Strange 304L HX Tube Construction and Failure

[Guest102023]

I attach a few photos of a failed refinery reboiler tube that has me puzzled.

Description: ¾"Ø x 16GA (0.062"), arc welded 304L. Chemical analysis of base metal is in spec, although P is surprisingly high at 0.034%.
Service: Tubeside steam, shell side cat cracker feed (heavy hydrocarbons). Chloride ions reported <10 ppm.

My first thought on seeing the double track on the OD was 'selective corrosion of HAZ', but the microscopy showed different.
- base metal looks heavily sensitized; possible sigma phase. Despite carbon of only 0.02%.
- failure was by a single pit originating on the STEAM side (NOTE: and adjacent tube failed by an apparently entirely different mechanism: 50% lack of fusion in the factory long seam followed by high cycle fatigue).
- the groove on the OD looks less like corrosion pitting and more like a dent.
- aggressive corrosion from the steam side but only at a single pit

Any thoughts?

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

 

[metengr]

How long has this tube section been in service?
This really looks like defects associated with the seam weld and possibly in the skelp. I couldn't tell in the first photograph if the OD crevices are actually dents because I was looking for locally deformed grains along the surface.

 

[Guest102023]

The trim on the skelp was indeed very poor, it was visible on the macro section of the other failed tube. I was surprised to see an arc welded seam on a small diameter tube, I had assumed that a high frequency process would have been used.
I've attached a second micrograph done with a different etchant; I cranked up the brightness to show things better. I don't see evidence of either corrosion or mechanical damage, it looks more like the weld just solidified that way. I will go back to the lab for clarification.

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

 

[Guest102023]

p.s., would the weld seam have been roll flattened? And if so are these grooves just errant tool marks?

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

 

[metengr]

The second micrograph of the seam weld metal does appear to show locally deformed remnant dendrites following the contour of the OD crevice, which implies local mechanical deformation. Regarding the pit on the ID surface, I would suspect an original defect in the tube at this location which resulted in a crevice and subsequent corrosion.

Poor quality HX tubing.

 

[EdStainless]

Stainless isn't HF welded, except for very low quality material, HX tubing is all fusion welded (TIG, Plasma, or laser). That is a very narrow weld, it might be laser (it looks wide for laser but it has the shape).
The lines do look like forming tooling marks, even if they were just grooves they could trap impurities and be a site for initiation of corrosion.
This look very much under annealed. I suspect a very rapid inline anneal that is just enough to say they annealed it, they didn't reach solution conditions at all. What is the elongation in the MTR?
Yes it looks sensitized, did they report passing a A262?
I wonder what the residual delta ferrite in the weld is?

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

 

[Guest102023]

Here's a macro of the weld, it might help you to identify the process. Also a picture of the weld root containing the pit, suggesting local LOF.
metengr, the bundle was installed 2003, replacing carbon steel. There is no MTR or other data about this, only that the client's standard requires seamless HX tube and that these were installed in error.

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

 

[Guest102023]

Weld root

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

 

[metengr]

Looks like long seam TIG to me with poor weld joint fitup with the skelp, and not laser.

 

[Guest102023]

The attached pics show the other tube - poor skelp (I think) and off-centre weld. Off-centre would appear to be the bigger problem.
I believe an intelligent welding system would have detected problems by the effect on the arc.

"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]

I agree, probably TIG, but the shape is very much like laser (hourglass).
The root issue could have been a non-metallic inclusion (polite for trash).
With that narrow of a weld it doesn't surprise me that you found lack of penetration in places.
And it doesn't look like the weld was cold worked as required by A249. It looks like it was just sized from the OD.
It might be good to close the loop on this to pull a few tensiles and check for residual ferrite. Let me know if you want a hand.



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

 

[Guest102023]

Just returned from my friendly local NDE purveyor, who lent me the use of their Ferrit Scope.
I can positively locate the weld, but everywhere I tested the result was in the range 0.1~0.2FN.
The tube geometry is not ideal, but for low ferrite numbers the curvature correction is very low.
I would think that if there was a normal amount of ferrite in the weld the scope would have found it.

EdS, when you say anneal, do you mean solution anneal? Not sure how that could be accomplished effectively onstream.

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

 

[Guest102023]

EdS,
I would like to know the manufacturing process for welded SS tube in detail, hopefully in flowchart form. I am especially interested in the annealing cycle.

"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]

The bulk of the SS pipe and tube made today is inline annealed. The hold times are <10 sec, but the heating and cooling rates are very fast (usually induction) so there should not be any sensitization, but way to short for any diffusion and grain formation in the welds.
The inline anneals work very well for ferritic alloys (439 and superferritics) since these alloys don't contain Ni and diffusion rates are very high.
Material that is specified as furnace anneal or welded and drawn would all be annealed in an off line furnace where hold times are >3 min. Here you will dissolve any preexisting carbides and secondary phases, and you will get some diffusion to break up dendrites (a little) and get some grains forming in the welds.
If the tube is cold drawn then the combined effects of the mechanical deformation and re-anneal will nearly eliminate any weld metal structures.
The specific steps can have some variation. On the weld mill you form, weld, bead condition (roll or forge), and size. If you inline anneal it goes before sizing. Other operations such as straightening and NDT can be done inline or offline.

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

 

[Guest102023]

So the non-existent weld metal ferrite, which I see in the micros as well as the meter, is normal in an austenitic welded tube.
That still leaves the severe sensitization; what could have gone wrong to cause that? Very slow cooling of some tubes in a furnace batch?

"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]

Could have been slow cooling, or too low of a temp.
The ferrite could be very low because of low FN raw material combined with very fast weld speeds.
With a weld that is only 0.020" wide this may be laser welded

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

 

[Guest102023]

I shouldn't think they add filler metal with any process, am I right? Pretty tricky at the weld speeds they run at I'm sure.
Nickel is at 9.2% which is generous, and N is 0.064%, so FN potential in the weld would indeed be low.

"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]

No filler (other than maybe a little N in the weld gas).
What country was this steel made in? You never see that chem here.
The FN would be low, if it is less than about 3 you get virtually zero retained ferrite in the welds.

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