On the sensitivity of the radial mode analysis (RMA) for the experimental evaluation of sound propagation in a low-pressure turbine

Abstract

The aeronautical industry is nowadays confronted with an important increase in the volume of passengers and transported freight. While this tendency encourages the industry expansion as well as further technology development associated with this branch, it brings along serious concerns related to the accomplishment of the stringent noise standards dictated by the governing authorities. This situation is especially critical for citizens living close to airports, as they are increasingly affected by the noise exposure. This problematic can be addressed by the development of innovative quiet and low-emission aircraft propulsion solutions. The development of such propulsion units demands however a deep understanding of the sound transport mechanisms taking place in real turbomachinery. Accordingly, the Institute of Turbomachinery and Fluid-Dynamics (TFD) studies the sound propagation through a low-pressure turbine with the purpose of developing quieter aircraft engines. Sound propagation is thus quantified by means of known established evaluation methods, one of these being the Radial Mode Analysis (RMA). The current work focuses on three central topics. In the first place, the RMA is implemented for the quantification of the associated acoustical mode amplitudes. The implementation of the RMA is achieved by means of a turbine-integrated rotating measurement unit. The development process associated with the abovementioned measuring device constitutes the second central topic. Thirdly, the RMA is optimized with respect to the quality of its output for a single operating point (m = 5 kg s-1, Ω = 3500 min-1). To this end, a sensitivity analysis of the parameters which influence the overall error associated to the RMA is performed. The analysis concentrates on the effect of two data acquisition parameters: the number of triggered rotor revolutions and the circumferential spacing of microphones. The results of the present RMA exhibit an acoustical modal structure composed of the modes m[1;0] and m[2;2] propagating at fBPF = 1750 Hz and at f2BPF = 3500 Hz, respectively. Regarding the selected data acquisition parameters, these do have an influence on the output of the RMA. An increase of the circumferential angular spacing (more azimuthal measurement positions) results in a decrease of the relative error associated to the circumferential mode amplitudes. In this way, a sufficient number of circumferential measurement positions assures a low relative error of the circumferential mode amplitude. This is only true if the number of triggered revolutions is high enough. The sensitivity analysis shows that only a sufficiently high number of triggered revolutions ensures satisfactory results independent of the circumferential mode order. As such, a realistic measurement setting for future experiments should be performed with maximal circumferential resolution (∆' = 1°) and at least N = 50 triggered revolutions. In this way, relative errors below 10% should be expected.