Stainless Steel

STAINLESS Q&A

BY DAMIAN J. KOTECKI

Q: For a great many years, our purchase specifications for common stainless steel bare wires such as ER308LSi and ER316LSi for GMAW or ER308L and ER316L for SAW have required 8 to 15 Ferrite Number (FN) calculated by diagram. We always referred to the DeLong Diagram, as that was included in the ASME Code, Section III, Division 1, as Fig. NB-2433.1-1. We hadn’t been paying much attention, but apparently that figure was changed to the WRC-1992 Diagram some time ago. We now feel we should “modernize” by referring to the newer diagram, but we wonder whether the calculated FNs will change. Should we make any changes to the FN requirements in our purchase specification?

A: Actually, the Fig. NB-2433.1-1 was changed from the DeLong Diagram to the WRC-1992 Diagram in the 1994 Winter Addendum to the ASME Code. So you are more than ten years behind. The change was made because it was shown, by comparing more than 200 measured Ferrite Numbers (FNs) vs. calculated FNs, that the WRC-1992 Diagram produced predictions consistently closer to the measured results, with a scatter band half as wide as that obtained with the same data using the DeLong Diagram. In particular, it was noted that the FN predictions of the two diagrams were closer together when the alloy composition was lean than when the alloy composition was richer. Keep in mind, however, that predicting ferrite in weld metal has a lot in common with predicting weather. We all admit to a degree of uncertainty in both activities. There is no simple answer to your question. One way to approach it is to pick bare wire compositions that produce FN predictions within your FN range according to the DeLong Diagram, then see what happens to the FN predictions when they are changed to the WRC-1992 Diagram. Bare wires can be produced with two silicon levels for the most part, the “normal” range of 0.30 to 0.65% and the “hi sil” range of 0.65 to 1.00%. Usual practice is to use “normal” silicon for SAW and “hi sil” for GMAW. I’ll consider both cases. I’ve arbitrarily chosen 10 FN by DeLong as the comparison point for the hypothetical compositions, and adjusted the nickel and chromium contents to remain at 10 FN for all alloys chosen.

18

MAY 2006

Fig. 1 — 10 FN from the DeLong Diagram transferred to the WRC-1992 Diagram as a function of alloy content. Table 1 was constructed by arbitrarily fixing several of the chemical composition variables of all wires at values commonly found, and then manipulating the remaining composition variables within the allowable ranges of the AWS A5.9 specification. The fixed composition variables for all hypothetical wires are 0.02%C, 1.2%Mn, 0.2%Cu, and 0.06%N. Compositions at the lean end of the AWS classification range and at the rich end were selected for all alloys except ER309L and ER309LSi to illustrate the trends. In the cases of ER309L and ER309LSi, it is necessary to raise the nickel higher than the specification allows, even with the Cr at the bare minimum of the specification, in order to force the FN calculated by the DeLong Diagram down to the arbitrarily chosen 10 FN. In the real world, ER309L and ER309LSi tend to be produced with higher nitrogen than the arbitrarily chosen 0.06% in order to ma-

nipulate the calculated FN of the wire while staying within the AWS A5.9 composition limits. Table 1 includes the calculated chromium and nickel equivalents, and calculated FN, according to both the DeLong Diagram and the WRC-1992 Diagram. For the lean ER308L composition in Table 1, the FN calculated by the WRC1992 Diagram is almost exactly the same as that calculated by the DeLong Diagram. However, as the alloy content increases, the WRC-1992 Diagram tends to predict lower FN than does the DeLong Diagram, and this tendency increases with increasing alloy content. It is illustrative to plot the FN calculated according to the WRC-1992 Diagram vs. the WRC-1992 chromium equivalent. This is done in Fig. 1 as two separate trend lines. One trend line is for the “normal” silicon compositions, and the other trend line is for the “hi sil” compo-

Table 1 — Compositions for 10 FN Calculated According to the DeLong Diagram Composition (%)

DeLong Diagram

WRC-1992 Diagram

Alloy ER308L ER308L ER316L ER316L ER309L ER317L ER317L ER308LSi ER308LSi ER316LSi ER316LSi ER309LSi

Si 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.8 0.8 0.8 0.8 0.8

Cr 19.60 20.61 18.46 19.90 22.97 18.96 20.33 19.50 20.07 18.10 19.90 23.00

Ni 9.55 10.90 11.10 13.00 14.03 13.10 14.90 10.24 11.00 11.45 13.85 14.62

Mo 0.2 0.2 2.5 2.5 0.2 3.5 3.5 0.2 0.2 2.5 2.5 0.0

CReq 20.4 21.4 21.6 23.0 23.8 23.1 24.5 20.9 21.5 21.8 23.6 24.2

Nieq 12.6 13.9 14.1 16.0 17.0 16.1 17.9 13.2 14.0 14.4 16.9 17.6

FN 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0

CReq 19.8 20.8 21.0 22.4 23.2 22.5 23.8 19.7 20.3 20.6 22.4 23.0

Nieq 11.5 12.8 13.1 14.9 16.0 15.1 16.9 12.2 12.9 13.4 15.8 16.6

FN 10.1 9.1 8.9 7.8 7.3 7.7 7.0 7.0 6.6 6.4 5.4 5.1

sitions. This is done to illustrate the point that the DeLong Diagram treats silicon as a ferrite-promoting element, while additional data used in developing the WRC-1992 Diagram led to the conclusion that silicon is not a ferrite-promoting element, at least up to 1.2% or higher. You will get slightly different results if you assume a different nitrogen content, because the coefficient for nitrogen in the nickel equivalent of the DeLong Diagram is greater than the coefficient for nitrogen in the WRC-1992 Diagram. But the general trend will be similar to that shown here. From Fig. 1 it should be quite clear that transferring a calculated FN requirement from the DeLong Diagram to the WRC1992 Diagram is not a straightforward operation. For a given calculated FN, an adjustment is appropriate. That adjustment depends upon the nominal alloy content of the particular filler metal classification. For 308L, you probably need to make no adjustment in order to obtain the same actual deposit FN you were getting previously. For 316L, you would get about the same actual deposit FN by reducing the requirement for calculated FN according to the WRC-1992 Diagram by about 2 FN from your requirement for calculated FN according to the DeLong Diagram. And for 309L filler metal, you would get the same actual deposit FN by reducing the requirement for calculated FN according to the WRC-1992 Diagram by 3 to 5 FN from your requirement for calculated FN according to the DeLong Diagram.♦

DAMIAN J. KOTECKI is technical director for stainless and high-alloy product development for The Lincoln Electric Co., Cleveland, Ohio. He is president of the American Welding Society, a vice president of the International Institute of Welding (IIW), a member of the A5D Subcommittee on Stainless Steel Filler Metals; D1 Committee on Structural Welding, D1K Subcommittee on Stainless Steel Welding; and a member and past chair of the Welding Research Council Subcommittee on Welding Stainless Steels and Nickel-Base Alloys. Send your questions to Dr. Kotecki c/o Welding Journal, 550 NW LeJeune Rd., Miami, FL 33126; or [email protected].

WELDING JOURNAL

19