Numerical simulation and modelling of the effect of film cooling on entropy noise in a stator blade row

Abstract

This thesis enhances the fundamental understanding of how film cooling flows interact with entropy waves, and their influence on entropy noise generation. A 2D numerical framework has been developed for preliminary understanding of the problem, and then analysis of the effect of film cooling on the entropy noise generation and entropy wave attenuation is carried out using a more complete 3D URANS simulation of a blade mid-span section.

The 2D investigation has determined that consideration of viscous effects and thermal conduction is important for accurate entropy noise predictions. The boundary layer increased the global acceleration in the blade passage, resulting in a blade exit velocity increase of 30-40 m/s, magnifying both the entropy noise generation and entropy wave attenuation. A cooling flow is found to have a similar effect. The layer of cool air on the blade surface displaces the incoming flow, increasing the convective acceleration of the entropy perturbed flow. The entropy wave is therefore accelerated at a higher rate, leading to additional entropy noise.

3D simulations demonstrate that a cooling mass flow increase of only 3.4% with respect to the freestream amplifies the entropy noise generation by over 12%. Since the cooling flow is contained near the blade surface, the planar entropy wave experiences limited extra attenuation due to increased shear dispersion. This is found to scale with the additional mass flow rate of the coolant.

An analytical model has been extended to account for a film cooling flow by relaxing mass conservation across the blade row and by assuming perfect mixing. Hence, entropy is not conserved across the row. The trend of entropy wave attenuation with increasing cooling flow is well predicted by the extended model, with a slight overestimation of ~4% compared to the high fidelity CFD simulations. The trend of rising entropy noise with an increasing cooling flow is also captured for the transmitted entropy noise transfer function. However, the prediction of the reflected entropy noise is over estimated by 6.5%, and diverges with increasing mass flow ratio. Analysing the impact of the temperature ratio between the coolant and the freestream, it has been shown that a decrease in the ratio leads to a monotonic increase in entropy wave attenuation.

Date of Award	Jul 2022
Original language	English
Awarding Institution	Queen's University Belfast
Sponsors	Engineering and Physical Sciences Research Council
Supervisor	Rob Watson (Supervisor) & Juliana Early (Supervisor)