Experimental analysis and characterization of flow turbulence regarding its effect on fan noise

Abstract

With the aim of making aircraft "greener'', researchers are continuously looking for solutions to make airplanes quieter and more efficient solutions to make airplanes quieter and more efficient. Actual and future aircraft the aim of making aircraft “greener”, researchers are continuously looking for concepts typically rely on highly integrated propulsion systems, aiming on the increase of flight efficiency, what typically leads to higher levels of turbulence ingested by the fan, and consequently the increase of fan noise generation. This work is focused in the understanding of fan noise generation due to distorted inflow. For the experimental investigation a series of measurements were performed in the low-speed aeroacoustic fan rig named CRAFT. Before experiments with variable inflow distortion may start, a baseline inflow with homogeneous and low-turbulence levels has to be established, as a prerequisite for fan self-noise investigation. This can be achieved by installing an Inflow Control Device (ICD) on the test rig’s inlet. For the CRAFT fan rig an ICD was especially conceived and constructed using flexible honeycomb covered on both sides with a thin wire mesh to ensure maximum turbulence reduction. Variable inflow distortions were generated using perforated plates combined with a honeycomb installed downstream of the inlet bellmouth. The aerodynamic performance of the ICD (baseline) and the inflow distortion tests were assessed by means of hot-wire and total pressure rakes. The fan noise emission was recorded by microphone arrays. A software to estimate turbulence parameters such as turbulence intensity and integral length scale (ILS) from hot-wire data was implemented. A novel technique was developed to estimate these turbulence parameters from the flow downstream of the fan by separating the statistics of the rotor wakes from the background flow. The ICD established a homogeneous inflow both in terms of turbulence intensity (as low as 0.1%) and the mean flow velocity. The turbulence distribution was characterized both in the inlet and between the fan and the stator section. Three inlet configurations were compared for different fan operating points: Without the ICD, with the ICD, and with ICD and an additional honeycomb installed upstream the fan. Acoustic analyzes revealed a reduction of the fan broadband noise levels and more stable rotor coherent tones with ICD compared to without. A reduction of narrowband components around the blade passing frequencies was also observed and is believed to be associated with the reduction of the low-frequency power content in the turbulence spectrum due to the ICD. These properties were further improved by adding a honeycomb in the duct inlet, with the penalty of a slight increase of the turbulence levels. The use of perforated plates combined with a honeycomb successfully generated distorted inflow profiles, which are relevant for fan noise investigations. Both the mean velocity profile and the turbulence field were altered by the screens. Acoustic measurements revealed an increase in broadband fan noise emission under distorted inflow conditions. The parametric tests with different levels of distortion allowed the assessment of the parameters that impact on the fan broadband noise generation. Analyzes suggest an approx. linear relationship between the fan broadband sound power emissions in decibels and the average ingested turbulence power. The turbulence profile generated inside the distorted area differs from what is typically seen in boundary layer ingesting fans.