Unsteady Loss Mechanisms in Low-Pressure Turbines With Diverging End-Walls Studied Via High-Fidelity Simulation

Abstract

High-fidelity simulations are used to conduct controlled numerical experiments to investigate the effect of periodically incoming wakes on profile and three-dimensional loss mechanisms. The present work considers the MTU-T161 cascade with spanwise diverging end-walls, representative of a high-lift, low-pressure turbine blade. All simulations are carried out at engine-relevant conditions, with exit Reynolds number of 90,000 and exit Mach number of 0.6. Upstream moving bars are used to generate incoming wakes which impinge on the blade and potentially alter its aerodynamic performance. Unlike previous studies, the incoming wakes are subjected to an additional axial pressure gradient when convecting through the passage, due to the divergence of the spanwise end-walls. The evolving secondary vortex systems around the bars periodically disturb the freestream end-wall boundary layer facing the blade leading edge. This ultimately influences the end-wall related losses downstream of the blade and governs the overall aerodynamic performance of the blade. Following validation against available experimental data, a systematic variation of flow coefficient and reduced frequency extends the parametric space studied to encompass engine-realistic operating conditions. The high-fidelity simulations reveal the impact of incoming wakes on blade boundary layer losses and wake-induced losses both at the mid-span and within the end-wall regions. Furthermore, by decomposing the total loss generation, the data-rich results shed light on the underlying physical mechanisms driving unsteady losses when applied to phase- and time-averaged flow fields. Secondary losses incurred in the end-wall region show little sensitivity toward unsteadiness associated with incoming wakes and are rather prone to the turbulence levels in the passage. On the other hand, profile losses show high dependency on bar wakes in the absence of wake fogging. While profile losses can be minimized by certain combinations of flow coefficients and reduced frequencies, they remain the dominant source of unsteady loss generation.