Numerical and Analytical Investigation of Heat Transfer Mechanisms and Flow Phenomena in an IP Steam Turbine Blading During Startup

As a result of an ever-increasing share of volatile renewable energies on the world wide power generation, conventional thermal power plants face high technical challenges in terms of operational flexibility. Consequently, the number of startups and shutdowns grows, causing high thermal stresses in the thick-walled components and thus reduces lifetime and increases product costs. To fulfill the lifetime requirements, an accurate prediction and determination of the metal temperature distribution inside these components is crucial. Therefore, boundary conditions in terms of local fluid temperatures as well as heat transfer coefficients with sufficient accuracy are required. As modern numerical modeling approaches, like 3D-Conjugate-Heat-Transfer (CHT), provide these thermal conditions with a huge calculation expense for multistage turbines, simplified methods are inevitable. Analytical heat transfer correlations are thus the state-of-the-art approach to capture the heat transport phenomena and to optimize and design high efficient startup curves for flexible power market. The objective of this paper is to understand the predominant basic heat transfer mechanisms such as conduction, convection and radiation during a startup of an IP steam turbine stage. Convective heat transport is described by means of heat transfer coefficients as a function of the most relevant dimensionless, aero-thermal operating parameters, considering predominant flow structures. Based on steady-state and transient CHT-simulations the heat transfer coefficients are derived during startup procedure and compared to analytical correlations from the literature, which allow the calculation of the heat exchange for a whole multistage in an economic and time-saving way. The simulations point out that the local convective heat transfer coefficient generally increases with increasing axial and circumferential Reynolds’ number and is mostly influenced by vortex systems such as passage and horseshoe vortices. The heat transfer coefficients at vane, blade, hub and labyrinth-sealing surfaces can be modeled with a high accuracy using a linear relation with respect to the total Reynolds’ number. The comparison illustrates that the analytical correlations underestimate the convective heat transfer by approx. 40% on average. Results show that special correlation-based approaches from the literature are a particularly suitable and efficient procedure to predict the heat transfer within steam turbines in the thermal design process. Overall, the computational effort can be significantly reduced by applying analytical correlations while maintaining a satisfactory accuracy.