Microphone Array Measurements for High-speed Train

Size: px

Start display at page:

Download "Microphone Array Measurements for High-speed Train"

Lambert Clement Perry
6 years ago
Views:

1 Microphone Array Measurements for High-speed Train Korea Research Institute of Standards and Science Hyu-Sang Kwon

2 2 Contents Railway Noise Sound Images Flow Noise

3 Railway Noise Measurement & Analysis v Railway rolling noise sleeper rail wheel sleeper rail wheel Cited from Railway noise and vibration written by D. J. Thompson Cited from a paper written by D. J. Thompson et al. (JSV, 193, pp , 1996) 3

4 Railway Noise Measurement & Analysis v Railway rolling noise ü Main sources Sleeper (< about 500 Hz) Rail (about 500 Hz ~1.5 khz) Wheel (> about 1.5 khz) v Rolling noise measurements ü Microphone array is often used to localise sources. ü However, measured results give less prominence to the rail and more to the wheel. ü It was reported that sound power measured by microphone array gives at least 10 db less than the theoretical ones. 4

5 Railway Noise Measurement & Analysis v Theoretical modeling (cont.) ü Rail noise prediction Series of coherent monopoles along rail D 5

6 Railway Noise Measurement & Analysis v Rail noise predicted ü Stationary load Distance from rail (m) Hz 400 Hz Distance from rail (m) Single load (stationary) at 400 Hz, Distance along rail (m) 800 Hz 1600 Hz Single load (stationary) at 800 Hz, Single load (stationary) at 1600 Hz, 5 5 Distance from rail (m) Distance from rail (m) Distance along rail (m) Distance along rail (m) 6

7 Railway Noise Measurement & Analysis v Rail noise predicted ü Stationary load Distance from rail (m) Freqency 125 = 125Hz Distance along rail (m) 800 Hz Freqency = 800Hz Distance along rail (m) Distance from rail (m) Distance from rail (m) Hz Freqency = 400Hz Distance along rail (m) 1600 Hz Freqency = 1600Hz Distance along rail (m) 7

8 v Rail noise predicted ü Moving load Railway Noise Measurement & Analysis Distance from rail (m) Hz 400 Hz Single load (moving, V = 360 km/h) at 125 Hz, Distance from rail (m) Single load (moving, V = 360 km/h) at 400 Hz, Distance along rail (m) 800 Hz 1600 Hz Single load (moving, V = 360 km/h) at 800 Hz, Distance along rail (m) 5 5 Distance from rail (m) Distance from rail (m) Distance along rail (m) Distance along rail (m) 8

9 Railway Noise Measurement & Analysis v Rail noise predicted ü Moving load with V = 111 m/s (400 km/h) Distance from rail (m) Hz 400 Hz Freqency = 125Hz, V = 100 m/s Distance along rail (m) 800 Hz Distance along rail (m) Distance from rail (m) Distance from rail (m) Freqency = 400Hz, V = 100 m/s Distance along rail (m) 1600 Hz Freqency = 1600Hz, V = 100 m/s Distance along rail (m) 9

10 Railway Noise Measurement & Analysis v Remarks ü Rail becomes a line source above 300 Hz ü For stationary load, Maximum pressure occurs 40 ~20 between 300~3000Hz Phase angle of the rail noise (plane wave) becomes around 20 between 300~3000Hz ü For a moving load of 400 km/h, Rail noise is radiated more to negative angles Maximum pressure -45 ~ - 30 between 300~3000Hz Phase angle of the rail noise (plane wave) becomes around 20 between 300~3000Hz at negative angles 10

11 Railway Noise Measurement & Analysis v Sound pressure from microphone array ü Microphone array configuration 17 microphones with 0.11m interval Distance from rail: 5m Force Cited from Kitagawa s PhD thesis 11

12 Railway Noise Measurement & Analysis v Beam forming of microphone array ü Delay and sum in time domain ( ) ( ) ( ) ( ) ( ) 1 ( ) M æ rij ö si t» rij p j ç t + M j= 1 è c ø å ( ) 2 2 r = x - x + r ij i j 0 12

13 Railway Noise Measurement & Analysis v Beam forming of microphone array ü Plane wave or spherical wave models Cited from Kitagawa s PhD thesis 13

14 Railway Noise Measurement & Analysis v Doppler effect due to a moving source Moving source Q observer R, Q s ~ function of time 14

15 Railway Noise Measurement & Analysis v Rail noise measurements using a microphone array ü Measurement at a test track 15

16 Railway Noise Measurement & Analysis v Measurement setup ü Setup The number of microphones 11 Microphone spacing m, m Spatial weighting function Distance from rail Excitation frequency Dolph-Chebyschev m 500, 625, 800 Hz (with m) 990, 1250, 1580, 2000, 2500 Hz (with m) m Force 16

17 Railway Noise Measurement & Analysis v Pressure distribution and directivity ü Results (90 ) 500 Hz Sound Pressure Level(dB) set 1 set 2 set 3 set 4 Sound Pressure Level (db) set 1 set 2 set 3 set Distance from rail (m) Beam Angle (degree) 800 Hz Sound Pressure Level(dB) Freq = 792 Hz set 1 set 2 set 3 set 4 Sound Pressure Level (db) Freq = 792 Hz set 1 set 2 set 3 set Distance from rail (m) Beam Angle (degree) 17

18 Railway Noise Measurement & Analysis v Remarks ü For a moving load of 400 km/h, the rail noise radiated to the negative angle is about 5 db louder that that to the positive angle ü Plane wave vs Spherical wave models Spherical wave model is not good for estimating the phase angle Spherical wave model, however, is not much different from the plane wave model in terms of the noise level ü Test track measurements were carried out to validate the features of rail noise predicted theoretically ü However, the measure results do not show consistent results and seem to be less reliable ü More powerful exciter may be demanded to obtain reliable results from the test track measurement 18

19 Sound Images (Microphone Array) v Measurements on an operational track with HEMU ü Array microphone configuration Ground level 19

20 Sound Images (Microphone Array) v Measurement configuration (plane view) Center Line Rail head End Trigger Start Trigger Microphone Microphone Array 20

21 Sound Images (Microphone Array) v Measurement setup Virtual source plane & relative coord. Q Q Measured mic. signals (140 ch.) - Positions of mic. array & moving source - View angle & signal measuring duration, array aperture, 21

22 Sound Images (Microphone Array) v Analyzed results ü All microphones of view angle 44 (train passage 15.5 m) (a) Reconstructed source image (whole freq. range) (b) Reconstructed source image (freq. range: 800 ~ 1500 Hz) (c) Reconstructed source image (freq. range: 200 ~ 400 Hz) 22

23 Sound Images (Microphone Array) v Analyzed results ü Whole freq. range, view angle 44 (train passage 15.5 m) (a) Reconstructed source image (use of all (140) microphones) (b) Reconstructed source image (use of 128 microphones) (c) Reconstructed source image (use of 116 microphones) (d) Reconstructed source image (use of 104 microphones) 23

24 Sound Images (Microphone Array) v Analyzed results ü Whole freq. range, all microphones (a) View angle : 44 (b) View angle : 30 (c) View angle : 15 (d) View angle :

25 Sound Images (Microphone Array) v Remarks ü Pass-by noise of HEMU measured on an operational track were analysed ü Data measured by a spiral two dimensional array with 140 microphones were investigated varying the view angles and array size ü It was found that the rail is displayed as a noise source in the scanned image ü However, the expected features of the rail noise predicted theoretically were not clearly shown in this measured data analysis ü More systematic and severe analysis for this measured data is required to identify the key parameters and configurations to measure rail noise with array microphones 25

26 26 Flow Noise Measurement & Analysis v Flow induced noise in high-speed train

27 Flow Noise Measurement & Analysis v Flow Noise Sources Pantograph FSI Inter-coach WPF in TBL 27

28 Flow Noise Measurement & Analysis v Acoustic Analogy ü For noise generated by flow over a surface F i (y,t) Observer q(y,t) 2 2 p x i p t = q + F x Curle s acoustic analogy (Proc Roy Soc, A231, 1955) i i + 2 x T i ij x j 28

29 Flow Noise Measurement & Analysis v Complete Governing Equations Theoretically, Fluid Flow r + Ñ ( rv) = 0 t ρv B + Ñ ( ρvv-τ) = f t re B + Ñ ( rve -t v + q) = f v + q t r = r( p, q ) and e = e( p, q ) B + Structure J u ( ) F u&& 2 mñ i + l + m + i = r xi + Compatibility equations Numerically, FVM FEM M t+dt U&& ( i) t+dt ( i) t ( i) t+dt t ( i-1) + C U& + KDU = R - F Mx && + Cx& + Kx = F - G T Ap 29

30 Flow Noise Measurement & Analysis v Simple Example 30

31 Flow Noise Measurement & Analysis v Flow noise source model: Corcos Model Uc p q Fully developed turbulent boundary layer Frozen flow 31

32 Flow Noise Measurement & Analysis v Measurement on the running train (planning) 32

(a) Flush mounted mic. array (horizontal 41 & vertical 13 mics.

33 Flow Noise Measurement & Analysis v Measurement setup ü Design of Wall plate Considering point pressure, spatial correlation, damping, etc. (a) Flush mounted mic. array (horizontal 41 & vertical 13 mics.) (b) Flush mounted mic. array (horizontal 21 & vertical 7 mics.) (c) A mic. at the center 33

34 Flow Noise Measurement & Analysis v Flow Noise Measurement by Microphones ü Flush mounting ü Pinhole mounting 34

001 m a = 1/8/2 in a = 1/4/2 in a = 1/2/2 in Pressure Sensitivity 0.6 0.4 0.2 Mic.

35 Flow Noise Measurement & Analysis v Microphones on the wall ü Flush mounting Sensitivity of microphone (circ. area) a = m a = 1/8/2 in a = 1/4/2 in a = 1/2/2 in Pressure Sensitivity Mic. Sensitivity relative to radius of diaphragm Frequency [Hz] 35

36 Flow Noise Measurement & Analysis v Microphones on the wall ü Pinhole mounting Helmholtz resonator Res. freq. of pin hole w.r.t. the radius of mic. a =11mm, L=5mm, h=5mm Res. freq. of pin hole w.r.t. the depth of neck φ=1/4in., a=1mm, h=5mm Res. freq. of pin hole w.r.t. the cavity depth in front of mic. φ=1/4in., a=1mm, h=3mm 36

37 Flow Noise Measurement & Analysis v WT experiments Measurement section (size): 300mm(W) X 300mm(H) X 600mm(L) Used microphone (ECM/ HJ06) 37

Flow Noise Measurement & Analysis v WT experiments ü Flow velocity fan (rpm) Flow vel. (m/s) 0 0 218 5 399 11.7 572 17.3 ü 743 23.3 910 29.2 976 31.

38 Flow Noise Measurement & Analysis v WT experiments ü Flow velocity fan (rpm) Flow vel. (m/s) ü Background noise measurement ü suction fan noise ü fan controller noise ü vibration induced noise ü Outside of the test section Power spectral density, (db, re. 20m Pa) m/s 5 m/s 12 m/s 17 m/s 23 m/s 29 m/s 31 m/s Frequency (Hz) 38

39 Flow Noise Measurement & Analysis v Measured spectra ü Flush mounting Power spectral density (db, re. 20mPa) m/s 5 m/s m/s 17 m/s 0 23 m/s 29 m/s 31 m/s Frequency (Hz) Power spectral density (db, re. 20mPa) m/s 5 m/s m/s 17 m/s 0 23 m/s 29 m/s m/s Frequency (Hz) HJ07 microphones Type 46AD microphones 39

40 Flow Noise Measurement & Analysis v Measured spectra ü Pinhole mounting Hz 900 Hz 120 Power spectral density (db, re. 20mPa) m/s 5 m/s m/s 17 m/s 0 23 m/s 29 m/s 31 m/s Frequency (Hz) Power spectral density (db, re. 20mPa) m/s 5 m/s m/s 17 m/s 0 23 m/s 29 m/s 31 m/s Frequency (Hz) 5 mm spacing between pinhole and mic. 10 mm spacing between pinhole and mic. 40

41 Flow Noise Measurement & Analysis v Measured data analysis 110 Channel : 1~ Channel : 22~41 Power spectral density (db, re. 20m Pa) Frequency (Hz) Power spectral density (db, re. 20m Pa) channel no. increases Frequency (Hz) 90 Channel : 1~20 90 Channel : 22~41 Power spectral density (db, re. 20m Pa) Power spectral density (db, re. 20m Pa) Frequency (Hz) Frequency (Hz) 41

Flow Noise Measurement & Analysis v Analysis (Convective velocity) ü Freq.

(rad/m) 400 300 200 Wavenumber (rad/m) 400 300 200 Wavenumber (rad/m) 400 300

(Hz) 0 0 500 1000 1500 Frequency (Hz) fan (rpm) Flow vel. (m/s) Conv. Vel.

42 Flow Noise Measurement & Analysis v Analysis (Convective velocity) ü Freq.-wavenumber analysis m/s 23 m/s 29 m/s Wavenumber (rad/m) Wavenumber (rad/m) Wavenumber (rad/m) Frequency (Hz) Frequency (Hz) Frequency (Hz) fan (rpm) Flow vel. (m/s) Conv. Vel. (m/s)

43 Flow Noise Measurement & Analysis v Analysis (Corcos model parameters) ü Coherence 43

44 Flow Noise Measurement & Analysis v Analysis: Flow induced vibration due to WPF ü Flow induced vibration of flat plate: Corcos model Cross power spectral density function Theoretical vibration response induced by TBL WPF 44

45 Flow Noise Measurement & Analysis v Analysis: Flow induced vibration due to WPF ü Flow induced vibration of flat plate: Corcos model Cross power spectral density function Cross power spectral density function 45

46 Flow Noise Measurement & Analysis v Comparison of microphone mounting conditions ü Spectral analysis relative to flow vel. (10, 20, 30 m/s) ü Measurement conditions Lines: 3200, Span: 25.6 khz, df: 8 Hz, Averages: 300 Overlap: 66.67%, Recording time: 12.58s, High Pass Filter: 7 Hz

47 Flow Noise Measurement & Analysis v Measurement setup (microphone arrays for flow noise) ü Flush mounting & Pinhole mounting simultaneously ü Horizontal 6 pts. & vertical 3 pts. (10 mm spacing)

48 Flow Noise Measurement & Analysis v Measurement setup (microphone arrays for flow noise) ü Reducing the noise of sonic nozzle (64 m/s) ü 1600 spectral lines, 25.6 khz span, 500 average,..

49 Flow Noise Measurement & Analysis Convection Velocity from Phase of Cross Spectrum 2 < CrossSpectrum between CH1 and CH5 > = ( ) Flow velocity : = [ ] Phase [rad] = ( ) Flow velocity : = [ ] Frequency Distance [Hz m]

50 Flow Noise Measurement & Analysis Coherence = w d / U c Stream-wise Decay Rate from Coherences

51 Future works Time Resolved Lasers 1kHz, max 20kHz) High Speed Camera(3260 fps, 1280 px X 800 px) PIV Experimental set-up 51

52 52

Localizing Noise Sources on a Rail Vehicle during Pass-by

Localizing Noise Sources on a Rail Vehicle during Pass-by J. Gomes 1, J. Hald 1 and B. Ginn 1 1 Brüel & Kjaer Sound & Vibration Measurement A/S, Skodsborgvej 307, DK-2850 Naerum, Denmark E-mail: Jesper.Gomes@bksv.com