1、I - ” h N OL U I ATMOSPHERIC ABSORPTION OF HIGH FREQUENCY NOISE AND APPLICATION TO FRACTIONAL-OCTAVE BANDS F. Dozglus Shields und H. E. Buss I Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-CR -2 760 Subtitle TECH LIBRARY KAFB, NM 1. Report No. 2. G
2、overnment Accession No. 3. Recipien lInIwIIIIIIll 00b1400 “- NASA 4. Title and 5. Report Date June 1977 ATMOSPHERIC ABSORPTION OF HIGH E PREQUENCY NOISE AND APPLICATION. TO FRACTIONAL -OCTAVE BANDS 1 6. Performing Organization Code 7. Author(s) 8. Performing Organization Report No. F. Douglas Shield
3、s and H. E. Bass 10. Work Unit No. 9. Performing Organization Name and Address University of Mississippi University, ,Mississippi 38677 I NAS3 -19431 I 13. Type of Report and Period Covered 12. Sponsoring Agency Name and Address National Aeronautics and Space Administration Washington, D. C. 20546 1
4、4. Sponsoring Agency Code I 15. Supplementary Notes Final Report. Project Manager, Orlando A. Gutierrez, V/STOL and Noise Division, NASA Lewis Research Center, Cleveland, Ohio 16. Abstract Pure tone sound absorption coefficients have been measured at 1/12 octave intervals from 4 to 1OOlcHz at 5.5 K(
5、lOO F) temperature intervals between 255.4 and 310.9 K(Oo and 100 F) and at 10 percent relative humidity increments between 0 percent and saturation. The measure- ments were made in a large cylindrical tube (i. d., 25.4 cm; length, 4.8 m). Special solid- dielectric capacitance transducers, one to ge
6、nerate bursts of sound waves and one to terminate the sound path and detect the tone bursts, were constructed to fit inside the tube. The absorption was measured by varying the transmitter receiver separation from 1 to 4 m and observing the decay of multiple reflections or change in amplitude of the
7、 first received burst. The resulting absorption was compared with that from a proposed procedure for computing sound absorption in still air, and the agreement was quite good. Absorption of bands of noise was numerically computed by using the pure tone results. The results depended on spectrum shape
8、, on filter type, and nonlinearly on propagation distance. For some of the cases con- sidered, comparison with the extrapolation of ARP -866A showed a difference as large as a factor of 2. However, for many cases, the absorption for a finite band was nearly equal to the pure tone absorption at the c
9、enter frequency of the band. A recommended prediction procedure is described for 1/3 octave band absorption coefficients. 17. Key Words (Suggested by Authods) Atmospheric sound absorption I Unclassified - unlimited 18. Distribution Statement Band absorption Atmospheric acoustics 1/3 Octave predictio
10、n procedure I 19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. NO. of PWS 22. mice Unclassified Unclassified 238 Ail * For sale by the National Technical Information Service, Springfield, Virginia 22161 Provided by IHSNot for ResaleNo reproduction or networking permitt
11、ed without license from IHS-,-,-r Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-TABLE OF CONTENTS Section Page 1 SUMMARY 1 . 2 INTRODUCTION 3 3 THEORY OF SOUND ABSORPTION IN AIR 6 3.1 Sound Absorption Mechanisms 6 3.2 Simplified Expressions for Pur
12、e Tone Absorption 11 3.3 Comparison with Prior Experimental Results . 15 4 EXPERIMENTAL PROCEDURE . 16 4.1 Experimental System 16 4.1.1 Sound source and microphone transducers . 16 4.1.2 Electronic equipment and sound burst generation . 21 4.1.3 Temperature and humidity systems 22 4.2 Test Procedure
13、 . 4.2.1 Signal handling procedure 4.2.2 Data analysis to obtain measured absorption coef- f icients 4.3 Correction for the Tube 5 DISCUSSION OF EXPERIMENTAL RESULTS . 5.1 Pure Tone Results . 5.2 Error Analysis . 6 ABSORPTION FOR BANDS OF NOISE 6.1 Analysis . 6.1.1 Known source spectrum 24 25 28 30
14、37 37 42 48 48 48 ili . . Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Page 6.1.2 Known received spectrum . 55 6.1.3 Correcting to standard conditions 56 6.2 Numerical Integration . 57 6.3 Simplified Technique for Estimating Band Loss . 57 6.4 Use
15、 of Tables to Determine Band Loss 61 6.5 Use of Graphs to Determine Band Loss 62 6.5.1 Known source spectrum 62 6.5.2 Known received spectrum . 66 7 PREDICTION PROCEDURE 72 7.1 Recommended Prediction Procedure 72 7.2 Comparison with ARP.866A 74 8 CONCLUDING REMARKS . 75 APPENDIX A . EXPERIWTAL AND C
16、ALCULATED ABSORPTION . 76 A.l Figures and Tables for Pure Tone Absorption 76 A.2 A Point by Point Comparison of Measured Values of Total Absorption in Original and Check Runs . 190 APPENDIX B . COMPUTER PROGRAMS USED IN THE STUDY 192 B.l Program Used to Acquire and Analyze Experimental Data . 193 B.
17、2 Program Used to Correct Data for Tube Losses 196 B.3 Subroutine AIRAB Used to Compute Pure Tone Absorption Coefficients 200 B.4 Programs Used to Compute Band Loss Coefficients 200 APPENDIX C . TABULATED LOSS COEFFICIENTS FOR BANDS OF NOISE . 209 C.l Tables.of Band Loss Corrections (A) 209 C.2 Corr
18、ection to Standard Atmospheric Conditions . 224 iv Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Sect ion Page APPENDIX D - SYMBOLS LIST. . 229 REFERENCES. 233 V Provided by IHSNot for ResaleNo reproduction or networking permitted without license f
19、rom IHS-,-,-ATMOSPHERIC ABSORPTION OF HIGH FREQUENCY NOISE AND APPLICATION TO FRACTIONAL-OCTAVE BANDS BY F, DOUGLAS SHIELDS AND HI E, BASS DEPARTMENT OF PHYSICS 8 ASTRONOMY THE UNIVERSITY OF MISSISSIPPI UNIVERSITY, MISSISSIPPI 38677 1 I SUMMARY This report presents the results of a NASA-Lewis sponso
20、red study of atmospheric absorption of noise in the frequency range of 4 kHz to 100 kHz, for temperatures from 255.4OK (OOF) to 310.9OK (10O0F) and at relative humid- ities from 0% to saturation. The measurements were made in a large cylindri- cal tube (25.4 cm I.D. by 4.8 m long). Special solid-die
21、lectric capacitance transducers were constructed which fit inside the tube. One of these trans- ducers generated bursts of sound waves and was mounted so that it could be moved inside the large sound tube. A second transducer of similar construc- tion terminated the sound path and detected the tone
22、bursts. The absorp- tion was determined from the decay rate for the burst measured as a func- tion of the propagation distance as the burst bounced back and forth in the tube. Pure tone absorption coefficients were measured at 1/12 octave intervals from 4 kHz to 100 kHz. The temperature was varied i
23、n 5.5K (1O0F) intervals from 255.4OK (OOF) to 310.9OK (100OF). The relative humidity was varied in 10% increments from 0% to saturation. The resulting absorption was compared to a proposed procedure for computing sound absorption for pure tones in still ah and the agreement was found to be quite goo
24、d under most conditions. The results for absorption of pure tones were then applied to the prediction of attenuation of bands of noise. The band absorp- tion was found to depend significantly on the shape of the noise spectrum and the type of filter used as well as the atmospheric conditions and pro
25、pa- gation distance. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-It was also found that the band loss coefficient does not depend on propagation distance in a simple way. However, for many cases considered the deviation between the 1/3 octave ban
26、d loss and the readily computed pure tone absorption coefficient at the center frequency of the band was found to be small. This report presents the proposed procedures for calculating pure tone and broad band atmospheric attenuation as well as the experimental data obtained. 2 Provided by IHSNot fo
27、r ResaleNo reproduction or networking permitted without license from IHS-,-,-2, INTRODUCTION Since the early measurements of Duff (ref. 1), absorption of sound in air has proven to be a fertile field of scientific investigation. The first systematic measurements were make by Knudsen (ref. 2) in the
28、1930s. The observed absorption was explained theoretically by Kneser (ref. 3) in terms of viscous and thermal conduction losses (classical absorption) and vibrational relaxation of oxygen. This theory did mch to explain the effect of humidity on the relaxation absorption. This and later work were st
29、imulated by studies of architectural acoustics, hence the frequency range of primary interest was that important in auditorium design, i.e., 200 Hz to 10 kHz. Greenspan (ref. 4) measured the absorption of sound in air at high frequencies (greater than 1 MHz) and established that rotational relaxatio
30、n also makes a significant contribution to sound absorption even at low frequencies. In the 1950s, increased interest in community noise in the frequency range from 100 Hz to 1 kHz and 1arge.propagation distances prompted further measurements. It was recognized that the simple model of Kneser did no
31、t provide reliable predictions under these conditions. As a result, in 1964, Committee A21 of the Society of Automotive engineers issued an empirical prediction procedure (ref. 5) which provided a significant improvement over the basic pro- cedure of Kneser. As is the case with any empirical techniq
32、ue, the accuracy of the prediction procedure was limited by the data on which it was based. In 1967, Harris (ref. 6) devised an improved empirical technique based on the impressive amount of data which he had collected. This data was limited to frequencies below 15 kHz, therefore, predictions based
33、on this empirical technique at higher frequencies or at largely different atmospheric conditions could not be considered reliable. Since 1967, several major developments have occurred which increase the accuracy of absorption predictions. In 1969 Piercy (ref. 7) recognized that vibrational relaxatio
34、n of nitrogen is a major source of absorption at audible frequencies. Monk (ref. 8) and Evans, et.al., (ref. 9) considered a kinetic model for air absorption which included 3 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-the effects of simultaneous
35、 relaxation of nitrogen and oxygen. During this period, experimental studies of atmospheric absorption under a wider variety of atmospheric conditions. (although still a limited frequency range) were accumulating. In 1971, the S1 Committee of the American National Standards Institute (ANSI) appointe
36、d the S1-57 Working Group to examine theoretical and experimental knowledge of sound absorp- tion in still air. This working group, chaired by Dr. Joseph Piercy, developed a prediction technique for pure tone absorption which is based on the fundamental physics of sound absorption and available expe
37、rimental data. This procedure is empirical only .in the sense that measured sound absorption was used to determine the microscopic energy transfer rates or vibrational relaxation times. Since it is firmly based on physical principles, there is no reason why this technique can not be applied outside
38、the region of conditions spanned by present experimental data. However, the numerical parameters used in the procedure becorhe less certain for frequencies above 10 kHz and temperatures far above or below 294.3“K. So far, only pure tone absorption has been considered. In principle, the absorption of
39、 a band of noise can be predicted if the variation of the pure tone absorption coefficient with frequency is known. However, in practice the process of converting from pure tone values to loss coefficients for bands of noise involves numerical evaluation of an integral. In order to avoid this compli
40、cation, ARP-866A (ref. 5) recommends using the pure tone absorption coefficient at band center at frequencies up to 4 kHz and at some frequency lower than the center frequency for bands with a center frequency above 4 kHz. This process recognizes that most spectra are falling off rapidly at high fre
41、quencies but is at best a first approximation. For most noise control applications, frequencies above 4 kHz are not very important so the procedure used in 3) where ., T = temperature in OK. -. Equations (3.1) and (3.2) can be simplified by making some approximations- First, if we use the Euken expr
42、ession, K = (15Rl.1/4) 4cv/ (15R) + 3/51 J- (kg-mole)-l-oK-kg-m-l-sec-l (3.4) where R= universal gas constant in J-(kg-mole) -OK -1 -1 and with values of y,c and c for air, equation (3.1) becomes P V a = /(Pc)1(1.881), 2 -1 c1 nepers-m . (3.5) Recognizing that with these substitutions, arotaC1 = 0.0
43、681 Zrot, (3.6) we can write the sum of classical and rotational relaxation absorption,a CR a = the experimental work reported here represents 6,847 points. Following the publication of the report mentioned above, three other papers have appeared which provide some additional information on sound ab
44、sorption is still air. In reference 12 Bass and Sutherland report a review of very high frequency (1 mHz to 100 dz) sound absorption in air over a range of temperatures (295OK to 773OK) which allowed them to deter- mine the temperature dependence of the rotational. relaxation time of air hence aCR.
45、Their results include those of Bass and Keeton (ref. 14) and are reflected in equation (3.15). A more recent paper by Bass, Keeton, and Williams (ref. 15) examines the temperature dependence of the vibra- tional relaxation frequency for oxygedwater vapor mixtures. Their results are consistent over t
46、he range with the lack of a temperature dependence in equation (3.16) of temperatures of concern in this report (255OK to 310OK). 15 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-4, EXPERIMENTAL PROCEDURE 4,l EXPERIMENTAL SYSTEM The experimental po
47、rtion of this study consisted in measuringsound absorption in air at frequencies from 4 kHz to 100 kHz at 1/12 octave intervals and at temperatures from 255.4OK to 310.9“K at 5.5OK intervals and relative humidities from 0% to 100% at 10% intervals. The system used to make these measurements is diagr
48、ammed in figure 4.1. A series of bursts of plane sound waves from 2 to 10 milliseconds long was generated at the sound source and then allowed to reflect back and forth in the sound tube. The tube was made of aluminum and has a 25.4 cm internal diameter, a 0.95 cm wall and is 4.8 meters long. There are suitable systems to control and maintain uniformity in temperature and relative humidity along the entire tube length. The amplitude of the sound wave in each echo bui-st was measured at each reflection by the microphone that terminates the opposite end of the sound tube. The path lengt
