Varestraint testing is commonly used to evaluate hot cracking susceptibility of materials. In this paper, the dependence of operators and evaluation technique on test results is studied for a high-temperature austenitic stainless steel (UNS S31035). Samples were tested at six different strain levels ranging from 0.7 to 3.8%. Four different operators evaluated the same samples following the same instructions on how to measure the cracks manually in an optical microscope at x 25 magnification. The largest variation among operators evaluation was found for low strain levels where small and few cracks were found. In addition, one of the four operators used image analysis to evaluate the samples at x 50 magnification. The average total crack length and total number of cracks in fusion zone and heat-affected zone were approximately 1.5 times higher when using image analysis compared with manual evaluation. Image analysis at x 50 made it possible to detect smaller cracks compared with manual evaluation at x 25 magnification, contributing to an increased number of cracks detected. The maximum crack length using image analysis at x 50 was similar to manual evaluation made at x 25 magnification and was the criterion that showed the least variation in this study. However, further comparisons using other magnifications are needed to verify the agreement between manual evaluation and image analysis found in this study. An advantage with evaluation using image analysis is that it provides traceable results. A harmonized standard for Varestraint testing, and especially for evaluation, would decrease the variation among operators and laboratories.

Sandvik Sanicro 25 (UNS S31035) is an advanced high temperature austenitic stainless steel that potentially can be used in super-heaters and reheaters in the next generation of advanced ultra-super critical power plants (A-USC). The material possesses both high creep strength and good corrosion resistance at temperatures up to 700°C. It is, however, well known that an austenitic solidification mode combined with a fully austenitic microstructure exacerbate susceptibility towards hot cracking. In this paper, the Varestraint test is used to investigate the susceptibility to hot cracking of Sanicro 25 and is compared to the well-known grade 310S (UNS S31008), which has been used in past generations of power plants. This investigation therefore attempts, based on the Varestraint weldability test, microstructural inspection, and thermodynamical calculations, to investigate if there are any differences in cracking mechanisms between these two grades due to the additional alloying elements used in Sanicro 25. The results show that the hot cracking susceptibility of Sanicro 25 is only slightly higher than that of 310S. The larger solidification interval for Sanicro 25 compared to 310S predicted by computational thermodynamics and the niobium-enriched phases in the grain boundaries of Sanicro 25 supported the hot cracking susceptibility results. Weldability of the two alloys is therefore judged to be comparable making the newer alloy a good candidate material to be used in A-USC power plants from a weldability perspective.

Two 617-type filler metals with different Carbon and Boron contents were used to deposit TIG all-weld metal. Longitudinal Varestraint testing was utilized to evaluate and compare their weld metal cracking susceptibility. Hardness, tensile, and impact toughness testing were conducted on the all-weld metal samples, and light optical microscopy as well as scanning electron microscopy, equipped with energy dispersive spectroscopy were adopted for the microstructural inspection of the Varestraint tested samples. Computational thermodynamics supported in calculating the solidification interval and predicting the phases formed. Results showed that the modified Alloy 617 (617mod.) with higher Boron and lower Carbon content is the preferred filler metal, because it showed lower hot cracking susceptibility than the regular version of Alloy 617. In addition, the impact toughness of the modified Alloy 617 was almost three times higher than for Alloy 617 but showed lower tensile strength. In terms of microstructure, the modified Alloy 617 disclosed less precipitates along the cracks than the regular Alloy 617. These observations were supported by computational thermodynamic calculations.