Not logged inAmerican Welding Society Forum
Forum AWS Website Help Search Login
Up Topic Welding Industry / Technical Discussions / Welding lean duplex
- - By john baxter (*) Date 08-05-2007 21:49
I am presently having problems with welding procedures for a lean duplex. The material is UNS S32003 and is welded with standard 22%Cr consumables. Procedures have been qualified with SMAW, GSFCAW and SAW. The problem is the ferrite content being too low. Specification requirement for the weld metal is 35% minimum but the SMAW and the GSFCAW procedures have given results of 18-24%. The SAW procedure gave results of ca 40%. The SMAW & GSFCAW procedures were qualified on single V butt welds 7.9mm thick and the SAW had no bevel just a  square edge. Clearly the high dilution from the SAW on the sqaure edge has given the weld a high proportion of base metal hence the higher level of ferrite. The manufacturer of the plate states that acceptable results can be achieved using the 22%Cr consumable. These procedures have been welded twice and tested in two different labarotaries but the minimum levels of ferrite cannot be achieved. All other tests, tensiles bends, hardness, charpies and corrosion tests were acceptable. Procedures were welded with correct interpass temperature and heat inputs. If the weld consumable and the base metal is plotted on the WRC diagram and compared to the line for the welding consumable with a 22%Cr duplex base metal it can be seen that as the chemistries of the 2003 and 22%Cr duplex are so different the effects of dilution must be fairly pronounced.
Has anyone else had experience of welding this material or could anyone shed some light on where the problem may lie. 
Parent - - By chuck meadows (***) Date 08-06-2007 01:43
John,
  I am assuming you are using a 2209 filler metal. If you will check in ANSI/AWS A5.4, you will see where the allowable nickel contant is higher than any other process (5.9 and 5.22) SMAW will normally give a ferrite content of around 30-35% at best. This is because ASME has determined that the lower ferrite % is less important than the approximately 70% austenite. The higher austenite will give added strength to counteract the oxidation in the flux of the SMAW process. So, 30-35% ferrite is quite normal for the SMAW, as you will be lucky to get any higher.

  Speaking from experience, you should be getting around 40% ferrite when using the gas shielded FCAW process, and about the same with the subarc process. You need to be very careful about keeping your interpass temperature below 300F, and your heat input should be in the range of 0.5-2.5 kJ/mm (12.7-63.5kJ/in) If you are not careful, this will definitely affect your ferrite percentage. One other thing that will help you when using the SAW process is to use a chromium enriched flux. This will replace the chromium you lose across the weld arc. In the SAW process, it is quite normal to lose up to 3-5% of the chrome across the weld arc. Since chromium is a ferrite former, it is important to replace the chrome by using a chromium enriched flux. The only process where ferrite is normally a problem is the SMAW process.  
Parent - - By js55 (*****) Date 08-06-2007 14:34
A couple of things here building on Chucks analysis. You first of all need to open a dialogue with your customer and hope someone there really knows what they are talking about. In line with Chucks statement about max ferrite for SMAW you may need to request a reduction in your spec somewhat.
Recognizing that I don't know your service it would seem to me that if your mechanicals are good with a lesser ferrite, then the 30 to 35 Chuck speaks of is plenty to accomodate SCC. And remember, duplex solidifies as ferrite first and then tramsforms to austenite. If you give the transformation more time you will get more transformaiton. In other words, find a wya to slow down the cooling rate.
You probably don't have the exact chemistry involved in your procedure qual and duplex is very chemstry sensitive when it comes to phase balance. Essentiially the very reason you need to qual a procedure.
Parent - - By chuck meadows (***) Date 08-06-2007 21:38
I agree. John should write a Letter of Variance. I just got a reply from Fluor amending the FN of Duplex ( Lean, Regular, or Super Duplex) when using SMAW. It appears John is only "out of line" with the SMAW process and that can definitely be amended. John, all you need is a copy of ANSI/AWS A5.4 to show your customer the higher allowable nickel content. Then, I can send you a paper I wrote a few years ago addressing this very issue of ferrites in Duplex Stainless Steels. That is all it took to change Fluor's specification.

Chuck 
Parent - - By john baxter (*) Date 08-07-2007 20:17
The 35% minimum ferrite content is being enforced to reduce the risk of sulphide stress corrosion cracking so no waiver will be given. The problem lies with both SMAW and GSFCAW processes. The consumables are all 2209 rutile and the heatinput was less than 1.5kj/mm and a maximum interpass of 100C. All text book stuff. I know that changing to the basic coated SMAW will make improvements (a problem since the consumables have been purchased) but is there any way to improve the GSFCAW or is that it? I could try forced air cooling but this would be difficult to control in the field and as has already been stated the factor of greatest influence is the chemistry. I have qualified many procedures on 22%Cr and 25%Cr duplex in the past and never had a problem meeting this requirement however this was always on pipe-work which was welded with the GTAW process to achieve optimum results. The lean duplex is for tank fabrication so this rules out this process.
I appreciate your advice.
Parent - - By chuck meadows (***) Date 08-07-2007 21:35
I don't know what to tell you, John. Speaking strictly from using the Avesta 2209 consumables, I will say this....the fabricators I have personally worked with have not had issues with the ferrites of the gas shielded FCAW. As a matter of fact, the ferrites are in the range of low to mid 40's. This was also on plate, or tanks, including pipe. The only issues have been with the electrodes, which are designed to give ferrites of around 30FN in the undiluted weld metal. Naturally, diluting with a base metal of approximately 40-50 ferrite will slightly raise the ferrites in both welding processes. Changing to a basic coated electrode will give you slightly better impact values, but will not improve the ferrites. It is what it is. The electrode ferrite content does not surprise me, but your excessively low ferrites in the FCAW process does somewhat surprise me. And, this has been observed on both pipe and plate.
Parent - - By john baxter (*) Date 08-07-2007 23:53
Chuck?JS55, I have tried to keep things simple as I am not very quick at typing but it is a bit more complicated. The procedures are for the welding of the containment walls of a series of counter current decantation units, part of a new nickel extraction plant in New Caledonia. The original procedures were welded in the States with Metrode basic 2209 SMAW Electrodes and ESAB SB 2209 cored wire. The test results gave a ferrite content (point count method) just below the 35% requirement. I am the Welding Engineer on site representing the Client however I am not the owner of the Engineering Specifications. Any deviation to the specifications is at the discretion of the design department. Despite the fact that the procedure tests all passed (except for ferrite estimations) the decision was taken by the design dept. to re-run the procedures. This was against my reccomendations as it is accepted or even promoted by organisations such as TWI/IIW that due to the inherent errors in ferrite determination these results in effect should not take precedence over the other tests if they are successful.
The procedures were re-run in the Phillipinnes under our witness (where the major pre-assembly work will be carried out) and the test coupons were sent to Singapore for testing. Unbeknown to myself the Contractor (a bridge and iron company out of Chicago) changed the consumables for some reason from a basic to a rutile SMAW electrode and changed the flux-cored wire from ESAB to Alloy Rods. Chemistry of the rutile and basic SMAW electrodes may look similar however the manufacturer of the electrodes has confirmed that there will be a significant drop in Ferrite content due to an increase in nickel content to offset the TiO2 coating. The contractor has stated that Alloy Rods Inc cored wire and the ESAB product are the same thing (through being part of the same company) however when I compared the data sent the wires seemed a bit different. The nickel euivalent of the ESAB product comes out about 13.5 whilst the AR consumable calculates less than 10.
Procedures were welded within adequate parameters however as I have stated previously the test results from Singapore showed good mechanical and corrosion results (again) however the ferrite was way down compared to the procedures done in the States (understandable for the rutile SMAW electrode but from the lower nickel equivalent of the Alloy Rods cored wire I would have expected the ferrite to be higher). There was also an anomaly with the hardness results from Singapore (above 330Hv10 in the base metal) so I asked for the samples to be sent to another lab for another opinion. Point counts have been done so far in Australia and they confirm the ferrite estimation (nothing else done so far).
As the manufacturer of the consumables produce well tried and tested consumables I thought that the material (AL 2003) may hold the key as I believe it is a relatively new product which is promoted more as a beneficial replacement for 316 (due to weight saving) rather than as a replacement for the standard 22%Cr duplex.
I appreciate, again, your advice.
Parent - By js55 (*****) Date 08-08-2007 13:51
First of all john, I need to apologize. I don't know what the hail I was thinkin. The primary thing, IMO, that needs to be done is INCREASING the cooling rate to prohibit the transformation of ferrite to austenite. Clearly you recognized this as evidenced by your forced air cooling statement, but were perhaps to polite to correct me.
I understand your 37kj/in and 212 interpass is low, and under normal circumstances would be well adequate. But I'm thinking this isn't normal circumstances. And the response from engineering that somehow 4% or 5% ferrite is critical to SCC is revealing. Not about the service "in my opinion", but the people responsible for the engineering. After all, continued testing is not going to be at their expense. It would be interesting however to see if they have definitive data stating the criticality of their ferrite % choice.
The way I see it you really only have two choices; change the filler (and thereby the chemistry) or speed up the cooling. Thats really all there is.
And here is another issue that could perhaps be used as ammunition. By speeding up the cooling rate too much N does not have time to diffuse to the austenite and can form Chromium nitrides(CrN and Cr2N), which are detrimental to corrosion as well. 
No, it is my belief, that without definitive data proving the criticality of the 35% ferrite standard (which I believe they don't have, and have created the requirement right out of the duplex handbook), the responsible engineering and metallurgical thing to do is reduce the requirement for ferrite just a little. Thats always been the problem with duplex. Its a window. And going too far one way to avoid problems from the other just creates a whole new set of problems.
And one other thing, if you run your heat inputs too low, what is going to happen to your HAZ? The chemistry of the BM is not designed to handle that fast of a cooling rate. You might find your weld metal ferrite rising, but your base metal ferrite will rise as well. Perhaps to a level that is too high. I am assuming of course that your spec has a top side as well as a bottom side. Have the engineers given any thought to service difficulties if the HAZ ferrite is too high? Your weld metal ferrite might, I say might, increase by 4% or 5%, but your HAZ ferrite might increase by 15% or 20%. Or more.
Parent - - By chuck meadows (***) Date 08-08-2007 21:09 Edited 08-08-2007 21:12
John,
  This is too little, too late, but I always recommend addressing the ferrite issue anytime 2209 stick electrodes are used on a job. Any time I'm asked to review a bid package, I always recommend the fabricator either directly talk to the Owner or some way of clarifying the issue of the almost certainty of lower ferrites when using stick electrodes. This can definitely eliminate a lot of cost (and grief) down the line. It should be perfectly clear to both the fabricator and the Owner of this issue. Of course I'm preaching to the choir in this case. But, a little communication can go a long way. P.S. The Avesta "How to Weld Avesta 2205" states that a minimum of 25% ferrite is adequate for strength and corrosion resistance.
Parent - By john baxter (*) Date 08-08-2007 23:08
Thank you Chuck for taking the time to assist me.
Parent - By john baxter (*) Date 08-09-2007 14:53
and JS55 too.
Parent - - By john baxter (*) Date 08-09-2007 15:14
JS55, I don't really disagree with our Engineering Department's statement about 35% minimum ferrite content to avoid SSCC as I think the optimum figure is ca. 40%. The problem is I have to take it as gospel from the design guys that the service is as they say. HAZ ferrite content was just over 50% and spec. allows 70%. All mechanicals and G48's were well in so there would be no problem with a slightly higher ferrite content in the HAZ. Material is thin (7.9mm) so I wouldn't think that too rapid a cooling would be encountered. You mentioned changing the filler but what could be used?
Duplex welding consumables are over alloyed with nickel to maintain a reasonable phase balance. I seem to recollect a matching consumable without the increased nickel for products that were to be solutioned annealed after manufacture.

Any ideas please.
Parent - - By js55 (*****) Date 08-09-2007 15:43
john,
You really are caught between a rock and hard spot. And yes, the fillers are Nieq enhanced. And this is the very reason that HAZ's, in my experience, can take greater jumps at extreme cooling rates than the fillers. And I emphasized ~20% ferrite in the HAZ as opposed to ~5% in the filler for two reasons (though clearly these are just very rough and worse case scenario estimates). The Nieq enhancement you mentioned and the fact that the HAZ will cool even faster than the WM. Your sittin at a good G48, good mechanicals, and some head room in the HAZ ferrite. So you do have some speace to push it a little farther if you don't get relief in another way.
As for changing fillers, the only hope you may have is to find actual chemistries in SMAW fillers that don't emphasize the Nieq enhancement as much within the specified ranges. Chuck can help us out here with more specifics but my guess is you don't have much room here to play with, but a very small change in nitrogen content can make one heck of a difference. But every little bit may help. Compare Nieq's. Another thought here since we're on the subject of nitrogn. Are you running SMAW over other processes. You could be realizing nitrogen dilution from either these processes or the BM if they happen to be higher than your SMAW filler. This won't help. If the BM, then increasing your bevel angle, or other measures, to reduce dilution could be a possibility.
The last resort is you may have to abandon SMAW all together if the engineers don't give you any relief. I'd say that increase coolng rate, decrease Nieq, check actuals on any source of dillution, and see what you get. If that don't make it you probably just won't be able to use SMAW.
Parent - - By chuck meadows (***) Date 08-09-2007 18:31
I believe John is where he is. ANSI/AWS sets the chemical range for 2209 fillers and I believe he is still gooing to get the minimum FN allowed. If he can find a manufacturer to run a complete heat where the nickel is at the bare minimum to Code and the Cr is at the highest allowable, but then you have the Mg and Nitrogen to contend with. Just try to get a huge amount of dilution with the BM and hope for the best, but I think hee is going to get lower FN's. The problem with high dilution is the fact that at higher heat ranges, the loss of Cr across the weld occurs, and there goes the ferrite former you so desperately need. I cannot imagine that 5% ferrite is going to affect the SSCC to the extent of acceptance or failure.
Parent - - By js55 (*****) Date 08-09-2007 19:01
Chuck,
I tend to agree with you on the SSCC. Seems to me if the alloy is that close to a threshold, a higher alloy altogether for SMAW might be an option the engineers can live with, if SMAW is critical to productivity. I hadn't thought of this (positive feedback brainstormin-though some might call it diamond mongering LOL). The higher alloy would render the sulphide (the first S of SSCC) not so critical.
Parent - - By chuck meadows (***) Date 08-09-2007 21:05 Edited 08-09-2007 21:37
Right Jeff..Again, all of this can be avoided if fabricators and Owners would just sit down and converse, especially relating to Duplex SMAW processes. I see it every day. The strange part is that it's been this way for 20 years. Accepting a job, starting the fabrication, then trying to solve the problem of ferrites, not the way to do it.
Parent - By Shane Feder (****) Date 08-10-2007 01:53
Hi guys,
Chuck, Jeff,
Sorry, I have to communicate with my boss via the forum as he is on R&R and this seems to be the only way to get through to him.
John,
Just to throw a real spanner in the works, as you were aware the panels were supposed to be 90% welded in the Phillipines using FCAW and SAW with SMAW on the 10% remaining welds on site. However the first 18 panels have arrived on site with 30% complete and that has left us with 70% of the welding to be done using SMAW.
I have sent you photographs and a report to your work e-mail, if they are too big to get through let me know,
Regards,
Shane
Up Topic Welding Industry / Technical Discussions / Welding lean duplex

Powered by mwForum 2.29.2 © 1999-2013 Markus Wichitill