5.1 Experimental Arrangement (1)
To validate the FDNESI technique, practical sources including a scooter and a wooden box model with a loudspeaker fitted inside were chosen as the test targets for experiments. Figure 8 shows the block diagram of the experimental arrangement. In the scooter experiment, the array configuration is a 4 4 URA, while in the wooden box experiment, the array configuration is a 5 6 URA. Two PXI 4496 systems [35]
in conjunction with LabVIEW [35]were used for data acquisition and processing at the sampling rate 5 kHz. A bandpass filter (20 Hz~1.7 kHz) is used to prevent aliasing and errors occurring in the out-of-band frequencies. The source amplitude, sound pressure, particle velocity and sound intensity reconstructed using FDNESI can be displayed on the monitor.
5.2 Scooter Experiment
In the experiment, a 125cc scooter served as a practical source to examine the capability of FDNESI in dealing with non-stationary sources. The scooter is mounted on a dynamometer inside a semi-anechoic room. The array parameters are selected to be M = J = 4 4 , d = df = 0.1m = λ/2 for 1.7 kHz and L = d/ 2 . The Frequency-domain FDNESI was used to reconstruct the sound field on the right side of the scooter in a run-up test. The engine speed increased from 1500 rpm to 7500 rpm within ten seconds. The unprocessed sound pressure received at the microphones is shown in Fig. 9(a), while the rms velocity reconstructed using the FDNESI is shown in
22
Fig. 9(b). These results revealed that the cooling fan behind the vented engine cover was the major noise source. Next, the virtual microphone technique is employed to see if it is possible to further enhance the image quality by increasing the number of channels from 4 4 16 to 11 11 121 . The inverse filters have been designed in the previous numerical investigation. The particle velocity was then reconstructed on the basis of the estimated source amplitude, as shown in Fig. 9(c). Total sound power level is 195 dB re. 1 10 12W. Clearly visible is a larger area of image with improved resolution than that of Fig. 9(b), where again the cooling fan is the major noise source.
5.3 Wooden Box Experiment
In this experiment, a wooden box model with loudspeaker fitted inside is used to validate the FDNESI technique by using a 5 6 URA. As shown in Fig. 10(a),
Figure 10(a) shows the unprocessed sound pressure picked up at the microphones within the band 200 ~ 1600 Hz. From the image, the noise sources were barely resolvable, particularly for the noise source at the edge - the circle, the slot and the square at upper left corner. Also, the square at the center was difficult to distinguish.
Virtual microphone technique was again applied to overcome this problem by interpolate and extrapolate the pressure field on the microphone surface and increase the number of microphones and focal points from 5 6 to 13 15 . With the new
23
setting, the particle velocity (rms) reconstructed using the frequency-domain FDNESI is shown in Fig. 10(b). It can be clearly observed from the result that the quality of the reconstructed image was significantly improved. Problems due to edge effect and insufficient resolution were basically eliminated. The FDNESI images apparently yielded more reliable information about noise sources than the unprocessed sound pressure. corresponding to fmax= 1.7 kHz) in the x and y directions and the distance between source surface and array surface L=0.1m. Two PXI 4496 systems [35] in conjunction with LabVIEW software [35] were used for data acquisition and processing at the sampling rate 5 kHz. A bandpass filter (20 Hz~1.7 kHz) is used to prevent aliasing and errors occurring in the out-of-band frequencies.
5.5 Loudspeaker experiment
In this experiment, two loudspeakers are used to validate the Fourier NAH, FDNESI, DAS, TR, MVDR and MUSIC algorithms in the nearfield with comparison.
The loudspeakers are situated at (0.1m, 0.1m) and (0.4m, 0.2m) on the source surface that produce random noise band-limited to 1.7 kHz. The observed frequencies in all the algorithms are chosen to be 1.2 kHz. The source images obtained by using the
24
six acoustic imagine algorithms with 5 6 URA are shown in Fig. 11(a)-(f). From the reconstructed sound pressure of Fourier NAH and FDNESI in Fig. 11(a)(b), Fourier NAH can find out the two sources roughly while FDNESI have clearer source distribution than Fourier NAH. From Fig. 11(c)(d), DAS have a poor result that is unable to distinguish the two sources and TR can localize the two source positions but with big main lobes. In high resolution MVDR and MUSIC algorithms, they can semi-anechoic room that the major noise is at the air intake position on the top of the compressor and the minor noise is low intensity vibration at overall body. Different with loudspeaker experiment, the source of this experiment is not on the planar surface. The observed frequencies in the algorithms are chosen to be 1.2 kHz. The noise images obtained by processing of the six algorithms with URA are shown in Fig.
12(a)-(f). From Fig. 12(a), Fourier NAH has a terrible noise source distribution, consistent with the theory that the source should be planar in Fourier NAH to identify successfully. From the reconstructed sound pressure of FDNESI shown in Fig. 12(b), FDNESI can identify the major source at the air intake and the vibration at overall body. The result of DAS is bad by wrong location and very big main lobe but TR provide an acceptable result as shown in Fig. 12(c)(d). In the noise images of MVDR and MUSIC, they identified the noise source at the air intake accurately and the result of MUSIC is aim at the major source.
25
5.7 Vibrations of free square plate experiment
The experiment is conducted to validate the FDNESI on detecting the mode shapes of vibrating plate. An aluminum plate with 0.2m side length and 0.002m thick was used to in the experiment. The aperture of the plate and the 5 6 URA are the same with the microphone spacing dx = 0.04m, dy= 0.05m in x and y directions.
The distance between source surface and array surface L = 0.02m and the retreat distance is Lr= L/ 2 = 0.01m. The edges of the plate were unbounded and the plate was excited by a shaker from its center using a random noise signal. The experiment was undertaken in a semi anechoic room and the experimental arrangement is shown in Figure 13. In order to confirm the experimental result, the salt is sprinkled on the vibrating plate for visualize the nodal lines, the salt is thrown off the moving regions and piles up at the nodes. The frequencies were chose by the peaks of the microphone’s frequency response. By exciting the plate at the frequencies we chose, we can find three mode shapes at 204 Hz, 226 Hz and 595 Hz, the real part of particle velocity in frequency domain after the processing of FDNESI are shown in Fig.
14(a)-(c). For getting better resolution of the results, interpolation of virtual microphone is applied in FDNESI at the frequencies of 204 Hz and 595 Hz that the result is shown as Fig. 14(d)(e). From the Fig. 14(a)(d), we can obtained a consistent result with simulation. From the experimental results that we can conclude the FDNESI can handle the non-destructive evaluations.
26