Study on the Analogue Feedback Active Soft Edge Noise Barrier

Study on the Analogue Feedback Active Soft Edge Noise Barrier

1Jacob Chia-chun Liu，2Po-Chien Lu
Department of Water Resources and Environmental Engineering
Tamkang University，Taiwan


Abstract: It is feasible to enhance the effect of reduction of the noise barrier on low frequency noise using the active control system. On the basis of the forerunners’ research, an Analogue feedback active control system that is easy to be carried out was arranged deployed on the top of a soundproof wall noise barrier to form an Analogue feedback active soft edge noise barriersoundproof wall in order to enhance the effect of reduction of the noise barriersoundproof wall on low frequency noise. Through a preliminary experimental study, this method was proved to be feasible and the following preliminary conclusions on the deploymentarrangement of active control systems were made: (1) when control sources were arranged deployed in a single row, the control effect was determined by the density of deploymentarrangement: the higher the density, the better the control effect; (2) with the same density of deploymentarrangement, the control effect of the double- row deploymentarrangement was better than that of the single- row deploymentarrangement; (3) with the same number of control sources, the single- row deploymentarrangement had better control effect than the double- row deploymentarrangement.
Key Words: Analogue feedback active soft edge noise barriersoundproof wall; new excess insertion loss
1. Introduction
Along with the social development, traffic noise pollution is becoming more and more serious. It is a recommendable way to reduce traffic noise with the soundproof wallnoise barrier. The noise reduction effect of the soundproof wallnoise barrier is related to the frequency characteristics of the noise. Both theoretical analyses and fieldexperimental measurements manifest that the noise reduction effect of the barrier on low frequency noise is poorer than that on high frequency noise. With regard to the noise with bigger low frequency component, in order to get a better noise reduction effect, it is necessary to increase the height of the soundproof wallnoise barrier, hence increasing the construction cost and inducing other difficulties in project implementation. However, active noise control system has a better noise reduction effect on low frequency noise. So the active control system is used to increase the noise control effect of the noise barrier.
Presently, a large number of studies are employing the active noise control system to enhance the noise reduction effect of the soundproof wallnoise barrier [1---7]. Omoto [1,2] validated the availability of the active control system for the barrier. Studies on increasing the noise reduction effect of the barrier with the active control system when the ground reflection exists is extended by Guo [3] through numerical simulations and experiments respectively. Guo pointed out that the ground reflection decreases the effect of the active control system to noise reduction, but the active control system is effective to decrease the diffracted sound all the time. About the influence of the geometry shape of the secondary sources, Shao [3] pointed the arc-type arrangement is more effective than line-type with the same number of the secondary sources through the numerical simulations. Based on Shao’s results, Yang [5] discussed how to choose the cost function when the secondary sources is arranged the arc-type. Recently, Ohnishi [6] et al creatively applied the active control system on the noise barrier along the 20 m’s high-way to reduce the noise of the “dark area”. The active control system gained 2 dB excess insertion loss in the factual situation. Berkhoff [7] used near-field error microphone to obtain the far-field error signals so as to control the far-field noise. This method improved the effectiveness of the active control system. The performance of the control algorithm is compared for far-field error signals, near-field error signals, and virtual far-field error signals. The simulations have shown that the performance of the virtual far-field error signals can gain the same results as the far-field error signals. Niu [8] et al gave out the error sensoring optimization in active noise barrier system.
In terms of control methods, there are two kinds of control modes adopted by the active controller: digital control and Analogue control. In digital control, the DSP chipset is used to process the signals collected by the error sensor and the adaptive algorithm is adopted to perform computation to produce appropriate counteracting signals so as to achieve a better noise reduction effect, while in Analogue control, a compensation network composed of Analogue circuits is used to reduce noise. Although the noise reduction effect of the digital control is better than that of the Analogue control, implementation cost of the digital control is too high to be practicable. Since the implementation cost of the Analogue control is considerably low, the Analogue active noise control system is usually adopted for occasions when a great quantity of active noise control systems are needed, such as active soundproof wallnoise barriers.
Based on the conclusion on of the error sensoring optimization of error sensor arrangement by Niu et al [8], Tthis paper studies the control effect of the Analogue feedback active soundproof wallnoise barrier on low-frequency noise through experimentation, analyzing the influence of the Analogue feedback active control system on the overall system in terms of channel number and deploymentarrangement method respectively.
2. RationalePrinciple
The rationaleprinciple block diagram of the active noise control system adopting Analogue circuit feedback control is shown in figure 1. The control sound signals sent by the control loudspeaker – u(t) and the external noise signals – d(t) (the noise signals at the error sensor before control) are added in the error sensor to produce the output signals – e(t) (the noise signals at the error sensor after control). Through feedback control circuits, these output signals produce secondary source input signals – u(t), which are added to the noise signals at the error sensor after control. This process continues until the noise at the error sensor is reduced to be within a reasonable scope. In figure 1, G(s) represents the transfer function of the feedback control circuits; K represents amplifier gain; C(s) represents the transfer function of the control loudspeaker. As shown in figure 1, the relation between the noise signals at the error sensor before and after control is [8]：
(1)
Assuming that the power spectral density of the external noise signals is , according to equation (1), the power spectral density of the noise signals at the error sensor after control is：
(2)
According to equation (2), if open loop gain is , the power spectral density of the noise signals at the error sensors after control – will reduce to a very low level, indicating the noise is well reduced. Presently, can be approximately shown as：
(3)
The noise reduction after control is defined as:
(4)
According to equation (4), the higher the open loop gain is, the bigger the noise reduction. However, due to the requirement for system stabilization, the open loop gain – K cannot increase infinitely. When the open loop gain – K increases gradually to a certain frequency point and the following two conditions are coincidently satisfied：
(5)
(6)
the system may become unstable and the self-excited oscillation (a phenomenon namely “howling”) may occur (the open loop transfer function of the active noise control system – usually has low-pass characteristic, therefore as the open loop gain – K gradually increases, the frequency point where the first “howling” occurs – is usually the lowest frequency point that satisfies the phase condition (5)). In phase condition (5), represents the phase response of the control loudspeaker at howling frequency point – while represents the phase response of the feedback controller at howling frequency point – . According to equation (4), the noise reduction of the system at frequency point ω0 is：
(7)
When the open loop gain – K increases to an extent where howling of the system is imminent, the noise reduction at the frequency point – ω0 is maximized. Combining equation (6) into equation (7), the following equation is reached：
(8)
According to equation (8), the maximum noise reduction reached at the frequency point – ω0 is determined by the difference value of gain of at frequency points ω0 and . The higher the difference value of gain is, the bigger the noise reduction. Therefore, it is possible to increase the difference value of gain of at frequency points ω0 and through adjusting the frequency response of the control loudspeaker and the feedback controller .
3. Experiments
This experiment was conducted carried out in a fullthe anechoic chamberof the institute of Acoustics . In order to meet the requirements of a semi-anechoic chamber, the chamber was laid with anti-moisture board, which has a density of 14.1kg/m2 and a sound insulation of 25dB at 160Hz, to serve as rigid floor. The measurements of the inner space of the resultant semi-anechoic chamber are 11.2m×7.8m×5m. The soundproof wallnoise barrier, which was set at a place 4 meters from the inner wall of the chamber, was 1.22m in height and made of double-layer anti-moisture board with a sound insulation of 31dB at 160Hz. Comparing with the diffracted sound at the sound shadow area of the soundproof wallnoise barrier, the transmitted sound can be ignored. As for the establishment of the coordinate system, the zero point was chosen at the center of the cross-line of the barrier and the floor; the positive direction of x-axis was the direction of the medial axis of the chamber facing the door; the positive direction of y-axis was the direction pointing from left to right after entering the chamber; the positive direction of z-axis was the direction pointing upwards from the zero point (as shown in figure 2). On top of the barrier, a thin plate of 24cm in width was placed for deploymentarrangement of the 16 loudspeakers (secondary source). The error sensors were arranged deployed on the center axis of the secondary source one to one correspondence.
The test equipment used was PULSE system from B&K: the measurement sensor was B&K4190 Smart Microphone, the primary source was a loudspeaker of 32cm in diameter, and the power amplifier was YE2706A.
With reference to previous studies, the error sensors of the feedback Analogue active noise control system were arranged deployed at the place 0.08m above the secondary sources [98], which were evenly arranged deployed on top of the barrier. The loudspeakers used in the experiment had resonant frequency around 150Hz. The noise chosen was one-third octave bandwidth white noise with 160Hz center frequency.
The primary sources were placed at (-2, 0, 0.16)m. The monitoring points chosen were: point 1 (4, 0, 0.1)m; point 2 (5, 0, 0.1)m; point 3 (6, 0, 0.1)m; point 4 (4, 0, 0.5)m; point 5 (5, 0, 0.5)m; point 6 (6, 0, 0.5)m. The loudspeakers of the secondary source, 6m in total length, were evenly arranged deployed on top of the barrier and both sides of x-axis. The sound pressure level of the diffracted sound at the sound shadow area behind the barrier of the Analogue feedback active control system before and after control was measured to obtain the new excess insertion loss of the Analogue feedback active control system.
4. Results and discussion
This section depicts the influence on the control effect of the Analogue feedback active noise control system when the control sources are arranged deployed in a single row or in a double- row array.
4.1 The influence on the control effect of the Analogue feedback active noise control system when the control sources are arranged deployed in a double- row array
The control sources of the Analogue feedback active noise control system were arranged deployed in a double- row array. In this experiment, 16 channel units were arranged deployed in two rows with 8 units in each row. The total length of each row was 6 meters. The distance between the centerlines of the two rows of channel units was 550mm. The two rows of channel units were symmetrically situated on both sides of the soundproof wallnoise barrier. For the sake of convenience, the two rows of channel units were defined as “near row” and “far row” based on their distance (near or far) from the primary source. Meanwhile, in this experiment, when the “near row” and “far row” were working simultaneously, the system was defined as “far-near system”; when only the “near row” was working, as “near system”; when only the “far row” was working, as “far system” (as shown in figures 3 and 4). The new excess insertion loss induced by these three deploymentarrangement methods was compared. The noise chosen was one-third octave bandwidth white noise with 160Hz center frequency.
As shown in figure 5, the new excess insertion loss was the highest when two rows of active systems were working simultaneously. When only one row of active system was working, the effect of the “near row” system was better than that of the “far row” system. Therefore, in terms of the deploymentarrangement of control sources in an active system, if the deploymentarrangement density of a single row is fixed, the effect of noise reduction is better with double- row deploymentarrangement than with single- row deploymentarrangement. This is because with double- row deploymentarrangement, the number of control sources of the active control system is bigger than that with single- row deploymentarrangement.
Figure 6 shows the comparison of new excess insertion loss between single- row deploymentarrangement and double- row deploymentarrangement with the same number of control sources. As shown in figure 6, the control effect is better with single- row deploymentarrangement than that with double- row deploymentarrangement. This is because the higher density of control sources with single- row deploymentarrangement enhances the control effect of the overall system.
4.2 The influence of the number of control sources on the control effect of the Analogue feedback active noise control system when the control sources are arranged deployed in a single- row
This experiment compared the influence on new excess insertion loss induced by the active control system of the soundproof wallnoise barrier when the channel number of the Analogue feedback control system changed. It also compared difference of new excess insertion loss, induced by the active control system, when the Analogue feedback active controllers arranged deployed on top of the barrier contained 16 or 10 channels (see figure 7 for a picture of the experimental system).
As shown in figure 8, it is feasible to enhance the effect of reduction of the barrier, made by the soundproof wallnoise barrier, on low frequency noise using the Analogue feedback active noise control system. According to figure 8, one-third octave bandwidth at 160Hz has the best effect. In comparison between the feedback active control systems with 10 channels or 16 channels, the new excess insertion loss induced by the 16-channel system is 3dB higher than that induced by the 10-channel system, which is in consistence with the previously mentioned theory that says the noise reduction effect is related to the frequency characteristics of the loudspeakers chosen. The channel number affects the control effect of the Analogue feedback active control system, the more the channels, the better the control effect. Particularly, at the resonant frequency of the sources, where the control effect is the best, the influence of channel number on the control system is even more obvious.
5. Conclusions
This paper studied the deploymentarrangement methods of the active control system of the Analogue feedback active soundproof wallnoise barrier. Base on the results of the preliminary experiment, following deploymentarrangement rules were proposed: (1) when control sources were arranged deployed in a single row, the control effect was determined by the density of deploymentarrangement: the higher the density, the better the control effect; (2) with the same density of deploymentarrangement, the control effect of the double- row deploymentarrangement was better than that of the single- row deploymentarrangement; (3) with the same number of control sources, the single- row deploymentarrangement had better control effect than the double- row deploymentarrangement.
Acknowledgement
The author would like to thank the anonymous reviewers for their helpful comments and suggestions.
References
[1] A. Omoto and K. Fujiwara, A study of an actively controlled noise barrier, J. Acoust. Soc. Am.，1993；94: 2173-2180.
[2] A. Omoto, K. Takashima and K. Fujiwara, Active suppression of sound diffracted by a barrier: An outdoor experiment, J. Acoust. Soc. Am.，1997；102(3): 1671-1679.
[3] J. Guo and J. Pan, Increasing the insertion loss of noise barriers using an active-control system, J. Acoust. Soc. Am.，1998；104(6): 3408-3416.
[4] J. Shao, J. –Z. Sha and Z. –L. Zhang, The method of the minimum sum of squared acoustic pressures in an actively controlled noise barrier, J. Sound Vib.，1997；204(2): 381-385.
[5] J. Yang and W. –S. Gan, On the actively controlled noise barrier, J. Sound Vib.，2001；240(3): 592-597.
[6] K. Ohnishi, T. Saito and S. Teranishi et al. “Development of the product-type active soft edge noise barrier,” Proceedings of the18th International Congress on Acoustics, 1041-1044, Kyoto, Japan(2004).
[7] A. P. Berkhoff, Control strategies for active noise barriers using near-field error sensing, J. Acoust. Soc. Am. 2005；118(3): 1469-1479.
[8] Feng Niu, Haishan Zou and Xiaojun Qiu et al. Error sensor location optimization for active soft edge noise barrie, J. Sound Vib.，2007；299(2): 409-417.
[9] C. Bao and J. Pan, Active feedback control in enclosures, Proceedings of the 5th International Congress on Sound and Vibration, 279-286 (1997).

Figures
Figure 1: RationalePrinciple block diagram of the feedback active noise control system


Figure 2: Coordinate graph of the anechoic chamber



Figure 3: Diagram of active system in double- row deploymentarrangement



Figure 4: Picture of active system in double- row deploymentarrangement



Figure 5: Comparison of new excess insertion loss of the active system with the same single- row density


Figure 6: Comparison of new excess insertion loss between single- row and double -row deploymentarrangements with the same channel number


Figure 7: Picture of the Analogue feedback active noise control system in single- row deploymentarrangement



Figure 8: Comparison of new excess insertion loss of the feedback active control systems with different channel number