In the present study, creep flow and damage behaviour of modified P91 steel weldments are investigated. Premature creep failure of weldments (with respect to base metal) occurs in the intercritical heat affected zone (ICHAZ). This microstructure is reproduced by thermal simulation applied to blanks cut from the base metal. Metallurgical investigations of what happens during the welding cycle show that the weakest part of the heat affected zone is heated slightly below complete austenitisation, with little (if any) carbide dissolution. During the post-weld heat treatment, extensive recovery is allowed by carbide coarsening. The intrinsic creep behaviour of the resulting microstructure is experimentally determined under controlled constraint conditions. The welding cycle strongly decreases the creep strength by increasing the creep strain rate, but not necessarily by decreasing the ductility, at least for lifetimes up to 3 500 h.KEY WORDS: 9Cr1Mo-NbV martensitic steel; intercritical heat affected zone; Gleeble 1 500 thermalmechanical simulator; High temperature creep flow and damage.in the microstructure of interest was not known, just as for cross-weld specimens. The key point of the present study is thus to determine the properties of the weakest HAZ while eliminating, as much as possible, the constraint effects due to the surrounding materials.
Material and Experimental Procedures
MaterialsThe study focuses on a V-shaped, circumferentially welded joint of two pipes of 295 mm in outer diameter and 55 mm in thickness. It was fabricated by submerged metal arc welding in seventeen runs followed by a post weld heat treatment (PWHT) of 2 h at 760°C. The base metal is a tempered, modified P91 martensitic steel (see Table 1 for chemical composition). Its microstructure consists of lath martensite packets with M 23 C 6 (100 nm in mean size) and MX precipitates. The filler metal has nearly the same chemical composition as the base metal (Table 1).
Simulation of Welding Thermal CyclesWelding thermal cycles were applied by using a Gleeble 1 500 thermal-mechanical simulator capable of heating specimens by Joule effect at very high heating rates, corresponding to those encountered in welding conditions (i.e. 100 to 250°C · s
Ϫ1). The temperature is controlled by a thermocouple spot welded onto the specimen surface. Round bars of 5 mm in diameter were used to optimise the thermal cycle and 12 mm round blanks were treated and used to machine mechanical testing specimens respectively. For each of these two geometries, the closed-loop parameters of the Gleeble 1 500 simulator were carefully adjusted in order to ensure an uncertainty lower than 2°C for the value of T peak and smaller than 1 s for the value of the cooling parameter Dt 800→500 . For specimens of 5 mm in diameter, phase transformations were continuously monitored using in-situ dilatometric measurement of the specimen diameter. To prevent from extensive oxidation of the specimen, heat treatments were performed under primary vacuum (0.1 Pa). Phase transformations moni...