Super Duplex problem

Date 01-11-2010 22:24

Just to brainstorm:
I don't see anything out of the ordinary for chemistry given dilution.
I assume its a plunger type bender?
My first guess is overall embrittlement or low elongation.
What is the tested elongation of the wire?
With plunger benders the first point of stress for low elongation will be the center of the weld. So with overall embrittlement it would make sense to fracture in the center, the first outer fibers to stretch. If you use a wrap around type it will fracture first along the first bending edge. After that it doesn't matter because the stress will be relieved.
If your elongation is less than 20% you can have trouble with ASME IX bending dimensions.
What were the weld parameters?
Heat input?
Did you measure the phase balance?
Can you tell where the fracture initiated? I know you said middle, but toward the root, the cap?
Is it possible it initiated in the HAZ near the root and propagated up the center? 2507 has a min elongation <20%. Given the high elongation of the wire (maybe) and the low elongation of the base metal the base metal may have absorbed the brunt of strain in the early bend angle.

By Jim Hughes (***)

Date 01-11-2010 23:07 Edited 01-11-2010 23:16

js55
25% elongation for the wire. I checked the bender set up and it is correct. It was a wrap around type bender. 160 amps at 10 IPM 25.8 volts, 24000 J. 35 CFH for gas. Dilution can't be much if any of a factor because I took the readings in the center of the weld cap which was approx. 1" wide. Crack looks like it started towards the root. The lab report shows unaffected structure-The ferrite has been etched without revelation of intermetallic phase. The interphase boundries are smooth.

So you don't think the Ti at .324 could not have effected the weld and made it brittle?

Jim

By Stephan (***)

Date 01-12-2010 08:12

Jim,

please forgive me the audacity to ask.

I know it's an idle task to reply on 'theoretical' questions as one has to fix a concrete practical issue. I venture nonetheless to ask, as this remembers me on a problem we had several years ago with a high-strength aluminium alloy, however.

Actually it's three questions, namely. Where might the Titanium may come from, have you carried out a PMI of the parent material and has the lab investigated the crack-surfaces?

And forgive me my ignorance. Does 'Fe' from the PMI literally mean 'iron'?

Just very humble asking, like I said.

Thanks and best regards,
Stephan

By nantong (**)

Date 01-12-2010 11:22

Jin, can I ask why you are welding superduplex with a nickel based welding consumable in the first place? You may have a good reason and i would be interested to know.

By Jim Hughes (***)

Date 01-12-2010 12:12

Thanks for your responces.
1. Stephan, Yes we did PMI the base metal and it was in compliance. and yes it is iron.

2. Nantong, It is a requirement in our spec for a scrubber project we are doing. This client has used ENiCrMo-4 filler material on other projects with success. We orig. qualified a procedure with 2553 flller material with success but the client came back to us and asked us to use ENiCrMo-4.

Thanks
Jim

By nantong (**)

Date 01-12-2010 12:50

Jim, did you do any all weld metal tensiles and have the elongation results? I take it as read you are using a first class reputable consumable manufacturer. I am always a bit sceptical of PMI with OES. I agree with Stephan where does the Ti come from if this accurate. I have had problems with titanium carbo-nitride formation in the past causing cracks, albeit not under these circumstances however you say the micros were clear. interesting! Nb bearing consumables were a no-no but what you are using and the problem you have has no apparent relation.

By Stephan (***)

Date 01-12-2010 18:55

Nantong,

just two words.

Titanium => flux?

Just a thought...

Regards,
Stephan

By js55

Date 01-12-2010 19:17

Stephen,
I can't imagine that even an acidic flux high in TiO2 could possibly impose that much elemental Ti on the weldment.
Also, fluxes formulated for Ni alloys, to my knowledge, are exclusively highly basic.

By Stephan (***)

Date 01-12-2010 21:37

Jeff,

cheers for the reply.

Just even a second back I read the response coming from Jim.

Very, very interesting... to say the least!

Stephan

By Stephan (***)

Date 01-12-2010 18:23

Thanks Jim + Nantong + Jeff!

On a first thought I would have considered Ti-oxides rather than titanium-carbides or similar (.02% C, at least in the filler analysis).

This also, as you've mentioned an unaffected structure and smooth interphase boundaries. Perhaps, so then my humble thought, the crack surface would have unveiled some more secrets.

However, I'm devotionally following this thread!

Thanks again and best regards,
Stephan

By js55

Date 01-12-2010 17:11

NiCrMo-4 is not uncommon for concern in reducing services. The high Ni and Mo will eliminate any concnerns for element segregation. And as you mentioned in another post 4 is stabilized with W instead of Nb so the Nb nitride problem is eliminated, though tons of NiCrMo-3 have been welded on AL6XN with ner'y a problem.

By js55

Date 01-12-2010 17:06 Edited 01-12-2010 17:21

Jim,
I am not leaning towards that conclusion right now. (However, I will concur with Stephen as to the source of Ti and Nantongs concerns about PMI's-I think the Ti number is bogus).
And I don't see any issues with the data you posted.
And Nantong asks a good question with elongation.
Do you have results for E in your tensiles?
The SDSS will have low elongation to start.
You said the cracks appear to initiate near the root.
Is it possible they intitiated near the root in the HAZ?
And if so, do they appear to have initiated on the first HAZ to undergo strain in the bends?

By js55

Date 01-12-2010 17:29

Also, the dilution you are reading at the cap will not be the dilution you are realizing at the root. Especially if your joint configuration includes a feather edge.
What is the joint config.?
Did you run a GTAW root?

By Jim Hughes (***)

Date 01-12-2010 21:32

Elongation is 17%. ASME requirements for 5.34 is 25%. js55 you are correct. After using a tiger paw on the weld and re-doing the PMI the Ti results are 0.98. It must be noted that ASME for this filler material allows 0% so the Ti is still an issue. If nothing else the filler material is out of compliance with ASME which is something I missed when I started this PQR. I should have reviewed the MTR in greater detail. My QA skills need some work.
As far as where the cracks intitiated its hard to tell but it looks like it started in the weld at the root. This is a FCAW procedure so I ground the back side to sound metal and back welded.

Thanks
Jim

By js55

Date 01-12-2010 22:14

I'm guessing the elongation at the root is less than the nominal 17% (ASME bend radii result in 20% outer fiber elongation).
This in reality doesn't necesarily mean you fail at 17%. I don't know why.
Except that, keep in mind elongation is measured from reference marks 2" (or 1") apart and the elongation within that length may not be even. In fact, I doubt it ever is. If the YS in any particular portion is lower than at the root it will take the brunt of the elongation.
What may have caused a reduction in elongation?
I need to do some more research. My memory fails me. But I'm gussing it has something to do with the SDSS diluting into the Ni filler.
Might try low dilution root techniques like wider root gaps and heavier landings.

By js55

Date 01-14-2010 13:42

Let us know if you get final resolution.

By Jim Hughes (***)

Date 01-15-2010 02:32

js55,
I am going to cut a sample out of the test plate and have some testing done. So I am on hold until that happens. The only thing that stands out to me is the presence of Ti. I am researching this to see how this would effect the weld. Again ASME Sec II part C 5.34 does not allow any.

thanks so much for your valued input js55

Jim

By js55

Date 01-15-2010 14:28

Jim,
Well, it doesn't seem like I helped much at all, but we just keep pluggin away.
Add up the posts from everyone, toss in some of your own, and you just might survive it. :)

By pgoswami (*)

Date 01-14-2010 20:37 Edited 01-18-2010 18:32

Jim,

I am follwing this thread for some time and would like to add a few notes,hope this would add some value to this discussion.
SFA-5.34 lists no "Ti" in undiluted weld metal,but per PMI and mill cert some level of "Ti" is there in the weld.The purpose of adding some Ti in the weld is to refine grains and to act as sacrificial element towards fixing "C" in the weld, through forming carbides or carbonitrides. There are many publications from Special Metals on this.

"Fe" in the weld (10.45%) looks high, higher level of "Fe" could reduce ductility and later reduced the corrosion resistance. "Fe" level in filler is within the range of SFA 5.34 ( 6.1% vs 4-7% as per the spec).Controlled welding parameters(lower heat input) may help to reduce Fe dilution in the weld and being a butt weld this would be easy. It may be a good idea to restrict "Fe" to the lowermost limit.

If you could perform any SEM/EDS on the broken specimen, the energy peaks would tell if had been any micro seggregation about the fractured surfaces.

Thanks

Pradip Goswami

Jim,

Further to prior suggestions I added a few documents as attachments on welding with of C-276 fillers in FGD environments.As usual most of the articles are from Special Metals .They heavily emphasize a lot on Inconel-686 CPT products. However a literature search on various types of Nickel alloys (C-22, C-276 ,C-686) indicate that for there could presence of an intermetallic phase e.g "Mu Phase" which significantly reduces corrosion and ductility of weld and alloy seggregation during solidification could embrittle the metals.

In the present case , the duclity and % elongation look very low.It's quite unusual for C-276 weld failure to start @20deg bend angle, because these alloys can easily pass 180 deg bend. The weld parameters as well as the heat input (24KJ/Inch) does not look very excessive.However if the weld puddle is large (single bead, edge to edge) there could be chances of cracking due to improper solidification and possibly excessive "Fe" dilution.

SFA 5.34 does not allow "Ti" in the weld.In the present context it looks "Ti" have ingressed in the weld from flux in FCAW core. As per the the paper Samuel Kaiser (Special Metals-See attached link) Ti, IN Alloy CPT-686 primarily improves corosion resistance. I could not locate any exact reference on "Ti in C-276 weld metal.

Also "Zn" is not allowed by SFA 5.34. The presence of this element would greatly embrittle the weld.

I would liketo suggest a few things:-

1-SEM/EDAX of the failed area as mentioned in the prior email.This would provide a through analysis of the alloying seements including any seggregation (of any)
2- Good metallography/microstructural evaluation around the failed area.
3--Hardness measurement in the weld, HAZ, especially around the failed area would be important.Micro hardness measurement would be good, this could be combined during step-2.
4- Measure the "C" level in the weld --more for customary requirement. Super Duplex contain low C(0.03),Filler (0.02). If "C" in the weld is more than above values, some thing must being contributed by CO2 in the shielding gas. In that case a change to 90/10 or other recomemmended choices by consumable manufacturer may be advisable.

I would look forward to your feedbacks.

Thanks

Pradip Goswami
Welding/Metallurgical Specialist
Onatrio Power Generation Inc.
pradip.goswami@opg.com

Attachment: FGDWeldingMetallurgyIssues.docx (16k)

Attachment: Nickel-Chromium-MolybdenumSuperalloys-FGDSolutions.pdf (890k)

By Jim Hughes (***)

Date 01-15-2010 02:40

pgoswami,
thanks for your input. So are you saying the Ti is helping the weld? Also the publications from Special Metals that you mention,
do you know how I can find them? Also do you think the heat input figures that I posted are out of line?

Thanks
Jim

By Stephan (***)

Date 01-15-2010 02:48

Jim,

forgive me.

pgoswami has - at least indirectly - agreed by writing:

(quote) "If you could perform any SEM/EDS on the broken specimen, the energy peaks would tell if had been any micro seggregation about the fractured surfaces." (unquote)

with the crack surface condition hypothesis.

If the lab would be able to identify some particular traces of segregation of precipitation upon the crack surface, you would have made a step towards greater clarity.

Just a humble thought again.

Regards,
Stephan

By Jim Hughes (***)

Date 01-15-2010 02:56

Thanks Stephen,
I missed that part of his responce. I will take his advise. Thanks for pointing that out to me.

Jim

By js55

Date 01-15-2010 17:07

A couple of things.
I don't believe the Fe issue is going to lead anywhere. NiMo's (obviously with high Mo) such as NiMo 8 (already with max 10% Fe allowed) is designed to weld 9Ni (~90Fe) alloys with very high Fe of course. NiFeCr alloys are common wherein the Fe is already high. Alloys with high Fe such as Alloy 20 or Alloy 800 demonstrate more than adequate ductility. Certainly in excess of the ASME 20% procedure qual min.
Fe is surely known to be problematic in high Cu alloys but, again, though Alloy 20 is not high Cu it has Cu in excess of the PMI stated above and a great deal of Fe, being a SS.
The G Alloys such as NiCrMo-9 and NiCrMo-11 have high Fe as well along with high Cr and Mo. In fact, take a look at the SFA chemistry of ERNiCrMo-10. It is almost dead nuts the PMI stated above.

Then again,,,,,,,,,,who knows? :)

By js55

Date 01-15-2010 17:12

Jim,
If you have a MIC-10 it might interesting to run a hardness comparison of the weld metal and base metal within the tangent of the bend.

By pgoswami (*)

Date 01-18-2010 18:33

Hi Jim,

I updated some of my inputs.

Thanks

By Jim Hughes (***)

Date 01-21-2010 02:20

pgoswami,
thanks for the help. I will review these items.

By Kix

Date 01-19-2010 02:14

Is the right bend radius being used for the test? Sorry, had to ask. Sometimes it's the simple stuff that will kick your butt. ;-)

Kix

By Jim Hughes (***)

Date 01-21-2010 02:22

Yes. One of the first things I checked because they were breaking so fast.

Thanks
Jim

I hope to have some resolve on this by next week.

By pgoswami (*)

Date 01-25-2010 18:25

Hi Jim,

Any update on this interesting topic. Appreciate your inputs.

Thanks.

By Jim Hughes (***)

Date 01-29-2010 02:33

Ok. Here is an update. If the attachment shows up on the post you will see the SEM test. It's obvious that my welding parameters are incorrect. I ran a new PQR with the following welding parameters. Volts=25 Amps=160 WFS=316 IPM=6.5 IPT was less than 250 Deg F. Instead of stringers I ran a weave. Instead of 8 passes I did it in 4. The plate was .500. Heat Input was 39000. Strap bent. These new parameters were given to me by someone else.

Jim

By Jim Hughes (***)

Date 01-29-2010 02:41

I tried to attach a file but could not.
Jim

By js55

Date 01-29-2010 13:32

What from the SEM is the indication?
Did the new straps bend?
What is the latest phase balance?
I would anticipate austenite volume percent going up with the increase in heat input.

By Jim Hughes (***)

Date 01-30-2010 16:27

js55,
for some reason I can not attach anything. The site keeps giving me an error message. I have copied and pasted everything off the SEM report except the pictures. I bent one strap out of the 16" X12" X.500 2507 test plate before I dropped it off at the lab. I did this to save material. If it did not bend I was going to cut the weld out and re-prep and then come back on the forum and wine and beg for answers. :) As mentioned the strap bent with no obvious defects. By contract we have to run a ASTM 923 AB&C. A=Sodium Hydroxcide Etch B=Charpy and C=Ferric Chloride Corrosion test. I will have those results by Friday Feb. 5th and will share them with the forum. If anyone can give me some advise as to why I cant' upload attachments please let me know. The file was small. Again to recap. It looks like it was a welding parameter issue. After welding new plate with recommended welding parameters the strap bent and the other tests are forthcoming.

Thanks again for everyones input. It's great to have this resource.

kV 20.0
Takeoff Angle 27.0°
Elapsed Livetime 30.0

Image 2-1: Matrix

Images 2-2 and 2-3: White precipitates/intermetallics

Elt. Line Intensity
(c/s) Error
2-sig Atomic
% Conc Units
C Ka 0.00 0.000 0.000 0.000 wt.%
N Ka 0.00 0.000 0.000 0.000 wt.%
O Ka 6.70 0.945 9.227 2.505 wt.%
Si Ka 4.02 0.732 0.976 0.465 wt.%
Cr Ka 154.40 4.537 18.925 16.701 wt.%
Fe Ka 73.93 3.139 10.684 10.126 wt.%
Ni Ka 242.50 5.686 47.768 47.582 wt.%
Mo La 89.28 3.450 10.812 17.605 wt.%
W La 5.48 0.855 1.607 5.014 wt.%
100.000 100.000 wt.% Total

kV 20.0
Takeoff Angle 27.0°
Elapsed Livetime 30.0

Matrix composition is typical of C276 weld metal

Elt. Line Intensity
(c/s) Error
2-sig Atomic
% Conc Units
C Ka 0.00 0.000 0.000 0.000 wt.%
N Ka 0.00 0.000 0.000 0.000 wt.%
O Ka 3.61 0.693 7.724 1.800 wt.%
Si Ka 1.71 0.478 0.383 0.156 wt.%
Cr Ka 142.95 4.366 20.072 15.197 wt.%
Fe Ka 47.94 2.528 7.797 6.340 wt.%
Ni Ka 150.46 4.479 31.274 26.728 wt.%
Mo La 239.41 5.650 29.607 41.362 wt.%
W La 10.52 1.184 3.144 8.416 wt.%
100.000 100.000 wt.% Total

White ppt is rich in Mo probably indicating a Ni-Mo-Cr intermetallic

Area near bend fracture showing precipitates rich in Mo. Note presence of precipitates along the fracture boundary

By Jim Hughes (***)

Date 02-11-2010 00:37

Well the bend test failed. I know I mentioned that I bent one strap and it passed. I recieved a call from the lab Monday telling me that three straps bent but one failed and it failed like the first test by breaking in half. I was so sure we had figured this out but now I don't know what is going on. I have to report this to the client and wait for their disposition. If there is still anyone following this thread I am very much open to input.

Thanks
Jim

By js55

Date 02-11-2010 18:14

Jim,
I didn't see if you listed what the phase balance was. But it seems to me, since the severity of precipitaiton is time dependent, that if you are imposing enough heat for precipitation at a level to cause ductility problems your phase balance would reflect this as well.
And so would toughness.
In a sense, there is always precipitation in duplex but it takes a relative great deal to be reflected in mechanical problems.
Seems to me I read some years ago that sigma chi can be present in as much as 2% vol before the mechanicals will be readily effected. Thats a lot.
This has traditionally been one of the arguments against impact testing for material process verification.

By nantong (**)

Date 02-16-2010 07:44

FCAW wire 0.45", on 1/2" material may be too much with interpass temperature you mentioned (remember centre line of the weld will remain at the highest temperature for the longest time). You mentioned precipitates found but i did not see it quantified. It used to be up to 5% acceptable (if mechanicals were ok!). Four runs on 1/2" material indicates a problem. I often find inspectors who run welding procedure qualification tests do not know how to measure interpass temperature, are you sure this has been done correctly? Finally, whose welding consumable and whose base material are you using? This is so important to know. They, especially the consumable manufacturer, should be helping you out big-style on this.