In the past, the immediate method for valve manufacturers to de- tect cavitation is from the high noise level that is induced by cavi- tation and often accompanies serious cavitation。 A few published manuscripts investigate cavitation in a valve using this approach。 Rahmeyer (1982) performed experiments to measure  the cavita- tion parameter, which is used to judge the inception  of  cavitation in a butterfly valve。 They also measured the noise from cavitation and used 85 dbA as the cavitation limit。 Rahmeyer et al。 (1995) employed a closed loop flow facility where they utilized a globe valve with a multipath cage, which was stacked by many disks creating a torturous path with right angle turns。 Using the multipath cage in a globe valve can effectively reduce noise and vibration caused by cavitation。 In terms of the size, distribution, and collapse of vapors in their results, cavitation in the globe valve with the multipath cage is not as serious as the traditional globe valve。 Recently, Jazi and Rahimzadeh (2009) made  an attempt  to study the acoustic waveform induced by cavitation in a globe valve。 They employed the fast Fourier transform (FFT) technique to analyze the acoustic waveform。 It indicated that the cavitation in higher opening percentage has higher amplitude, frequency, and energy level from their results。

To reduce cavitation, there were different designs of valve cages to reach this purpose in a number of patents (Luthe and Hays 1978; Bates and Cain 1986)。 For example, Luthe and Hays (1978) proposed a special design of circumferential flow passageway  in a valve cage for a globe valve。 Because of the distribution of passageways, the fluids can  be  guided  and  form  jets  toward the center of the cage。 As a result, the area where cavitation incepts is restricted in the cage of the globe valve。 Bates and Cain  (1986)

JOURNAL OF ENERGY ENGINEERING © ASCE / MARCH 2013 /  25

revised the design of the passageway in the cage。 Passageways of the innovation of Luthe and Hays (1978) had sudden contraction, but Bates and Cain (1986) designed a torturous path in a passageway of the cage。 The valve cage proposed by Bates and Cain (1986) consists of many cylindrical sleeves。 Each sleeve contains the same amount of orifices to construct a torturous path。 The passageway design guides fluids to make a right-angle turn and controls the pressure drop effectively。 Bates and Cain (1986) design can also restrict the area of cavitation in the cage。 Nevertheless, the flow details in different passageways of those valve cages, which are intended to restrict cavitation, are not clear。 Hence, this study investigates the fluid flows in cages using numerical simulations。

Experiments are commonly used to analyze and test a valve, but there are some difficulties to perform those valve experiments。 For example, to observe the flow field in a valve, the valve body has to be transparent。 Nevertheless, it is difficult to fabricate a transparent valve because of its complex configuration。 Also, the transparent valve must be able to work under a high pressure environment and the materials must endure the high pressure in experiments。 Those problems raise the difficulty of the experiments for a valve。 Therefore, computational fluid dynamics (CFD) has become an alternative approach to simulate the flow field inside a valve。 Huang and Kim (1996) investigated three-dimensional and incompressible flow characteristics inside a butterfly valve using numerical simu- lations。 They described the influence of the valve disk angle on the flow field。 Kerh et al。 (1997) conducted a two-dimensional transient analysis for the interaction of a control valve under the alteration of a periodic inlet flow by the finite-element method。 Merati et al。 (2001) established numerical models and investigated flow fields in a V-sector ball valve。 Davis and Stewart (2002a, b)   undertook a CFD study and experiments for flows in five various openings of a globe valve。 They found good agreement between the CFD results and realistic experiments data in the variation of the flow coefficient。 Chern and Wang (2004) studied fluid flows inside a ball valve and calculated performance coefficients including the flow coefficient, loss coefficient, and cavitation index using the finite-volume method。 Cavitation phenomena inside the ball valve were reported in their experimental results。 They also utilized a V-port, which is installed behind the ball valve to control the fluid flow。 The influences of the V-port on the performance coefficients were presented by Chern and Wang (2004)。 Moujaes and Jagan (2008) demonstrated the three-dimensional CFD prediction of turbulent flow in a ball valve。 They predicted the characterized coefficients and made a comparison with ASHRAE experimental data in different partial openings。