Periodic wakes affect not only the surface boundary layer characteristics of low-pressure turbine blades and profile losses but also the vortex structures of the secondary flow and the corresponding losses. Thus, understanding the physical mechanisms of unsteady interactions and the potential to eliminate secondary losses is becoming increasingly important for improving the performance of high-lift low-pressure turbines. However, few studies have focused on the unsteady interaction mechanism between periodic wakes and endwall secondary flow in low-pressure turbines. This paper verified the accuracy of computational fluid dynamics by comparing experimental results and those of the numerical predictions by taking a high-lift low-pressure turbine cascade as the research object. Discussion was focused on the interaction mechanisms between the upstream wakes and secondary flow within the high-lift low-pressure turbine. The results indicated that upstream wakes have both positive and negative effects on the endwall flow, where the periodic wakes can decrease significantly the size of the separation bubble, prevent the formation of secondary vorticity structures at relatively high Reynolds numbers (100,000 and 150,000), and reduce the cross-passage pressure gradient of cascade. In addition, periodic wakes can improve the cascade incidence characteristic in terms of reducing the overturning and underturning of the secondary flow at downstream of the cascade all of which are beneficial for decreasing the endwall secondary losses, whereas more endwall boundary layer is involved in the main flow passage due to the wake transport, resulting in increased strength of the secondary flow at low Reynolds number of 25,000 and 50,000. Compared with the results without wakes, the total pressure loss for unsteady condition at the cascade exit decreases by 2.7% and 6.1% at high Reynolds number of 100,000 and 150,000, respectively. However, the secondary loss at unsteady flow conditions increases at low Reynolds number of 25,000 and 50,000.
