Abstract Scope |
* Introduction
Residual lifetime assessment of welded 9Cr steel components is crucial for power plant operation and maintenance. Not only should the creep fracture mechanisms across heat affected zone (HAZ) be fully understood, but also the creep properties of each sub-region within the HAZ itself—including the coarse-grained HAZ (CGHAZ), fine-grained HAZ (FGHAZ), and intercritical HAZ (ICHAZ)—should be obtained. However, the highly inhomogeneous HAZ (2–3 mm in width), generated from nonequilibrium austenitization on heating and martensitic transformation on cooling during fusion welding processes, pose many challenges to achieving these goals. The infamous Type IV cracking in HAZ used to be widely reported to be caused by a faster deformation in a soft zone (SZ) with the lowest hardness. This SZ has been inconsistently identified as either an FGHAZ, an ICHAZ, or even over-tempered base metal (BM) in different reports. Non-uniform creep deformation in sub-regions of the HAZ cannot be directly measured with standard creep testing methods. To address these issues, we purposefully designed a series of integrated experiments, including a finite element (FE) model–assisted Gleeble simulation, high-temperature creep testing with in-situ strain distribution measurement by digital image correlation (DIC), and high-resolution electron microscopy microstructure analysis.
* Experimental Procedures
In the present work, Grade 91 steel (plate) (Oak Ridge National Laboratory (ORNL) heat 30176) used as the base metal was normalized at 1050 °C for 1 h and tempered at 760 °C for 2 h. A coupled electric-thermal FE model was developed and applied to customize a thermal gradient (peak temperature profile) with a specially designed specimen geometry. Various HAZ microstructures were reproduced in a single designed specimen with the Gleeble simulator. The entire thermal history of the specimen was recorded with thermocouples and an infrared (IR) camera. The simulated HAZ specimen was then subjected to a post-weld heat treatment (760 °C for 1 h) and tested (creep temperature: 650 °C, stress: 80 MPa) using an ATS 2330 series lever arm testing frame integrated with a purpose-built DIC system. Both full-field nominal creep strain over the gauge length and the local creep strain over the multiple HAZ sub-regions were determined by post-processing the DIC data with the VIC-2D® software. Vickers hardness measurements and advanced metallurgical analysis with SEM/EBSD/TEM/EDS were conducted on the specimen before and after creep testing.
* Results and discussion
The peak temperature profiles predicted by the electrical-thermal FE model and measured by the IR camera exhibit s high agreement with each other over the entire temperature range (500–1200 °C). This enlarged 26.2 mm HAZ specimen is about 10 times the 2.5 mm HAZ in a typical weld. The measured widths of the HAZ sub-regions, including CGHAZ (1050–1200 °C), FGHAZ (950–1050 °C), and ICHAZ (850–950 °C), were 13.0 mm, 6.8 mm, and 6.4 mm, respectively. The soft zone with the lowest hardness (256 HV0.5) was identified at a 25.1 mm distance from the specimen center. The most creep-vulnerable zone (CVZ) with a hardness of 380 HV0.5 prior to creep is identified at 20.8 mm from the specimen center. The results show creep fracture did not occur in the soft zone. The exact fracture location (the CVZ) is in the ICHAZ exposed to a peak temperature of 932 °C. Creep curves of each HAZ sub-region extracted from the high-resolution DIC strain measurements clearly show the ICHAZ deforms faster than the FGHAZ and CGHAZ in all creep stages. There is no obvious secondary creep stage with a steady creep rate for the ICHAZ. Localized creep strain in the ICHAZ reaches beyond 100% right before fracture, while the nominal strain over the gauge length is only about 8.6%.
The characterization results reveal that the SZ exhibited a higher microstructural stability than the CVZ. EBSD and TEM results shows untransformed but over-tempered martensite grains observed in the as-simulated CVZ exhibit a lower internal strain energy or dislocation density. It is also observed that incomplete dissolution of Cr-rich M23C6 carbides led to an inhomogeneous Cr distribution in the CVZ. Local Cr enrichment as high as 12 wt.% was the critical factor responsible for retaining those untransformed grains. The matrix grains in the CVZ recovered into nearly ferritic grains (average grain size: 4.98±2.52 µm), a 15.5 % increase from the PWHT stage. The high-angle grain boundary (HAGB) fraction increased from 24.3 % to 38.4 % accordingly, while the SZ experienced a slower microstructure degradation during creep. Grain sizes in the SZ increased only to 7.61±4.88 µm, and the HAGB fraction increased from 23.4 % to 26.5 %.
* Conclusion
In summary, the developed Gleeble-Creep-DIC approach successfully quantified localized creep properties within the HAZs of 9Cr steels. In-situ DIC strain measurements provided additional views of creep deformation behaviors of multiple HAZs with varying microstructures, which further clarified some uncertainties regarding high-temperature creep deformation and fracture mechanisms in creep-resistant steel weldments, especially Type IV cracking. The most CVZ was in the high-temperature end of the ICHAZ exposed to a peak temperature of 932 °C close to AC3. The CVZ was not the soft zone with the lowest hardness before creep. This result indicates that using room-temperature hardness values across welds measured before creep can lead to inaccurate prediction of creep lifetimes and potential creep rupture locations. Significant microstructural degradation in the CVZ during creep led to its largest hardness reduction and eventually the lowest hardness after creep testing. Fast matrix grain growth due to reduced precipitation strengthening and grain boundary strengthening directly led to rapid creep strength degradation in the CVZ. This integrated testing technique can easily be adapted to test other high-temperature alloys used in advanced power plants. |