Engineering topics
Publications and source records attributed to Saule, A. V..
The far-field radiation from the end of an exhaust duct is studied using both approximate and exact methods. Experimental data of narrow-band tone noise from static tests are compared to a multimodal radiation pattern. It is pointed out that possibly the exhaust noise in the far-field is inherently more difficult to attenuate than an inlet noise using duct suppressors.
Approximate expressions were developed for the noise radiation from the aft duct. The results of approximate aft radiation equation compare favorably to more exact Wiener-Hopf radiation results. Refraction as well as convective effects in the multiple flow streams is considered. The peak in the radiation pattern, which occurs nearly at engine sideline, is composed of modes with relatively large cut-off ratios. This implies that aft fan radiation will be inherently more difficult to suppress that the fan inlet noise. The theoretical multimodal radiation pattern is compared to experimental data for the first two harmonics of blade passage frequency for three full scale fans at two speeds. The agreement between theory and experiment is quite good.
Explore the source record for details and available documents.
Interaction tones were intentionally generated by circumferential arrays of equally spaced rods that protrude radially from the inlet wall near the face of the 28-blade fan. Arrays of 28 and 41 rods, selected to give specific far field radiation properties, were tested. The expected properties were readily apparent in the measured radiation patterns. A more detailed analysis of the test data showed both the precision and limitations of the applied acoustic theory. Rods protruding 23 percent of the radius predominantly generated only lowest radial order modes, as expected. Measured and predicted radiation patterns were generally in good agreement. The agreement, however, depended on a significant degree of implied refraction due to inlet velocity gradients. Refraction, if present, would impact static-flight noise comparisons.
A comparison is made between approximate equations for far-field acoustic radiation patterns and exact equations for single and multimodal excitations in order to determine the validity range for the approximate approach. It is found that for single-mode cases: (1) the gross behavior of the primary lobes is adequately described by the approximate equations, (2) some error is found for lower and zero radial order modes, and (3) some agreement is yielded between exact and approximate sidelobes, although the approximate equation was not intended to simulate sidelobes. Multimodal approximate equations are compared to exact equations for various distributions of modal power; for all cases excellent agreement is found. For multimodal patterns it is noted that many modes influence the final level and shape of the directivity curve, although the major contributions are from the higher radial order modes.
Approximate equations for the far field acoustic radiation patterns in the forward quadrant from a flanged circular duct were compared with exact equations for both single and multimodal excitations. The single mode comparison showed good agreement between the exact and approximate equations for the principal lobes of higher radial order modes. For lower and especially for zero radial order modes, there was some error in the angular location and decibel level of principal lobe peak pressure obtained from the approximate equation. Some agreement of sidelobes was also observed although the approximate equation was not intended to simulate the sidelobes. The multimodal approximate summation equations consisting only of a simple function of directivity angle and an acoustic power biasing function were checked against the exact equations for several distributions of modal power and showed excellent agreement with exact equations for all cases. Although many modes contribute to the final level and shape of the directivity curve, the major contributions appear to come from the higher radial order modes.
Turbofan noise directivity calculated for two directivity models (equal modal amplitude and equal modal power or energy) was compared with experimental blade passing frequency data from two fans at 60 and 90 percent speeds. Experimental data indicated similar directivity patterns which were well represented by a single average data curve. Calculated points using the equal amplitude model showed over-prediction near the fan axis and near the 90-deg position. Calculated points using the equal power model showed a very good match with the average data lending support to theory of equipartition of modal power from a random source such as the interaction of the rotor with inlet flow distortion. The equal modal power model also gave good agreement with individual data points.
Turbofan noise directivity calculated for two directivity models was compared with experimental, blade passing frequency data from two fans at 60 and 90 percent speeds. Experimental data indicated similar directivity patterns which were well represented by a single average data curve. Calculated points using the equal amplitude model showed over-prediction near the fan axis and near the 90 degree position. Calculated points using the equal power model showed a very good match with the average data lending support to theory of equipartition of modal power from a random source such as the interaction of the rotor with inlet flow distortion. The equal modal power model also gave good agreement with individual data points.
Qualitative and quantitative spectral analyses are presented that give the broadband-noise, discrete-tone, and multiple-tone properties of the noise generated by a full-scale high-bypass single-stage axial-flow transonic fan (fan B, NASA Quiet Engine Program). The noise components were obtained from narrow-band spectra in conjunction with 1/3-octave-band spectra. Variations in the pressure levels of the noise components with fan speed, forward-quadrant azimuth angle, and frequency are presented and compared. The study shows that much of the apparent broadband noise on 1/3-octave-band plots consists of a complex system of shaft-order tones. The analyses also indicate the difficulties in determining or defining noise components, especially the broadband level under the discrete tones. The sources which may be associated with the noise components are discussed.
Design analysis and general characteristics of flat plate central fin-tube sensible-heat space radiators under thermodynamic and fluid mechanic conditions
Fortran program for sensible-heat space radiator design