Is 308H sufficient for 304H in 1100F in SO2 service or should we use something else? 1

[jeevesme]

All,

We have a vessel that has had 4 repairs in the same area and just cracked again. We have repaired it using 308H filler for the past repairs. Should we use an Inco or is 308H fine?

Vessel Information:
Vintage- 2005
Head Thickness- .315"
Head Material- 304H
Design Temp- 1250F
Design Press- 14.2psig
The head was made by welding two plates together, then forming it.


Operating Information:
Service- SO2/SO3 Converter
SO2 comes in from the top 24" welded piping at ~1100F at pressures of ~13psig to be catalyzed via exothermic reaction. The top head should only see SO2. However, it is possible that it sees small amounts of SO3 from time to time.

Repairs:
2009
-The weld that attaches the piping to the top head cracked.
-Likely from the piping orientation, it was creating a cantilever at the joint, as there were no supports.
-The crack was ground out and rewelded from external only.
-Filler metal used unknown.
-Inspection & Testing unknown.


2017
-The weld that attaches the piping to the top head cracked
-The weld that made the head a single plate cracked.
-Likely from the piping orientation, it was creating a cantilever at the joint, as there were no supports.
-The cracks were ground out and rewelded from external only under nitrogen purge.
-308 Filler Metal used
-Preps PT'd, Roots PT'd, Finals PT'd

2018
-The weld that made the head a single plate cracked.
-Likely from the piping orientation, it was creating a cantilever at the joint, as there were no supports.
-The crack was ground out and rewelded from external only under nitrogen purge.
-308 Filler Metal used
-Preps PT'd, Roots PT'd, Finals PT'd

2019
-Pipe support spring can was installed with the SO3 Cooler Upgrade Project USIC.014386

2021
-The weld that made the head a single plate cracked.
-Unknown cause. Possibly still from the piping support issue. Maybe Spring can was not properly designed or installed.
-The crack was ground out, drill-stopped, and rewelded from external only under nitrogen purge.
-308 Filler Metal used
-Preps PT'd, Roots PT'd, Finals PT'd, Final had UTPA performed on it

 

[EdStainless]

A 304 vessel seeing these temps will move a lot, getting the pipe system and supports right can be very difficult.
The trick to handling thermal expansion and stresses is not to make it stronger but to make it more flexible.
How confident are you that they have actually been grinding out the entire crack?
When I hear people say that they just ground deep enough to get the crack what I here is 'we ground out most of the crack".
Are the cracks in the weld, at the toe, or in the HAZ?
Going to a higher Ni filler helps in some ways but it can cause other problems.
This is a design/installation issue.
I would hate to be trying to make the fifth weld repair on highly sensitized material.
The 304H and the welds will have very little ductility when they are cold, so start-ups and shutdowns are likely where most of the damage occurs.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[TugboatEng]

At only 13 psi it shouldn't be too much trouble to isolate the vessel from pipe strain with metal bellows expansion joints.

 

[jeevesme]

@edstainless,

I'm fairly confident that we are getting out the cracks entirely. If we don't get a good PT, we go until it's basically 1/16" opening making it a root pass.

The cracking has always been in the center of the weld. I was very careful with the welding and NDE (i even added PAUT for the final weld)the last time because I did not want to have to weld on it again.

I double checked the calculations for the support done by a 3rd party engineering firm and they seem to be good. We've checked the measurements of the spring can in the cold position as well as the hot position. There is 5.5" travel. The calculations show up to 7" travel may be possible. Would 1.5" cause an issue like this?

 

[EdStainless]

More likely the issue is the direction of the loading; it is probably a bending and not just the weight.
What FN is the filler?
What is the measured ferrite in the welds?

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[jeevesme]

I'm not sure what the FN number is on the filler. I know they use 308h Lincoln wire.
I am trying to get the weld(s) ferrite tested to see what they are.
I am inclined to say it is the design as well. Each time it has cracked, it has been in the throat of the weld in the same place.

 

[0707]

We have used 308H filler metal with good results in the past in FCC flue gas lines service. The ferrite number before and after welding shall be controlled.

 

[jeevesme]

What Ferrite numbers should we be at? Is there a range or specific number?

 

[weldstan]

Do you see any distortion at the crack sites? Like small sinking of the head at the nozzle attachment or at the plate to plate seam weld near pipe supports? Check the carbon content of the original welds as well as the ferrite number.

 

[EdStainless]

You need enough ferrite to suppress hot cracking of the welds, but not so much that the microstructure becomes a mess with long term high temp exposure. One issue with H grades is that the carbon is an austenite stabilizer (reduces ferrite) and it is easy to end up with very low FN (<3) and have weld cracks as a result.
This is case where it is good to check both the calculated value and measure the actual.
I would think that a range of 5-10 is suitable.


= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Guest102023]

When is a crack not a crack?
When it is an incipient crack. Not detectable today but fully developed by next inspection.
This is how some degradation happens.

Some ferrite is normally considered a must but has a downside in high temperature operation, namely sigmatization. If repeat failures are happening in younger weld metal as you say, then ferrite might actually be part of the problem.

Sound, ferrite free deposits are obtainable with 310 and 320LR fillers.


"If you don't have time to do the job right the first time, when are you going to find time to repair it?"

 

[EdStainless]

This is a fine line.
You likely can find 304H or 308H filler that will get you near 0 FN, but risk weld shrinkage cracks.
You can add more Ni (309, 310, 320) and get good ductile welds with near zero FM.
But these higher Ni welds will have quite a bit higher thermal expansion than 304 and that will impose additional stresses.
You need to cut samples out and have them analyzed.
Are your failures from lack of ductility (sigma)?
Are they from thermal fatigue?
Are they from overstressed welds?
Each of these will have a different response.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Guest102023]

You’re exactly right EdS, it’s a fine balance.
You don’t want the weld overpowering the surrounding base metal, and differential CTE introduces a new source of stress.

I have found that particular SS classifications across different manufacturers deposit consistently similar FN. E308 around 8 and E316 around 6.

"If you don't have time to do the job right the first time, when are you going to find time to repair it?"