Microstructural Evolution and Fatigue Performance of Welded Joints in Thick-Wall Pressure Vessels

Thick-wall welded pressure vessels operate under combinations of internal pressure, thermal cycling, structural restraint, corrosive or hydrogen-containing environments, and long service durations. In such components, fatigue performance is not governed by a single nominal material property but by a manufacturing-dependent internal state comprising spatially variable microstructure, hardness, local constitutive response, and residual stress. This review critically synthesizes the mechanisms by which welding processes and post-weld heat treatment determine fatigue performance in thick-wall pressure-vessel joints, with emphasis on SA508-type reactor pressure-vessel steels and 2.25Cr-1Mo- 0.25V/Cr-Mo-V steels used in petrochemical and hydrogenation reactors. Conventional multi-pass arc welding, narrow-gap welding, electron-beam welding, and laser-based high-energy-density processes are interpreted as distinct thermal-history generators. Their associated peak temperatures, cooling rates, reheating cycles, and through-thickness thermal gradients control fusion-zone solidification, HAZ subdivision, prior-austenite grain evolution, martensite-austenite constituent formation, carbide/carbonitride precipitation, segregation, and local mechanical mismatch. The review further evaluates how welding residual stresses originate from thermal contraction, plasticity, phase-transformation strain, restraint, and wall-thickness effects, and how post-weld heat treatment redistributes stress while simultaneously modifying phase stability and precipitate morphology. Fatigue crack initiation and propagation are discussed as coupled microstructure-residual-stress phenomena involving cyclic strain localization, small-crack interaction with microstructural barriers, crack-tip constraint, and environmental acceleration. Advanced characterization methods, including EBSD, SEM/TEM/APT, neutron diffraction, nanoindentation, and small punch testing, are assessed as enabling tools for spatially resolved microstructure-property mapping. Finally, thermometallurgical- mechanical simulation, crystal plasticity, representative-volume-element modelling, fracture mechanics, physics-informed machine learning, and digital-twin frameworks are synthesized into a process-microstructure-stress-fatigue roadmap for predictive integrity management. The central conclusion is that fatigue assessment of thick-wall welded pressure vessels must transition from averaged weld-region properties toward history-dependent, spatially resolved, and experimentally validated frameworks capable of linking fabrication parameters directly to fatigue life.