fixed operation parameters, variation of the phase between the sources 

 provides the additional sets of relations needed to find [z]. In practice 

 combinations of phase and amplitude were used. The excitation consisted 

 of a "siren" like valve which provided a variable hydraulic resistance to 

 the flow circuit. 



Another notable feature of the Ng experiments was the use of laser 

 velocimetry to measure up and downstream flows. This necessitated certain 

 flow smoothing and conditioning not common in hydraulic circuits. These 

 experimental features are shown in Figure 40, and an example of some of the 

 first experimental results are shown in Figure 41 (Ng et al. 1976). 



This rather involved figure shows two sets of data, one basically 

 noncavitating and the other cavitating. The noncavitating results show 

 that the terms z , z , and z„ are small; in the absence of cavitation 

 and elasticity of the fluid and supporting structure, and experimental 

 error, these terms should be strictly zero. The term z „ is seen to have 

 a negative real part. This corresponds fairly well to the slope of the 

 total head performance curve, the angle 3 of Figure 38a, but shows some 

 change with frequency. The term z also has an imaginary part which is 

 equivalent (within experimental accuracy) to the inertia of a certain 

 length of fluid. The fully wetted data are rather as expected. 



The cavitating transfer function is, however, quite different. As 

 our touchstone, the slope of the performance curve decreases (in magnitude) 

 at higher frequency and the inertance part practically vanishes. We may 

 mention that the cavitation number, O = 0.046, results in a condition of 

 extensive cavitation in the inducer portion of the pump. There is attached 

 blade cavitation, but the extensive cavitation seen under strobe illumi- 

 nation is due primarily to bubbly-cavitating tip-clearance flows (see 

 Figure 41) . This extent of cavitation is a normal condition of operation 

 for many inducer pumps. We see some change in the customary pressure 

 rise-flow characteristic (z 19 ) due to cavitation. But there are more 

 important and striking changes in the other terms of the transfer function; 

 the term z.. , for example, exhibits a large change with frequency, both in 



85 



