Effect of inlet reflections on fan noise radiation

Abstract

The purpose of this initial investigation into the effect of inlet reflections on turbofan noise radiation is to discover whether the contribution from reflected waves is sufficient to warrant the extra complexity of including reflection in prediction methodologies. In making evaluations, two noise prediction codes have been used, the Ventres-Theobald-Mark (turbofan source noise generation) code (VTMC), and the Eversman radiation code (ERG), which predicts far-field acoustic radiation from a turbofan engine inlet. To first evaluate reflection, the ERC alone was used. Then a coupled code using both the ERC and a modified version of the VTMC that accepts reflected waves and generates updated stator loading was developed and used for exploratory runs. Results have shown that reflection from the inlet is significant in bands of frequencies about modal cutons. Reflected amplitudes are high enough to indicate that full coupling of the source and radiation fields is needed for accurate noise predictions. I. Introduction I T is known from the mathematics of duct acoustics that reflections will occur at points in a duct where there are changes in geometry. Therefore, when turbofan noise is generated, part of the energy from the source waves is reflected by the inlet back to the source region where it can influence the noise generation process. The purpose of the present work was to study reflection and its effect on fan noise radiation to see whether the contribution from reflected waves is sufficient to justify including this information in source noise prediction models. This study was based on use of two noise prediction codes, the Ventres-Theobald-Mark (turbofan source noise generation) code (VTMC), and the Eversman radiation code (ERC), which predicts far-field acoustic radiation from a turbofan engine inlet. The programs treat the regions indicated in Fig. 1 (Ref. 1). The two regions are joined at an interface that we call the source input boundary. In the ERC, this interface is called the fan face, although the user may place this interface anywhere in the inlet. Previous analysis by Topol2 has considered the two regions in Fig. 1 with coupling of waves only from the source region to the radiation region. A complete match would require knowledge of the reflected waves at this surface. These reflected waves are the accumulated effect of reflection in the duct from the source input boundary to the inlet. Ignoring the reflections is justified if they are small. When reflections are large, a more sophisticated method might be preferable, one that would recognize waves passing the interface in both directions. This approach, called a coupled approach, could arrive at a solution, for example, by iterating between source and radiation analyses. Such an approach has been developed and used for work here by modifying the VTMC to accept reflected waves and generate updated stator loading. The plots and tables presented here supercede those in Ref. 3 upon which this work is based. As part of more recent investigations of the coupling process, some cases run originally were rerun. The results of these runs were different than those obtained previously. The difference was caused by errors in the previous versions of both the ERC and the VTMC, which have since been corrected. Subsequently, all pertinent cases have been rerun. Based on reruns, the conclusions made in the original report are not affected by these differences. Note that real fans have rotors that would be expected to attenuate the sound through rotor reflection and scattering. However, the rotor has been neglected because the purpose of the present work is to