**7.1 Velocity distribution**

A straight length in front of a butterfly valve is needed to obtain a normal velocity distribution as shown in Fig.11(a). In a turbulent flow, the entrance length to obtain fully developed flow is about fifty or one hundred times the pipe diameter. However in practical engineering, more than two valves are installed in series or bends are installed just ahead of a valve to save space. Under such conditions, the upstream velocity distribution is very different from the normal turbulent velocity distribution. In this study, the cavitation noise measurement was performed under such a velocity distribution that there was a large velocity difference between the nozzle side and the orifice side.

Fig. 11. Velocity distribution upstream of the butterfly valve and the Circular Plate for biased velocity.

(The nozzle side velocity is larger than distribution the orifice side velocity.)

A straight length in front of a butterfly valve is needed to obtain a normal velocity distribution as shown in Fig.11(a). In a turbulent flow, the entrance length to obtain fully developed flow is about fifty or one hundred times the pipe diameter. However in practical engineering, more than two valves are installed in series or bends are installed just ahead of a valve to save space. Under such conditions, the upstream velocity distribution is very different from the normal turbulent velocity distribution. In this study, the cavitation noise measurement was performed under such a velocity distribution that there was a large

 (a) Normal velocity distribution (b) Biased velocity distribution (A) (The orifice side velocity is larger than the nozzle side velocity.)

0

r/R

1

<sup>0</sup> <sup>2</sup> <sup>4</sup> <sup>6</sup> <sup>8</sup> <sup>10</sup> -1

velocity(m/s)

 (c) Biased Velocity distribution (B) (d) The Circular Plate for biased velocity (The nozzle side velocity is larger than distribution the orifice side velocity.)

Fig. 11. Velocity distribution upstream of the butterfly valve and the Circular Plate for

**7. The effect of the upstream velocity distribution on butterfly valve** 

velocity difference between the nozzle side and the orifice side.

<sup>0</sup> <sup>2</sup> <sup>4</sup> <sup>6</sup> <sup>8</sup> <sup>10</sup> -1

velocity(m/s)

<sup>0</sup> <sup>2</sup> <sup>4</sup> <sup>6</sup> <sup>8</sup> <sup>10</sup> -1

velocity(m/s)

**cavitation** 

0

r/R

1

biased velocity.

0

r/R

1

**7.1 Velocity distribution** 

A plate with a half circle hole shown in Fig.11(d) was installed at 3D (D: pipe diameter) upstream of the valve to obtain an biased velocity distribution. The velocity distributions were analyzed by numerical simulation using a commercial code Star-CD because the measurement of the velocity distribution was difficult behind such plate. Fig.11 shows three types of velocity distributions. The vertical axis shows the location divided by the radius of the pipe and the horizontal axis shows velocity. Fig.11 (a) shows an example of a normal turbulent velocity distribution at 1D upstream of the valve. Fig.11 (b) and Fig.11 (c) show the biased velocity distributions. In each example, the average velocity was 3 / *m s* . In Fig.11 (a), the velocity distribution mostly agreed with Blasius's law. In Fig.11(b), the velocity on the orifice side of the valve was larger than that on the nozzle side of the valve. The velocity distribution of Fig.11(c) is the reverse of Fig.11 (b).

In many practical industry plants, the flow rates are controlled by adjusting the opening of the butterfly valve and are determined by head curves of pumps and pressure losses of the whole piping systems. Accordingly, the effect of valve shape on cavitation noise should also be investigated under the condition of a constant pressure loss coefficient.

In the flow of the normal velocity distribution, the cavitation at the orifice side becomes severe since the contraction at the orifice side becomes intense and the pressure around the edge becomes very low. As shown in Fig.12, the cavitation inception of the Type-B valve was observed at 38 , but in contrast the cavitation inception of the normal valve was observed at 58 . Moreover, the increase of cavitation noise of the Type-B valve was suppressed after inception. As for the maximum noise, the Type-B valve was lower than the normal valve. Therefore, the effect of semi-circular fins is clear.

Fig. 12. Cavitation Noise under normal velocity distribution ( 45 , 12 ).

Figure 13(a) shows the cavitation noise in the flow of the biased velocity distribution (A). In this case, the flow rate of the orifice side was much larger than that in the normal velocity distribution. Accordingly, the cavitation noise of both the normal valve and Type-B valve were increased remarkably in the range of 30 to 50 compared with the results of

Noise Reduction in Butterfly Valve Cavitation

(a) Inception condition

except for the near wall.

diameters larger than 230

The inception condition was at 48

to be photographed.

diameters over 100

Fig. 14. Cavitation Bubbles of the normal valve ( 45

**7.2.2 Diameters and numbers of cavitation bubbles** 

*m* , and a bubble with a diameter between 50

Bubbles with diameters were smaller than 20

normal velocity distribution, the diameters ranged mainly from 20

*m* increased at 35.9

Fig.15(b) had already become more intense at the same cavitation number.

condition reached the flashing condition at 24.2

*m* . For example, a bubble with a diameter between 20

30

by Semicircular Fins and Visualization of Cavitation Flow 495

Under the flashing condition as shown in Fig.14(c), a large cavity was formed behind the valve, therefore the individual cavitation bubbles were not be able to be photographed

Fig.15 shows the diameters and numbers of cavitation bubbles of the normal valve under the normal velocity distribution. The diameter of the bubbles are shown at 8 stages at every

experiments. Such size nuclei are presumed to be contained even in water under noncavitation conditions, and will not affect cavitation noise greatly. The bubbles with

number of bubbles is shown as zero because the individual cavitation bubbles were not able

in Fig.15(a). However, under the biased velocity distribution (A), the cavitation bubbles with

the cavitation condition was in the growth stage and the cavitation bubbles occurred numerously behind the valve body. Though it looks as if the amount of the bubbles decreased compared with Fig.8, this is due to the fact that the photographing was performed at one position. Compared with the results of Fig.15(a), the cavitation noise in

).

(c) Flashing condition

50 (b) Growth condition

50

*m* did not occur in our experiments. In Fig.14, the cavitation

*m* and 80

and the flashing condition was at 25

*m* and 50

. As for the flashing condition, the

as shown in Fig.15(b). Under this condition,

*m* could not be visualized in our

*m* is expressed as 65

*m* to 100

*m* is expressed as 35

. Under the

*m* as shown

*m* .

50

Fig. 13. Cavitation Noise under biased velocity distribution.

Fig.12. However, the noise of the Type-B was smaller than that of the normal valve. The effect of the fins remained in the biased velocity distribution.

Figure 13(b) shows the cavitation noise in the flow of the biased velocity distribution (B). In this case, the cavitation noise was relatively smaller than that in the biased velocity distribution (A), but much larger than that in the normal distribution. However, the noise of the Type-B was smaller than that of the normal valve. Therefore, it is clear that the effect of the semi-circular fins holds even in the biased velocity distribution and that the cavitation noise is larger when the velocity distribution is biased.
