Acoustic Characteristics of Flow Noise Generated from Air-Conditioning Duct of Rail Vehicle

Vehicle Engineering (VE) Volume 1 Issue 4, December 2013

www.seipub.org/ve

Acoustic Characteristics of Flow Noise Generated from Air-Conditioning Duct of Rail Vehicle Kunihiko Ishihara*1, Masashi Ichimiya2, Tomoji Yoshida3 *1

Department of Clinical Engineering, Tokushima Bunri University 1314-1 Shido, Sanuki, Kagawa 769-2193, Japan

Department of Mechanical Engineering, The University of Tokushima, 770-8506 Minami-Josanjima, Tokushima, Tokushima 769-2193, Japan *2

*3

Department of Clinical Engineering, Tokushima Bunri University 1314-1 Shido, Sanuki, Kagawa 769-2193, Japan

k-ishihara@fe.bunri-u.ac.jp; 2ichimiya@me.tokushima-u.ac.jp; 3yoshida@fe.bunri-u.ac.jp

*1

Abstract Noise from air conditioning duct of rail vehicle is annoying for passengers especially when the rail vehicle is in a resting state. Then the noise reduction is very important from the point of view of keeping the car room comfortable and is one of the differentiating techniques. This paper described experimental results of flow noise characteristics. The air conditioning duct has some holes to ventilate the car room.

In this study, the flow noise characteristics generated from the air conditioning duct have been studied experimentally by using the duct with four holes. As a result, it was clarified that the noise reduction effect of the absorbing treatment could be neglected when the noise source was made only by the flow and much when the fan noise is greater than the flow noise.

Keywords Flow Noise; Acoustic Characteristics; Straght Duct

Introduction Air conditioning equipment is usually set in rail vehicle in order to keep comfort. However, the noise it creates is offensive to the ear especially when the rail vehicle is in a resting state. Therefore, noise reduction is so important to keep the compartment comfortable, and is one of the differentiating techniques of rail vehicles.

The air conditioning equipment consists of a fan and ducts with holes. Noise sources are fan noise generated from the fan itself, flow noise generated from bended and expanded parts of the ducts which are made out of necessity, and flow noise generated from open holes in the ducts that is released directly into the compartment. However, the present situation cannot clarify which source is dominant.

Therefore in this study, a degree of contribution of

each source to the compartment noise was investigated by regarding a duct with holes as an air conditioning duct. There are many studies related to the noise generated from ducts [1]~[6]. However, studies on the acoustic characteristics and generated flow noise from ducts with open holes like air conditioning ducts are lack in number.

In a previous paper[7], acoustic characteristics of a duct with holes were studied and the difference of acoustic characteristics was clarified between ducts with and without holes. In this paper, flow noise characteristics of ducts with holes will be studied and the absorbing effect in existing flow will be also examined experimentally by using the straight duct with some holes.

Experimental Method Experimental Apparatus Figure 1 shows the experimental apparatus. In the present experiment, first of all, the flow noise characteristics of the straight ducts without holes were clarified, and secondly, the flow noise characteristics of ducts with four holes were obtained. Then, test ducts A and B were set in the nozzle of a low noise wind tunnel (cross section is 200 mmĂ—200 mm) and the flow noise was measured at positions as shown in Fig. 1. Next, the signals were processed by the FFT analyser. All of the sound pressure level were Aweighted here.

The low noise wind tunnel is an open flow type and the specifications are as follows; ď Ź

The background noise is 61 dB (A) at 500 mm from the center of open flow area in 50 m/s of wind velocity.

81

www.seipub.org/ve

Vehicle Engineering (VE) Volume 1 Issue 4, December 2013



The turbulent degree is less than 1% and nonuniformity of the average flow velocity at the open flow area is less than 1%.



The background noise is 23 dB (A) in no flow.

Where duct A (Fig. 1 (a)) has no holes and duct B (Fig. 1 (b)) has four holes in one side as shown in Fig. 1. Furthermore, the end of duct B is closed and the cross section of the duct is 200 mm×200 mm.



The microphone was positioned outside in a 45 degrees diagonal direction of 100 mm from the end in duct A as shown in Fig. 1 (a) and the microphones were positioned outside of 45 degrees diagonal direction of 100 mm from the open flow area in duct B as shown in Fig. 2. In addition, in the experiment of the duct A, the noise measuring position without the duct was outside of 45 degrees diagonal direction of 100 mm from the exit of wind tunnel.

t0 200 200

Mic1



Test total pressure: 50 Pa, 170 Pa, 430 Pa and 750 Pa



Open area size:

(a)Duct A Flow velocity measuring point

75mm Speaker

 Holes

Mic.4 Mic.3 Mic.2 Mic1 0.5m 1.0m 1.5m 2.0m 2.5m

(b) Duct B FIG. 1 TEST APPARATUS

Closed end of duct

Duct

No.1 Hole 45

100mm

45

Mic.

Flow FIG. 2 MEASURING POSITION OF DUCT B

1) Experimental Method The total pressure was measured at the position 82

200 200

The experimental parameters are as follows;

100mm

No.2 Hole

20 0 200

t0 (Punching metal)

FIG. 3 KINDS OF ABSORBING TREATMENT OF DUCT

Speaker

Pitot tube

t20 (Punching metal)

2) Experimental Parameters

Cross section

75mm

100 mm far from the upstream of the duct and the rotating speed of the fan was adjusted to obtain the test total pressures (See 2)) by watching the digital manometer. After obtaining the test total pressure, the measurement of the noise and the velocity were started. The outputs of each microphone led to FFT analyzer via the noise measuring system and then spooled up the personal computer after frequency analysis. The frequency range is 200 Hz～5000 Hz and the frequency resolution is 12.5 Hz. The velocity was measured at three points in the duct cross section and nine points in the open area as shown in Fig. 1(b) and they were averaged.



50 mm×50 mm (50 square),



75 mm×75 mm (75 square),



100 mm×100 mm (100 square)

Acoustic absorption treatment: (See Fig. 3) 

absorbing material with 20 mm thickness and specific weight is 48 kg/m3 (t20)



no absorbing material (t0)

Experimental Result and Discussions Noise due to Flow (Case of Duct A) Figure 4(a) and Fig. 4(b) show the sound pressure level (Overall value) and the root total pressure against the root velocity respectively. The parameters are existence or non existence of the absorbent and the perforated plate. The following findings can be obtained from Fig. 4(a) and Fig. 4(b). 

The sound pressure level with duct (t0) becomes 5~10 dB larger than that without duct due to the duct with 2.5 m length. Because the background noise is 23 dB (A) in this unechoic room, the sound pressure level at 5.5 m/s becomes 25 dB (A) in case of no duct. Therefore, the data is not on the straight line made by other points.

Vehicle Engineering (VE) Volume 1 Issue 4, December 2013

We will discuss the rise of the noise below.

∫

S

Where Sc is the correlation area of pressure fluctuation on the solid surface, ω is the angular frequency.The root mean square of pressure fluctuation on the wall in the plane boundary layer pw2 and the correlation area Sc

Sound pressure level dB(A)

80

The radiation acoustic power level due to the boundary layer noise is given as follows[8]: 1 2 (1) W= [ pw (y )] ω 2 Sc (y, pw )dS (y ) 3 12πρ c

www.seipub.org/ve

(a)

70 60 50 40

without duct t0(Punching metal t20(Punching metal) t0 U^7Law

30 20 10 0 1

10 Root velocity

100 U

m/s

(b)

are approximately expressed as follows: ρ  2 4.72 × 10−5  0 U 2  p= w  2 

2

(2)

ω 2 Sc = 0.338U 2 [8]

Where U is the flow velocity and ρ0 is the air density.



The flow velocity dependency follows the 7 powers law. The flow velocity dependency of the flow noise is said to be generally 5 ～ 8 powers law[6].



The sound pressure level in perforated plate (■) is 2～3 dB larger than that in flat and smooth plate (●). This is because the vortex is easy to generate the noise in perforated plate than in flat and smooth plate. And there is no difference between the sound pressure level with and without absorbent in the case of having the perforated plate. That is to say, the effect of the absorbent does not appear in that case. According to the investigation by Watanabe et.al, the acoustic power level at the end of the straight duct with arbitrary length L (m) LWL is given as follows.

(7) LwL =+ LW 1 10 log {(1 − e − kL ) / k }

FIG. 4 DEPENDENCY OF WIND VELOCITY ON SPL AND Pt

Substituting these equations to Eq. (1), the acoustic power becomes = W 1.06 × 10−7

ρ0 c3

(3)

U 6S

Where S is the radiation area. Expressing this equation by the power level, we have Eq.(4) ρ  LW = 50 + 10 log  30 U 6  + 10 log( S ) c 

(4)

On the other hand, the acoustic power level LW 0 when no duct is calculated by (5)

LW 0 = L p 0 + 20 log r + 11 (r = 0.2m)

And using the acoustic power level LWT when duct exists calculated by power summing Eq. (4) and Eq. (5), we can obtain the sound pressure level L pT when duct exist by the following expression.

L pT = LWT − 20 log r − 11 (r = 0.2m)

(6)

These values are shown in shadow part of table 1. Where S is the inside area of duct. That is to say, S=0.2m×4×2.5m=2m2.

These values are in good agreement with the measured values ( ● symbol in Fig. 4(a)).

Therefore, the power level of the boundary layer noise generated in the duct might be evaluated by the expression (4).

TABLE 1 SPL OF NOISE GENERATED FROM DUCT EXIT

Velocity 5.5 10 15 20

Lp0 25 33 50 64

Lw0 22.0 30.0 47.0 61.0

Lw 22.3 37.9 48.4 55.9

LwT 25.2 38.5 50.8 62.2

LpT 28.1 41.5 53.8 65.2

Where LW1: Power level of generated sound per unit length (dB/m); R: Attenuation per unit length in a duct (dB/m); k: Attenuation coefficient per unit length (=0.23R) According to the paper [7] previously published by the author, the attenuations per unit length are R=3 dB/m (without absorbent) and R=7 dB/m (with absorbent) respectively. The acoustic power level of generating sound can be obtained by using these values and Eq. (7), and the result is shown in Fig. 5.

Comparing the result of Fig. 5 and Fig. 4(a), it can be found that the effect of absorbent is hard to appear in the case of existence of flow. This is profoundly different from the speaker test result in which the effect of absorbent is large. This is because the sound source is generated in a duct from the inlet to the outlet of duct in the case of existence of flow.

83

www.seipub.org/ve

Vehicle Engineering (VE) Volume 1 Issue 4, December 2013

FIG. 5 COMPARISON OF POWER LEVEL OF DUCT WITH FLOW BETWEEN WITH AND WITHOUT ABSORBENT

Comparing results of with (■) and without (●) the perforated plate in the case of no absorbent in Fig. 4(a) and Fig. 4(b), the sound pressure Lp in the large total pressure is larger than that in the small total pressure at the same flow velocity. That is to say, the larger the pressure loss is, the larger the sound pressure is.

According tothe reference [9], the acoustic power level of generating sound due to the pressure loss is given by 30logξ. Where ξ=2ΔP/ρu2. The sound pressure rise calculated by the expression becomes as shown in table 2. ΔP is the pressure drop, namely pressure change due to the parameter change. For example, when U=10 m/s, ΔP=30log(72.6/51.8)=4.4 dB can be obtained between parameter 2 and 1 as shown in table 2. The result in the case of no absorbent (Right column) is different in value from the experimental result (value in ( )) and is the same in tendency. From this result, the small pressure loss, that is to say, to be small the flow separation is effective to reduce the sound pressure due to the flow. On the other hand, the inverse tendency could be seen in the case of existence of flow as can be seen in the second column from the right in table 2.

highest and it becomes lower while going upstream. This is because the velocity at the open area in downstream is large. The frequency where the sound pressure level is large becomes higher with a small open area size in a low frequency range. That is to say, they are 220 Hz in 50 mm×50 mm, 150 Hz in 75 mm×75 mm and 110 Hz in 100 mm×100 mm. The peak frequency of the jet noise is given by the expression f=StU/d (St=0.44[9]).

According to this expression, when an open area size is small, that is to say, when d is small the frequency becomes large. For example, the peak frequency in the case of 750 Pa and 50 mm×50 mm becomes 204 Hz (f=0.44×26/0.056) and this value is almost the same as the peak frequency 220 Hz of Fig. 6(a). Where d=0.056 m (Equivalent diameter) and U=26 m/s. (a) (a

)

220Hz

(b) (b

) 150Hz

TABLE 2 RELATION BETWEEN PRESUURE DROP AND SPL

U (m/s)

(1).t0 (Pa)

10 15 20

51.8 157 315

(2).t20 Punching (Pa) 72.6 192 355

(3).t0 Punching (Pa) 88.9 210 362

ΔLp (2)-(1) dB 4.4(1.3) 2.6(2.5) 1.6(2.8)

ΔLp (3)-(1) dB 7.0(3.2) 3.8(2.9) 1.8(1.9)

(c)

110Hz

Noise due to Flow (Case of Duct B) 1) Frequency Characteristics of Sound Pressure Level Fig. 6 shows the spectra of sound pressure at each open area outlet when the root pressure is 750 Pa in the duct without an absorbent. These figures are results of (a) 50 mm×50 mm. (b) 75 mm×75 mm and (c) 100 mm×100 mm respectively. From these figures, the tendency can be seen that the sound pressure level at Open1 (Downstream side) is the

84

FIG. 6 SPECTRA OF NOISE AT EACH HPLE IN CASE OF NO ABSORBENT

Fig. 7 shows the spectra of sound pressure at each open area outlet when the root pressure is 750 Pa in the duct with an absorbent of 20 mm thickness of glass wool. In the case of an absorbing treatment, the effect of the open area position on the sound

Vehicle Engineering (VE) Volume 1 Issue 4, December 2013

pressure level becomes small in a high frequency range and strongly appears in a low frequency range as same as the case of not being absorbent.

This might be concerning the fact that the effect of an absorbent is small in a low frequency range and large in a high frequency range. And the effect of the open area position on the sound pressure level becomes small when the open area size is small.

This is because the flow velocity is independent of the open area position when the open area size is small. (a)

www.seipub.org/ve

mm respectively. (a)

(b)

FIG. 8 7OVER ALL SPL OF EACH HOLE TO VELOCITY (50SQUARE)

(a)

(b)

(b)

FIG. 9 OVER ALL SPL OF EACH HOLE TO VELOCITY (75SQUARE)

(c)

(a)

(b)

FIG. 7 SPECTRA OF NOISE AT EACH HOLE IN CASE OF HAVING ABSORBENT

From these results, it is considered that the flow noise generated at the open area outlet is dominant in low frequency noise. On the other hand, the boundary layer noise is dominant in high frequency noise.

FIG. 10 OVER ALL SPL OF EACH HOLE TO VELOCITY (100SQUARE)

2) Flow Velocity Dependence of Open Area Noise

The (a) of each figure is the result of no absorbent and (b) is that of an absorbent of a 20 mm thickness glass wool. From these figures, the following findings can be obtained.

In order to grasp the acoustic characteristics of the noise generated from the open area, the overall sound pressure level raised to the maximum flow velocity has been examined against each open area size and its position. The results are shown in Fig. 8~Fig. 10. Figure 8, Fig. 9 and Fig. 10 are results of 50 mm×50 mm, 75 mm×75 mm and 100 mm×100

(1) The sound pressure level is larger with a larger open area size. This is because the sound intensity is proportional to the open area size for the same velocity later mentioned in 3).

Generally, the velocity at large hole is considered to be smaller than that of small hole.

However, in this case, the velocity is almost

85

www.seipub.org/ve

independent of hole’s area as shown in table 3.

The velocity is determined by the pressure difference between inlet and outlet. The inlet pressure is the root pressure and the outlet pressure is atmosphere. These pressure difference s are the same in three different hole areas. Then the noise from large area is larger than that from small area. For your information, seeing the case of Open 1 and no absorbent, the sound pressure levels are 79 dB in 50 mm×50 mm, 82 dB in 75 mm×75 mm and 85 dB in 100 mm×100 mm. These results follow the expression ⊿Lp=10log(S/S50). Where S and S50 are arbitrary open areas and open areas of 50 mm×50 mm respectively.

(2) The absorbing effect of the absorbent strongly appears insomuch as the small open area size.

This is because the absorbing force of the absorbent appears effectively insomuch as a small open area size as long as the absorbing rate of the open area is 1.0. The effect of absorbent does not almost appear in the case of 100 mm×100 mm.

(3) The absorbing effect of the absorbent strongly appears insomuch as the open area downstream. This is because the attenuation becomes large while going downstream.

(4) In total, the sound pressure level of Open1 (mostly downstream) is large and that of Open 4 (mostly upstream) is small. The reason is that the flow noise generated in the duct and the flow velocity of flowing out from holes become large with going downstram as shown in table 3 and Eq.8. Table 3 shows flow velocities from each hole to the root total pressure for three open areas. Eq.(8) shows the defference between the flow noise powers generated at Open1 (L=2 m) and Open4(L=0.5 m) in the duct.

For your information, substituting k=0.69 (without absorbent) and k=1.61 (with absorbent t20), We can get 4.1 dB and 2.4 dB as ∆Lw respectively. On the other hand, evaluating using the 6 power law and velocity defference between Open1 and Open4, we can get 7.5 dB in case of 100square and Pt=750 Pa in table 3(C).

It results in 11.6 dB(wihout absorbent) and 9.9 dB(with absorbent) in total. This values(red number in table 3(C)) correspond with the experimental results as shown in Fig.10. From these considerations, it can be understood that the sound pressure level of Open1 (mostly

86

Vehicle Engineering (VE) Volume 1 Issue 4, December 2013

downstream) is large and that of Open 4 (mostly upstream) is small. The tendency becomes conspicuous with a larger flow velocity. As a result, the flow velocity dependency of the sound pressure level (Exponent n in I ∝Un) becomes larger downstream. For your information, the flow velocity dependencies are of the 7 powers law for downstream cases and 6 powers law for upstream cases.

(8) ∆= Lw= LW ( L 2)= − LW ( L = 10 log {(1 − e −2 k ) / (1 − e −2 k )} 0.5) TABLE3(A) VELOCITY OF EACH HOLE(50SQUARE)

Root Pt 50 Pa 170 Pa 430 Pa 750 Pa

Root 1.9 3.4 5.9 8.0

Open1 7.73 14.1 20.7 26.3

Open2 7.62 13.9 20.4 25.4

Open3 7.56 13.5 20.2 26.0

TABLE3(B) VELOCITY OF EACH HOLE(75Square)

Root Pt 50 Pa 170 Pa 430 Pa 750 Pa

Root 3.5 6.8 11 14.3

Open1 7.40 13.4 20.0 26.6

Open2 7.29 13.2 20.3 26.1

unit:m/s

unit:m/s

Open3 7.49 13.0 20.2 25.2

TABLE3(C) VELOCITY OF EACH HOLE(100Square)

Root Pt 50 Pa 170 Pa 430 Pa 750 Pa

Root 5.5 10.0 15.1 20.0

Open1 7.14 13.1 19.1 24.0

Open2 7.02 12.0 17.9 22.6

Open4 7.18 13.1 19.8 25.5

Open4 7.63 12.8 20.1 23.8 unit:m/s

Open3 6.96 11.0 16.0 18.6

Open4 6.64 10.0 15.0 17.9

3) Relation between Open Area Size and Sound Pressure Level In order to examine which relates to the sound pressure level, the open area size or the circumferential length, the open part at the farmost end of the downstream which has the largest sound pressure level was chosen and examined as one example. The result is shown in table 4.

TABLE 4 RELATION BETWEEN SPL AND HOLE AREA

Kind Length AreaRatio Of Area Ratio (L) (S) 50□ 75□ 100□

1 1.5 2

1 2.25 4

10log 10log (L) (S) 0 1.75 3.0

0 3.5 6.0

Measured SPL difference 79dB(+0dB) 82dB(+3dB) 85dB(+6dB)

The differences of measured sound pressure level among the three open area sizes under the conditions without an absorbent and a maximum flow velocity (20m/s) are obtained as shown in table 4 (Right column). On the other hand, the differences of sound pressure level due to the difference of the circumferential length and the open area size are shown in the columns of 10log(L)

Vehicle Engineering (VE) Volume 1 Issue 4, December 2013

and 10log(S), respectively as shown in Table 4.

As can be seen from this result, the value of 10log(S) is in good agreement with the measured sound pressure level difference. Then it can be said that the sound pressure level is proportional to the open area size.

Flow Noise Characteristics of Bending Duct

www.seipub.org/ve

correspond to four different ducts shown in Fig. 12 and the numbers show the O.A. values (dB) of each duct. From these figures, it can be seen that the dominant frequency for high root pressure Pt becomes higher than that of low root pressure Pt. In order to examin this fact in details, the relation between the root velocity U and the dominant frequency fp was obtained.

90

In general, an air conditioning duct has a bending part.

It can be thought that the flow noise characteristics of the bending duct may be different from that of the straight duct. Then in this chapter, the flow characteristics of the bending ducts and the noise reduction effects of various countermeasures will be described. Fig. 11 shows the experimental setup. The bending duct has not any holes on the duct wall like the duct described before. The four different ducts as shown in Fig. 12 are used to examine the flow noise characteristics and reduction effects of the flow noise.

SPL [dB(A)]

30 20 10 10

10000

(a) (b) (c) (d)

100

200

A

A

100 1000 Frequency [Hz]

110

200

0.4m

Microphone Total pressure tube

50 40

0.4m Origin

60

71.4 70.5 65.3 61.1

FIG. 13 SPECTRUM OF SPL (Pt=170Pa)

Absorbent

Flow

70

SPL [dB(A)]

Speaker 75mm 0.65m

(a) (b) (c) (d)

80

0.1m

90

94.0 92.6 89.7 82.0

80 70 60 50 40 30

FIG. 11 TEST APPARATUS

10

100

Absorbent

1000

10000

Frequency [Hz]

FIG. 14 SPECTRUM OF SPL (Pt=750Pa)

(a) (b)

Guide

(c)

Airfoil

(d)

Fig. 15 and Fig. 16 show the relation between the root velocity U and the dominant frequency fp. The gradient of these figures is the Strouhal number defined by St= fpD/U and these values become 1.12 for the 1/2 duct and 0.93 for the total duct respectively.

That is to say, the dominant frequency of the bending duct can be predict by using St=1.0.

140 100

Frequency fp [Hz]

The experiment was carried out for two ducts. One is the same as the duct as shown in Fig. 11, called “Total duct”. The other is the duct which is removed A-A downstream side, called “1/2 duct”. Fig. 13 and Fig. 14 show the experimental results of the frequency characteristics of the flow noise of four different ducts for Pt=170 Pa and 750 Pa respectively. These results are for “1/2 duct”. The symbols (a) ~(d) in these figures

D=0.2m, 1/2 duct, St=1.2

120

FIG. 12 KINDS OF DUCT

80

(a)

60

(b)

40

(c)

20

(d)

0 0

5

10

15

20

25

Root velocity U [m/s]

FIG. 15 RELATION BETWEEN fp AND U (1/2DUCT)

87

www.seipub.org/ve

140

Conclusions

D=0.2m, Total duct, St=0.93

120

Frequency fp [Hz]

Vehicle Engineering (VE) Volume 1 Issue 4, December 2013

The flow noise characteristics of the lined ducts with four holes like air conditioning ducts were examined experimentally. As a result, the findings obtained were as follows:

100 80 60

(a)

40

(b)

20

(c) (d)

0 0

5

10 15 Root velocity U [m/s]

20



The absorbing effect of the absorbent is significantly reduced when comparing one obtained in the speaker test in the case of only the sound generated in the duct due to the flow being the sound source.



The sound pressure level at the open area is larger when going downstream and the absorbing effect of the absorbent is also larger when going downstream.



The sound pressure level is larger with the larger open area in the same flow velocity and the absorbing effect of the absorbent becomes smaller.



The flow velocity dependency ( Exponent n in I∝Un) is larger while going downstream. In the present experiment, n is 7 downstream and n is 6 upstream. Incidentally, n is 6 downstream and n is 5 upstream when measuring the sound

by C characteristics.



The flow noise generated at the open area exit is dominant in the low frequency noise. On the other hand, the boundary layer noise generated inside the duct is dominant in the high frequency noise.



The noise of the bending duct is about 30 dB larger than that of the straight duct and its cause may be thought of the flow separation.



The dominant frequency can be calculated by using St=1.0 and fp =StU/D.



The separation protection device like airfoil brings much reduction effects of flow noise.

25

FIG. 16 RELATION BETWEEN fp AND U (TOTAL DUCT)

Next the noise reduction effects of these four different ducts will be discussed.

Fig. 17 and Fig. 18 show the experimental results for the 1/2 duct and the total duct. The symbols (a)~(d) in these figures correspond to four different ducts as shown in Fig. 12. As can be seen in Fig. 17, the noise of the bending duct is about 30 dB larger than that of the straight duct. From these figures, it can be seen that the duct(c) is the best countermeasure to reduce the flow noise. That is to say, the flow noise can be suppressed by removing the flow separation by using the airfoil type devise and the effect of the absorbing material is not good for the flow noise reduction.

∝U7

Straight duct without absorbent(9)

FIG.17 SPL(O.A) FOR FOUR DUCTS (TOTAL DUCT) 110

SPL [dB(A)]

90

ACKNOWLEDGMENT

1/2duct

(a) (b) (c) (d)

100

This study was conducted under the grant of (C) of Research Grant-in aid for Scientific Research.

80 70

REFERENCES

60

∝U7

50

Kaname WATANABE et al., “A Study on Noise Generated

40

by the Air Flow in Duct Systems”, Journal of the

30

S.H.A.S.E. Vol.37, No.5 , 22-33 , 1963 1

10

100

Root velocity [m/s]

FIG.18 SPL(O.A) FOR FOUR DUCTS (1/2 DUCT)

88

Hiroyoshi SHIOKAWA, Morimasa ITAMOTO, “On Sound Characteristics of Straight Glass Fiber Ducts”, Journal of Archit Plann. Environ Engng, AIJ, No453, Nov., 9-

Vehicle Engineering (VE) Volume 1 Issue 4, December 2013

Ohyo Gizyutu Syuppan( in Japanese), 1987, 217

15,1993 Morimasa

www.seipub.org/ve

ITAMOTO,

Hiroyoshi

SHIOKAWA,

“On

O. Mochizuki, Y. Maruta, “Introduction to Flow Noise

Generated Noise by Air Flow at Straight Glass Fiber

Engineering―Obtaining

Ducts”, Journal of Archit Plann Environ Engng, AIJ,

Asakura Shoten, 6, 1996

No.428, Oct, 21-27, 1991 Morimasa ITAMOTO et al.” On Air Flow and Sound Characteristics of Straight and Bend Glass Fiber Ducts”, Proceedings of the 1989 meeting The institute of noise control engineering of Japan, 213-216, 1989 Morimasa ITAMOTO et al., “Study on Air Flow and Sound Characteristics of Straight and Bend of Lined Ducts”, Proceedings of the 1989 meeting The institute of noise control engineering of Japan, 459-460, 1990 INCE/J Edition, “Noise Countermeasure of Duct Sysyem”, Gihodo Syuppan ( in Japanese),1999 Kunihiko Ishihara, “Study on Acoustic Characteristics of Straght Duct with Some Holes”, Transaction of the JSME, 74-734, 332-338, C(2008) K. Shiraki et al., “Noise Protection Design and Simulation”,

rich

sound

environment”,

Kunihiko Ishihara born in Kurashiki City, Okayama Prefecture, Japan in1947, and graduated from Kobe University in 1969 and Graduate school of Kobe University in 1971; then he got Ph.D of Engineering from Osaka University in 1985. He has been studying Mechanical Vibration and Noise Control, especially Flow-Induced Vibration and Noise.

He worked in Kawasaki Heavy Industry for thirty three years and researched vibration and noise problems of many products. After that, he became a professor of The University of Tokushima and has been studying vibration and noise problems for eight years. He has published many academic articles more than 80. Almost of these articles were listed in Journal of JSME (Japan Society of Mechanical Engineer).

Professor Ishihara is a member of JSME and ASME, and he won many awards.

89