ALBERT

Description: Recent experience using ANOPP to predict turbofan engine flyover noise suggests that it over-predicts overall EPNL by a significant amount. An improvement in this prediction method is desired for system optimization and assessment studies of advanced UHB engines. An assessment of the ANOPP fan inlet, fan exhaust, jet, combustor, and turbine noise prediction methods is made using static engine component noise data from the CF6-8OC2, E(3), and QCSEE turbofan engines. It is shown that the ANOPP prediction results are generally higher than the measured GE data, and that the inlet noise prediction method (Heidmann method) is the most significant source of this overprediction. Fan noise spectral comparisons show that improvements to the fan tone, broadband, and combination tone noise models are required to yield results that more closely simulate the GE data. Suggested changes that yield improved fan noise predictions but preserve the Heidmann model structure are identified and described. These changes are based on the sets of engine data mentioned, as well as some CFM56 engine data that was used to expand the combination tone noise database. It should be noted that the recommended changes are based on an analysis of engines that are limited to single stage fans with design tip relative Mach numbers greater than one.

Description: An extension of a prior study has been completed to examine the potential reduction of aircraft flyover noise by the method of active noise control (ANC). It is assumed that the ANC system will be designed such that it cancels discrete tones radiating from the engine fan inlet or fan exhaust duct, at least to the extent that they no longer protrude above the surrounding broadband noise levels. Thus, without considering the engineering details of the ANC system design, tone levels am arbitrarily removed from the engine component noise spectrum and the flyover noise EPNL levels are compared with and without the presence of tones. The study was conducted for a range of engine cycles, corresponding to fan pressure ratios of 1.3, 1.45, 1.6, and 1.75. This report is an extension of an effort reported previously. The major conclusions drawn from the prior study, which was restricted to fan pressure ratios of 1.45 and 1.75, are that, for a fan pressure ratio of 1.75, ANC of tones gives about the same suppression as acoustic treatment without ANC. For a fan pressure ratio of 1.45, ANC appears to offer less effectiveness from passive treatment. In the present study, the other two fan pressure ratios are included in a more detailed examination of the benefits of the ANC suppression levels. The key results of this extended study are the following observations: (1) The maximum overall benefit obtained from suppression of BPF alone was 2.5 EPNdB at high fan speeds. The suppression benefit increases with increase in fan pressure ratio (FPR), (2) The maximum overall benefit obtained from suppression of the first three harmonics was 3 EPNdB at high speeds. Suppression benefit increases with increase in FPR, (3) At low FPR, only about 1.0 EPNdB maximum reduction was obtained. Suppression is primarily from reduction of BPF at high FPR values and from the combination of tones at low FPR, (4) The benefit from ANC is about the same as the benefit from passive treatment at fan pressure ratios of 1.75 and 1.60. At the two lower fan pressure ratios, the effectivness of treatment is much greater than that of ANC, and (5) No significant difference in ANC suppression behavior was found from the QCSEE engine database analysis compared to that of the E3 engine database, for the FPR = 1.3 engine cycle. The effects of ANC on EPNL noise reduction are difficult to generalize. It was found that the reduction obtained in any particular case depended upon the frequency of the tones and their shift with rpm, the amount of ANC suppression received by each tone (which depended on its protrusion from the background), and the NOY-value of the tone relative to the NOY-value of other tones and the peak broadband levels, because PNL is determined from the sum of the NOY-values.