A dissimilar metal weld (DMW) is created whenever alloys with substantially different chemical compositions are welded together — for example, when a low-alloy steel such as Grade 22 (2¼ Cr-1Mo) is welded to an austenitic stainless steel such as TP304H (18Cr-8Ni). Many DMWs are commonly present in fossil-fired power plants, examples being material transitions in boiler furnace tubes, stainless steel attachments welded onto ferritic steel tubes or pipes, and stainless steel thermowells or steam sampling lines in ferritic steel pipes. The chemical composition gradients associated with DMWs present unique issues relative to their design, in-service behavior, and life management, particularly for those DMWs operating at elevated temperatures where solid-state diffusion and cyclic thermal stresses are factors, which was previously presented in News and Views Issue 43.
With the now widespread use of Grade 91 steel (9Cr-1Mo-V-Nb) for elevated-temperature applications in modern power plants, DMWs involving this material have become common, and increasing service experience has revealed some unique characteristics and failure mechanisms, especially in thicker-section DMWs with austenitic materials. This article presents a short overview of Grade 91 DMWs: their design, fabrication, and failure, with emphasis on current industry issues.
There are two basic classes of DMWs in Grade 91 steel: ferritic-to-ferritic and ferritic-to-austenitic. The first type corresponds to Grade 91 welded to another ferritic steel with a lower chromium content, such as Grade 22; the second type corresponds to Grade 91 welded to an austenitic stainless steel such as TP304H. Each of these types has unique concerns and considerations.
There are several potential issues which must be considered in the design and fabrication of DMWs of Grade 91 to other ferritic steels, perhaps the foremost being the migration of carbon atoms from the material with lower chromium content to the Grade 91 material with higher chromium content. The chromium content difference across the weld interface creates a driver for carbon diffusion which will occur during both any post-weld heat treatment (PWHT) applied or during service at temperatures greater than approximately 900°F. This diffusion creates a weak decarburized layer in the lower-chromium material which can result in premature cracking and failure (Figure 1). There will also be a significant difference in strength between the two materials, which can result in strain localization in the weaker material if the joint is not designed correctly. The strain localization will be enhanced by any decarburization related to carbon migration. A third issue is selecting a PWHT temperature, since the code-required ranges for the different materials do not always overlap.
DMWs between Grade 91 and lower alloy ferritic steels are generally more problematic than DMWs between different low-alloy steels (e.g., between Grade 22 and Grades 11/12) and DMWs between low-alloy steels and carbon steels. This is because ferritic-to-ferritic DMWs involving Grade 91 have greater differentials in chromium content, higher PWHT and operating temperatures (higher carbon diffusion rates) and greater strength mismatches.
While every configuration is unique, it is generally best to make a Grade 91-to-ferritic DMW using a filler metal matching in composition to Grade 91 (e.g., type B9 or B91) rather than to the lower-alloy material (e.g., type B3), although both are options depending on the specific configuration and requirements. It is not advisable to make DMWs between Grade 91 and Grades 11/12 steels or low-carbon steels due to the large mismatches in strength, chromium content, and PWHT temperature.
Because there are typically transitions in both material and component thickness at a DMW, it is critical to design the DMW accounting for both types of transitions and the expected operating conditions of the component (service temperature, the frequency of thermal cycling, etc). Figure 2 shows examples of good and bad transition designs. In the good design a separate piece is used to make the thickness transition in the stronger material and the DMW is placed in the high thickness region; in the bad design the low-alloy material (and the decarburized layer) is present at the lower thickness (and correspondingly higher stress levels) of the Grade 91 component.
DMWs connecting Grade 91 piping to steam turbine stop/control valves have been found with premature service damage and cracking. In these cases, the Grade 91 pipe is typically connected to a steam turbine OEM proprietary CrMoV valve body casting, sometimes using a relatively low strength filler metal such as type B3, or even B2; a typical and the “bad” design configuration shown in Figure 2. In such configurations, a Grade 91 matching filler metal would perform better. See the article; Example Grade 91 High Energy Piping DMW Joint Stress and Metallurgical Analysis, (page 39).
DMWs joining Grade 91 to austenitic stainless steels are relatively common, perhaps more so than realized. In addition to butt welds, connecting stainless steel furnace tubing to Grade 91 outlet header tube stubs in modern conventional boilers, Grade 91 to austenitic DMWs are present where stainless steel attachments have been welded to Grade 91 tubes, where stainless steel thermowells or steam sampling lines have been used in Grade 91 piping systems, and where stainless steel support lugs are welded to Grade 91 headers. There are also more isolated cases of DMWs in which Grade 91 pipe has been welded to austenitic stainless steel pipe, for example where stainless steel in-line flowmeters are present in a Grade 91 piping system, or boilers in which stainless steel outlet headers connect to a Grade 91 piping system.
To date, DMWs joining Grade 91 to stainless steel have been made using nickel-base alloy weld consumables such as ERNiCr-3/ENiCrFe-3 (nickel alloy 82/182) using welding parameters (preheat, interpass, and PWHT temperatures) suitable for Grade 91. In thin-section welds, the two materials are typically joined directly and the completed weld given a PWHT at Grade 91 conditions (Figure 3a). In thicker-section welds the Grade 91 side of the joint is sometimes first “buttered” with several layers of nickel-base filler, subjected to a PWHT, and then a final closure weld is made to the stainless side of the joint (again with a nickel-base filler). No PWHT is performed after the closure weld (Figure 3b). Note that nickel-base fillers which strengthen or embrittle at typical Grade 91 service temperatures should be avoided. These include alloys 625 (E/ERNiCrMo-3) and 617 (E/ERNiCrCoMo-1).
Use of a nickel-base filler metal avoids the issue of carbon migration in the DMW, since carbon does not readily diffuse into the nickel-base material. However, the significant differences in thermophysical properties (especially thermal expansion coefficient) between the components of the DMW can lead to adverse thermal cycling effects. Since the thermal expansion coefficient of Grade 91 is slightly lower than low-alloy steels such as Grade 22, thermal stresses will be higher in Grade 91 to austenitic DMWs than in Grade 22 to austenitic DMWs. As with the Grade 91 to ferritic DMWs, strength and thickness mismatches typically exist, although in the austenitic DMW case Grade 91 is the weaker material, at least at the service temperature. Hence similar attention must be paid to overall joint design and thickness transitions.
The service experience with Grade 91 to austenitic DMWs made as butt welds in thin sections (e.g., tubing) has so far been positive; the few reported failures in these welds have been primarily attributed to factors other than the DMW itself, for example high thermal stresses due to sootblower impingement.
In contrast, however, many premature failures have occurred in Grade 91 to austenitic DMWs with thick sections (e.g., pipe butt welds) or significant weld constraint (e.g., fillet welds). An EPRI report documenting recent service experience with thick-section Grade 91 DMWs was prepared by Structural Integrity in 2014 (Report 3002006759). The most common failure mode for these cases has been cracking along the weld fusion line between the Grade 91 and the nickel-base weld metal (Figure 4). Through-wall cracking has occurred in relatively short service durations, fewer than 20,000 hours in one case. The cracking appears to be creep-related, in that cavities are frequently observed along the fusion line ahead of the crack tip. This is particularly true for DMWs operating at temperatures around 1000°F and slightly above. In contrast, creep cavitation was found at both the fusion line and in the Grade 91 heat-affected zone (“Type IV” damage) in a pipe DMW which operated at 1050°F.
The risk presented by stainless steel in-line flow elements was identified and documented in 2010 by Matherne, et al., and in 2011 by Paterson, et al.; these flow elements were installed in many early Grade 91 piping systems because Grade 91 flow elements were unavailable. If still present, they should be replaced at earliest opportunity because of the risk of unpredictable catastrophic rupture along the stainless to Grade 91 DMWs.
More attention is now being paid to Grade 91 to stainless DMWs in fillet weld applications such as tube attachments, thermowell fittings, and steam sampling line fittings because these are increasingly found to be cracked in piping inspections. Cracking is again typically along the Grade 91 to nickel-base weld metal fusion line; examples are shown in Figures 5 and 6. The thermowells are of particular concern because of the significant projectile safety hazard associated with their liberation (this has occurred). Any such DMWs present in a Grade 91 piping system should be eliminated where possible by replacing the stainless steel components with equivalent Grade 91 components; if this is not possible these DMWs must be regularly inspected (10,000 to 20,000 hour intervals). EPRI plans to perform a service experience review for thermowells and steam sampling lines in 2018.
The metallurgical root cause for these premature failures in thick-section Grade 91 to austenitic DMWs is not fully understood and is the topic of on-going research by EPRI. It is not currently possible to make life predictions for these welds based on their nominal service loading; early failures have occurred in welds operating at very low service stresses. The best strategy for utilities is to identify where such welds are present in a plant, eliminate them where possible, and frequently inspect the remainder.
In summary, Grade 91 DMWs present an additional set of concerns and considerations compared to low-alloy steel DMWs. When properly designed and fabricated, DMWs between Grade 91 and lower-alloy steels such as Grade 22 should not present a particular risk for premature failure, and current service experience suggests that this is also the case for Grade 91-to-stainless tube butt DMWs. However, for Grade 91-to-stainless DMWs in thick sections or other configurations with high weld constraint present a high risk for premature failure; these welds should be identified, carefully monitored, and eliminated where possible.
 Matherne, C., D. DeRouen, and W. Baxter. “Catastrophic Failure of a Dissimilar Metal Weld in a High Pressure Steam Venturi”. in 2010 AIChE Spring National Meeting. March 21-25, 2010. San Antonio, TX.
 Paterson, S.R., R. Gialdini, and M. Cronin. “Early Service Life Cracking of Steam Piping Welds in Combined Cycle Power Plants”. in 2nd International EPRI Conference on Welding and Fabrication Technology for New Power Plants and Components. June 21-24, 2011. Orlando, FL.