Crack corrosion in stainless steel 3

[prc]

In a fabricated part in a transformer (non magnetic steel), crack developed in prime material (12 mm thick 304 steel)after nearly 6 years. The site is in a thermal coal power station with a steel blast furnace nearby. The crack started near to a weld (stainless steel to stainless steel) but spread in to prime material for nearly 10 centimeters.Is this the knife line attack ? What is the short term and long term solution?

 

[desertfox]

Can you post a picture of the crack and is it above or below the insulation media within the transformer?

 

[EdStainless]

What is the environment? Ambient air both sides?
What max temperature would this see?
How humid is the local?

What NDT was done at fabrication?
What is the residual ferrite in the welds?

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

 

[prc]

It is a bushing turret.Below transformer oil at 60-70C and top air. Bus duct is terminated on air side.Yes ambient humid and pollution from steel mill and coal thermal station.NDT was not done.Three bushings on a common turret.Cracks on either side of central bushing. One crack starts from welding, other middle of the two bushings,away from any weld.Please see the attached photos of two cracks.

 

[metengr]

The surface is not clean for detailed inspection but as a rough approximation these look like fatigue cracks emanating from the openings in the shell. It could be thermal or mechanical. I would perform surface NDT.

 

[EdStainless]

Do some careful inspection, If there are a lot of fine cracks then it might be CSCC. If there are only the major cracks I would suspect thermal or mechanical fatigue.
I would suspect that the defects at the origin were there in the original fabrication.

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

 

[desertfox]

Hi

My ten pence worth is its not very clear for me to see the cracks sufficient to say what they are and the other posters might well be correct, however there doesn't appear to be any leakage of oil, that makes me wonder because the highest stress occurs on the inside of pressure vessels normally, so if it were cracks due to fatigue or thermal cycling I would expect the crack to grow from inside the tank to the outer side of tank and as a consequence a leakage of oil.

Also you mentioned that no NDT was carried so how do you know those cracks haven't been there for a long time?[highlight #FCE94F][/highlight]

 

[prc]

I have a clear picture of the crack after removing the paint. It is a continuous crack with oil profusely coming out,one big drop every second.What type of surface NDT will help? What is CSCC? This is in service for more than 6 years and hence no defect is suspected at the time of fabrication.After fabrication the shell is subjected to vacuum and pressure test. The 12 mm plate has a copper shield welded to it inside.3mm sheet is spot welded to stain less steel at a pitch of 150-200mm.

 

[EdStainless]

CSCC is short for chloride stress corrosion cracking, and this not what you have.

I see a couple of possibilities.
1. Significant weld or base metal defects at the time of manufacture that allowed cracks to propagate, driven by thermal cycling.
2. Crack initiated by the welding of the Cu to the SS. The molten Cu will cause liquid metal embrittlement of SS. And then over time with cycling these cracks grew.

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

 

[prc]

Thank you Edstainless. The copper plate is spot welded to stainless steel by drilling 12 mm dia holes at 200 mm pitch. So it is a surface contact weld and not any buttwelding.To detect any such defect in base metal what is the cost effective NDT method used ?

 

[desertfox]

Try this link

http://en.m.wikipedia.org/wiki/Ultrasonic_testing

Alternatively you could use a spray penetrant however if the cracks are growing from inside tank to outer side the dye pen won't help.

 

[Sammytotheh]

Based on crack propagation path and the weld source, it appears that the welding of the component sensitised the material allowing for Stress corrosion cracking CC to occur. Despite it not having lots of fine cracks, this only eliminates one type of SCC which is inter granular SCC.
If this crack was present in a component in a marine environment then yes i would say CSCC but the lack of Cl ion in this environment eliminates this.
Instead it is quite possible transgranular SCC.

Either way, the cause of this originates from the welding. A post weld heat treatment is needed to dissolve the carbides precipitated during the welding in order to stop Chromimium depletion at the weld.

 

[prc]

But the crack is near to weld only on one side (as seen in clear photo) but on the opposite side of bushing it is away from weld, some where between the two bushings.What is the solution for Chromium depletion? 304 L ?

 

[EdStainless]

If you are welding without post-weld anneal you should be using "L", preferably something less than 0.015% C.
How magnetic are these welds? Probably very magnetic. That will cause more stress over the life of this unit.
I believe that there were defects in the original structure, and that they eventually grew into these cracks.
Buy better quality steel, inspect the fabrication (NDT), and pay more attention to details.

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

 

[AndrewFinAustralia]

Hi prc,

Blast Furnace nearby = sulfur in the air, as well as zinc fume. (working on the furnace was always a great cure for a head cold)

I'll completely back Ed and Metengr, as well as the other replies. You would be well advised to call in a local metallurgical consultancy to take a carbon replica of the crack, particularly if this is a critical piece of equipment.

In particular, you want to know if it's cracking along grain boundaries (assuming austenitic stainless,) or if cracking is intragranular.

Is this component under hoop or tensile stress? From the photo, which I can not see clearly, the quality of the cleanup on the original weld would maybe have gotten my *bottom* kicked when I was trained as a basic welder. Potentially look to a local undercut of the parent metal where the crack started also, unless it's shadowing of the image.

What is the oil - transformer oil ? If so, does it contain chlorine? (or other halogens, yet another attack mechanism)

This would be one I'd refer to a local consultancy - could be one of a number of attack mechanisms, crevice corrosion may be an issue.

First base but - spend USD20 on a dye penetrant kit and learn how to use it - key rookie mistake is to use a lot of developer, don't drown the crack, you only need a light coating. Use this kit to review growth of crack monthly, and potentially drill ahead of crack tip - once again - get your local consultancy to advise.

I'd be looking to repair or replace next maintenance window according to pre-determined specifications for the parent metal/copper. As Ed said, you don't want to melt the copper. I'd also be getting your consultant metallurgist to do a full NDT of the entire weld in the tank - would you bother with repair if there were 10 of these waiting to happen, 20? 50?. At what point would you consider replacement, especially if the metal's already sensitised.

Regarding NDT, X-ray would be best, but this requires draining the tank which would be unlikely to get the goahead. NDT (ultrasonic with a skilled operator preferably) or Dye Penetrant would be best.

Regards,

 

[prc]

Thank you Andrew for the inputs. This is a proven design and fabrication that I have been following may be, for the last 40 years and in service at various power stations. Oil has no chlorine and corrosion from that end is ruled out.

Please see the photo that I put on 22nd April. It is clear. Paint was removed for repair and the crack is clear. The nature of crack is not matching with inter granular crack pattern as seen in literature.May be experts can comment. But please see the color change at the maximum crack region. Can it be defect in material? Why it took so long to come out? Please note it has not started from any weld,but occurred in base metal.If it is from external corrosion,will not the paint protect it? But it was said the paint at this position was peeling off easily. The temperature at this position during service is less than 60C.

Did contact local metallurgists,but could not throw much light.Patch welding is advisable ? Or should we seal with metallic epoxy putty?

 

[AndrewFinAustralia]

Hi prc,

I can't shed any more light on this unless I was there - it'd need closer inspection. Try a strong Rare Earth magnet and see if it sticks near the weld - this'll indicate the presence of ferrite - looking at the photo, if you have created ferrite, it'd be approx 6-8mm from the weld pool, maybe a bit closer. (not a foolproof test for sensitisation [edit], by the way)

The grain size of the austenite is approx 0.1mm to 0.15mm, so inter- vs intra-granular fracture is hard to pick by eye.

The new and intriguing bit of info is the fact this is a tried and tested design. Looking at the amount of weld metal, the base metal was fairly well heated with the size of the weld, I'd still stick to weld sensitisation as a possible *guess* but to comment further, I'd have to see it in person. I was a plant metallurgist/process engineer in heavy industry for a number of years.

From your description of the design - the copper internal skin makes welding problematic as you wouldn't want to dissolve the copper into the weld pool or diffuse it into the surrounding metal by accident.

As you hint at, you're looking to stop the leak by whatever means possible at the moment, and I'd also monitor the growth of the crack. I'd put no more heat into the tank metal/weld it at the moment. In my experience, you'd need something with a bit more flexibility than epoxy to fill the crack, but anything to stop the leak would be good at the moment as a leak that large would (I'm guessing) mean shutting down the transformer frequently to top up oil.

Most telling is that the crack needs tensile stress to grow that large - whatever repair you do would have to take the tensile stress into account.

Sorry I can't be of any more help - a carbon replica is what's needed to give you the answer in my opinion.

 

[mrfailure]

You really need to understand the nature of the crack, as some of the possibilities listed above mean other cracks can develop even if you can repair. Is it feasible to cut out a section containing the full length of the crack including the weld? If yes, you can then have metallurgical failure analysis performed to first identify where the crack started by examining the opened fracture lips, followed by identification of crack initiation (and propagation mode) to help establish the cause. From what I see, I cannot give you any good comments at the moment because it is not even clear where the crack first began. Good luck!

 

[prc]

Unfortunately we cannot cut a piece from the failure point.After a couple of months later,it will be possible for us to remove painting once again and have a close look of repaired area.I thought of a close look of area (with a magnifying glass) for any spider net like cracks near to failure area.With a magnet see the relative magnetism at crack area and other areas. Cut a small piece from the edge of plate. Have a look from inside of tank to see the nature of crack -from inside to outside or max crack width at inside or outside. Condition of copper weld on inside surface of plate at the area of crack.Using any mobile NDT in-situ inspection of the plate from top.Any thing more can be done?

From the look from photo can we not conclude that the crack started where the width of crack is maximum?

 

[69Firebird]

Xray is not a good choice to detect cracks. It will find only a small percentage of them. It is too dependant on orientation.