**2. Reduction of cavitation noise by using fins**

Figure 1 shows a typical case of intense cavitation. A butterfly valve is not usually used under such conditions, but this cavitation appears when a valve repeatedly opens and closes. The separation of the flow and the vortex region around the butterfly valve during light or moderate cavitation is easily presumed in this case.

In Fig. 1, the upstream side appears on the left. The leading edge of the butterfly valve is on the upper left and the trailing edge (nozzle side) is on the lower right. The flow from the lower left region (nozzle side) of this photograph passes along the wall surface and interferes intensely with the flow from the upper left (orifice side) at the top of the pipe. It has been pointed out (Itoh et al.,1988) that this interference from both flows causes intense cavitation and brings about the erosion of the wall surface. Therefore, the fins were attached to the valve body in this study to avoid the interference caused by these flows. The interference of the flows and the separation of the flow behind the valve were expected to be reduced when the fins were attached to the downstream surface of the valve.

Fig. 1. Cavitation around a butterfly valve (Valve opening:45deg, Flashing Condition).

al.,1988). To avoid this interference, semicircular fins were attached to the valve body. In this paper, it was confirmed based on the experimental results that the attachment of the fins

In this study, cavitation bubbles were photographed by Hi-speed camera and the size and number were measured from those photographs and the effect of the fins and the upstream velocity distribution were investigated. In past studies, photographs focusing on the aspect of the butterfly valve cavitation have been reported in great numbers. For example, it was pointed out that the most erosive cavitation around a butterfly valve is the vortex cavitation on the orifice side by means of the pressure-sensitive films and high speed photography (Tani et al,1991). An observation of butterfly valve cavitation and the measurement of cavitation noise were performed to diagnose the cavitation condition (Kimura and Ogawa, 1986). However, since the measurements of number and size of the cavitation bubbles were not carried out, the details of the cavitation growth were not clear. In this study, close-up photographs of cavitation bubbles were taken and their number and size were analyzed. The difference

In actual piping arrangements, straight lengths in front of butterfly valves are not sufficient to obtain normal velocity distribution in many cases. In extreme cases, more than two valves are installed in series or bends are installed just ahead of or behind those valves. Under such conditions, the upstream velocity distribution is different from the usual turbulent velocity distribution. Therefore, under the condition that the velocity distribution is biased, confirmation

Figure 1 shows a typical case of intense cavitation. A butterfly valve is not usually used under such conditions, but this cavitation appears when a valve repeatedly opens and closes. The separation of the flow and the vortex region around the butterfly valve during

In Fig. 1, the upstream side appears on the left. The leading edge of the butterfly valve is on the upper left and the trailing edge (nozzle side) is on the lower right. The flow from the lower left region (nozzle side) of this photograph passes along the wall surface and interferes intensely with the flow from the upper left (orifice side) at the top of the pipe. It has been pointed out (Itoh et al.,1988) that this interference from both flows causes intense cavitation and brings about the erosion of the wall surface. Therefore, the fins were attached to the valve body in this study to avoid the interference caused by these flows. The interference of the flows and the separation of the flow behind the valve were expected to be

between the normal valve and the valve with the fins is reported in the following.

of whether or not the attachment of fins is useful for the noise reduction is necessary.

reduced when the fins were attached to the downstream surface of the valve.

Fig. 1. Cavitation around a butterfly valve (Valve opening:45deg, Flashing Condition).

**2. Reduction of cavitation noise by using fins** 

light or moderate cavitation is easily presumed in this case.

was very effective in reducing cavitation noise.

Figure 2 shows the test valves. The valve shown in Fig.2(a) is a normal valve without a fin

Fig. 2. The test valves with fins.

The author clarified that the fin must be installed in the downstream of the valve in order to reduce cavitation (Ogawa & Uchida,1995). In this study, a test valve is called "TYPE-A" when a semicircular fin is attached to the downstream surface of the valve. When two semicircular fins are attached to the downstream surface of a valve, the test valve is called "TYPE-B". On the "TYPE-C" test valve, three semicircular fins are attached to the downstream surfaces of the valve. For each valve, each fin was fixed perpendicular to the valve stem.

Cavitation noise measurements were performed in a closed-type cavitation tunnel, using water as the fluid. There was a pump on the downstream side of the test section. The upstream pressure was kept at atmospheric pressure and the valve opening was fixed during the experiment. The flow velocity was increased by controlling the frequency of the pump using an inverter. Cavitation noise was measured at each flow velocity and a visualization was created by a high-speed camera. Cavitation noise was measured using a noise meter placed close to the outside surface of the test section duct. The frequency range of this noise meter was 20-8000Hz.

Noise Reduction in Butterfly Valve Cavitation

**pressure loss coefficient)** 

test valves are nearly equal.

constant pressure loss coefficient condition.

Fig. 4. Pressure Loss Coefficients for each Valve Type.

0

10

20

Pressure Loss Coefficient (

30

40

50

ζ)

was intense.

by Semicircular Fins and Visualization of Cavitation Flow 487

In the cases of TYPE-A and TYPE-C, the cavitation number at cavitation inception was larger than that in the case of the normal valve. The fin promoted the occurrence of cavitation because the middle fin was fixed on the orifice side, where cavitation occurrence

**5. Effect of the fins on cavitation noise reduction (Comparison with constant** 

The results shown in Fig. 3 were obtained by measuring cavitation noise at each flow velocity, while flow velocity was increased using a pump controlled by an inverter. The upstream pressure was kept at atmospheric pressure in each experiment, and the effect of the saturated steam pressure was minimal. Therefore, the flow velocity was also almost the same when the cavitation number was same. However, in many cases, the flow rate was not controlled by changes in the frequency of the pump, and the flow rate was controlled by adjusting the opening of the valve. Additionally, the flow rate in the actual plant was determined by the head curve of the pump and the pressure loss of the piping system as s whole. Accordingly, the effect of cavitation control should also be investigated under the

When a fin is attached to a valve body, the flow rate may be changed by the variation in the pressure loss of the valve. Figure 4 shows the pressure loss coefficients for each type of valve. For an example, the pressure loss of TYPE-C was larger than that of the normal valve at the same Reynolds number and at the same valve opening. When the butterfly valve is used to control the flow rate, the valve opening of TYPE-C must be larger than that of the normal valve in order to obtain the same flow rate. Therefore, the experimental results must be compared with a constant pressure loss coefficient. The following discussion is based on a comparison of cavitation noise under the condition that the pressure loss coefficients of the

1 2

Reynolds Number (Re)

[10<sup>5</sup> ]

**Normal(**θ**=35**゜**) TYPE-A(** θ**=35**゜**) TYPE-B(** θ**=35**゜**) TYPE-C(** θ**=35**゜**) Normal(**θ**=40**゜**) TYPE-A(** θ**=40**゜**) TYPE-B(** θ**=40**゜**) TYPE-C(** θ**=40**゜**) Normal(**θ**=45**゜**) TYPE-A(** θ**=45**゜**) TYPE-B(** θ**=45**゜**) TYPE-C(** θ**=45**゜**) Normal(**θ**=50**゜**) TYPE-A(** θ**=50**゜**) TYPE-B(** θ**=50**゜**) TYPE-C(** θ**=50**゜**)**
