P91 to SS Weld 3

[dbsmith]

Does anyone have experience joining SA335 Grade P91 to SA312 Type 304L stainless? In the absence of other advice, I plan to butter the P91 with ERNiCr3 filler metal and apply PWHT, then weld stainless to the buttered section again using ERNiCr3. Am I headed in the right direction? Thanks.

 

[mqas]

DBSmith,

Have welded A217-C12A to A376-TP316

GTAW Buttering using Metrode 20.7Nb and 250 Deg. C Preheat

PWHT @ 750 deg. C

SMAW weld using Metrode Nimrod 182KS

 

[GRoberts]

mqas: Did you mean Metrode 20.70.Nb?

If so, that is ERNICr-3. The Metrode Nimrod 182KS is NiCrFe-3. I think either combinations would be fine. However, since P91 has nitrogen, I would use the one that has the lowest Nb available. Nb is supposed to cause brittle zones where it formes nitrides. It isn't always a problem, as super austenitic and duplex has often be welded sucessfully with Inconel 625 (NiCrMo-3), but why not aviod the chance if possible?

 

[unclesyd]

Look at Inconel 617 weld rods. Check thier web site or get a Inconel booklet from your consumables supplier.

 

[davefitz]

Ome problem discovered with hese dissimialrmetal welds i\on tubing was that a manual welder would reduced thamps to prevetn weld spater, but this woudl also prevent proper weld penetration. The most successful welds of P22 or P91 to SS304 H with incolnel 625 have been with orbital welders wherein the amps were specifec by the welding engineer, and a very wide weld pass was used.

Normally P91 is used for high temperature creep strength, and that would imply comparable use of 304H not 304 L. Also, because of the high cyclic shear stress imposed on the weld after cyclic heating to design temp followed by cool down to ambient, cracks may eventually appear in the mid wall , detected by angle UT. One important measure to prevent failure is to limit stresses caused by bending moments at the weld to less 20% allowable stress for the weaker alloy.

 

[mqas]

DBSmith

Sorry you are correct 20.70Nb

 

[bngreene]

You are on the right track using the ERNiCr-3 wire for welding the P91 to an austenitic stainless steel. This is one of the methods recommended by Vallourec & Mannesmann, who are major suppliers of P91 tubular products. You can contact them in Houston (V & M Tubes 713-479-3200) if you want more specific information.

While Inconel 617 and Inconel 625 fillers may also work, I would not suggest them. As you are welding the P91 to 304L, you apparently don't need the high temperature strength these alloys provide. The ERNiCr-3 is as strong or stronger than the 304L, and is therefore strong enough for your application.

Even at higher temperatures, there are some cracking mechanisms that sometimes make the use of Inconel 617 and Inconel 625 problematic. It's also hard to imagine that the dlution effect of the niobium in Inconel 625 will be helpful to the P91 weldment.

 

[dbsmith]

Thanks to all for the advice. The application is HRSG drains, and as bngreene noted, the weld is outside of the temperature range where P91 provides hi-temp strength and creep resistance. I'm not positive as to why 304L was specified - I suggested 304H, which usually takes PWHT better than L grades, but the customer wanted 304L.

 

[stanweld]

Your proposed method has been used many times.

I,however, do not recommend such joints for use in todays combined cycle HRSG power generating stations or any other application where thermal cycling is anticipated. Weld failure by thermal fatigue is inevitable. Dissimilar butt welds in pipe have failed in fewer than 50 full temperature cycles. Repairs using best EPRI practices have failed in less than 10 cycles.

If the P91 material is operating below 1000 F, I would recommend welding a P22 transition to the P91 alloy and then welding the 304L to the P22 transition. This way you avoid having to postweld heat treat the P22 to 304L weld.

 

[davefitz]

As per the Mannesman Vallourec manual, the preferred method to weld P91 tubes to 304 is to fabricate a "safe end" in a ashop ahead of time.

"first nickel base weld metal is deposited on one side of the T91 piece followed by a NT treatment similar to T91 base material. This transition piece can tehn be welded on site to T91 and to the austenitic steel. The T91 side is a regular similar weld with matching consumable and local PWHT, whereas the austenitic side is welded to the nickel base weld deposit using a nickle base consumable without PWHT"