A case study investigation into the effects of spatially dependent convection coefficients on the fatigue response of a power plant header component

Societal and environmental pressures are forcing thermal power plant operators to deviate greatly from the generation strategies of the past. The application of high frequency start up/shut down/partial load cycles to components that may well be outside their design life makes research into the long-term integrity of at risk assets paramount. Decoupled thermal/mechanical analyses have been used in the literature to estimate anisothermal fatigue in header components, with convective boundary conditions typically assumed for internal surfaces in order to determine heat fluxes and hence a temperature field. In reality, convective coefficients are heavily dependent upon local velocity profiles. In the present work, computational fluid dynamics is used in order to better approximate the steam flow in a real power plant header, leading to a convection coefficient field that is used to solve the thermal problem. Anisothermal fatigue analysis is finally conducted using a Chaboche type model. The results of computational fluid dynamics have illustrated that heat transfer coefficient values can vary (spatially) by a factor of 5.49 over the internal header wall, with noticeable hot spots in the wake of the stub penetrations. Peak differences of 6.47% in accumulated plastic strain levels have been observed between simulations conducted with constant (simplified) and variable (computational fluid dynamics derived) thermal boundary conditions.