the results - the end result is you have a "procedure" that either works or it doesn't.
Cheers,
DD

 

[metengr]

tc7;
I would agree with DD above that you use a mock-up and evaluate the weld region before jumping off of the cliff.