Experimental Investigation of Fan-Stage Interaction Noise of a Turbofan Engine

Abstract

Future designs of turbofan engines are heading towards High By-Pass Ratios (HBPR) to satisfy the requirements of improving the overall efficiency, but there is an inherent acoustic penalty related to this. An increase in the by-pass ratio imposes structural constraints on the engine leading to a decrease in the axial distance between the rotor and the stator. The turbulent fan wake thus becomes more coherent on the OGVs, resulting in an increased tonal noise production. A test setup which can emulate the noise production mechanisms on a smaller scale and study the noise mitigation process is currently lacking. A characterization experiment of such a simplified test setup to assess its ability to predict the vortex-cascade interaction noise as in a realistic fan-stage is performed in this master thesis. The setup consists of an arrangement of linear cascade of OGVs, and a rod which mimics the rotor of a turbofan engine. Two experimental campaigns are undertaken. The first one consists of a wing (NACA 5406) and a rod mounted on flat plates. The wing mimics the mean aerodynamic loading of the central blade of the cascade. This experiment is performed to study the individual characteristics of the OGV blades, when placed in a cascade. The second setup is that of the Rod Linear Cascade model. Far-field directivity of noise sources are analysed using microphones placed on directivity arcs. Ability to quantify the upstream propagation of noise from the cascades is realized using a beamforming array, which is recessed on the test section wall and facing the pressure side of the OGVs. The flow-field at different locations in the test section is analysed using Hot-Wire Anemometry. The Hot-Wire measurements are also used to supplement the acoustic results. The test section is observed to successfully emulate the vortex-cascade interaction noise and this noise is observed to scale with the sixth power of flow speed. The beamforming array is found to help in improving the SNR of acoustic measurements in the test section and hence allows the quantification of the upstream propagation of vortex-cascade interaction noise.