1. Σ SWL(100-3500Hz) [dB] ϕ [mm] 98.9 99.4(+) 96.2 (-) 0.304 2.4 114.1 115.5(+) 113.3(-) 0.194 2.4 109.8 111.2(+) 101.7(-) 0.095 2.4 99.7 99.6(+) 96.5 (-) 0.306 0.6 102.2 103.6(+) 98.2 (-) 0.163 0.6 110.3 110.9(+) 104.6(-) 0.097 0.6 Table 1 gives an overview of all experimental results. The sound power was summed over the frequency range 100-3500 Hz. The calculated sound power of the new method described in this paper is slightly, but significantly higher than the results obtained from the sampling tube measurements. ISO 5136 does not give the power of both transmitted and reflected modes. Hence, this result must be compared with the sum of SWL(+) in flow direction and SWL(-) against flow direction. It can be concluded, that the broadband mode detection method delivers between 1 and 2 dB higher total SWL in the investigated frequency range. The inferred reflection coefficients are much lower than one, but seem to be higher than one would expect them to be, as the duct end is equipped with an anechoic termination. One reason could be the noise floor in the time data sets. Broadband noise data does not have the same quality of signal to noise ratio for each frequency component, as is the case with dominant rotor tones. A longer measurement time mightimprovethis situation.
2. L. Enghardt, Y. Zhang, W. Neise: "Experimental verification of a radial mode analysis technique using wall-flush mounted sensors",137thMeetingoftheAcousticalSocietyofAmerica,Berlin,March15-19,1999