**3.3 Experimental study and numerical analysis of cavitating flows in three inducers 3.3.1 Experimental tests**

Fig. 13. Bank of experimental tests from DynFluid laboratory (Arts et Métiers ParisTech), loop of the industrial inducers

The hydrodynamic bank used for the experimental tests consists of three independent closed loops. The first loop is adapted for test on the industrial inducers, see Fig. 13. The second loop is used for experimental tests of centrifugal pumps alone or coupled to an inducer. The third loop is adapted for aircraft-type inducers.

Numerical and Experimental Study of Mass Transfer Through Cavitation in Turbomachinery 197



The formation of various type of cavitation is caused by the kinematics of the flow in the inducer. The numerical and experimental analyses suggest that the high velocity regions (favourable zone to cavitation inception) are localized to the inlet at partial flow rates, and

Fig. 15. Vapour behaviour for various cavitation conditions and its corresponding performance

The numerical simulations, in steady and unsteady regime, were carried out on a two-blade aircraft inducer with a blade tip angle of 4°. In order of accelerate the calculation time, for the steady numerical simulations, only one third of the inducer was modelled. By contrast, this simplification could not be considered for the unsteady calculations because the instabilities of the cavitating flow are influencing by the neighbour blades (system

β*=8°* (Φ*=0.057*)

drop curve obtained experimentally on a three-blade inducer with

the main flow, see Fig. 14(e-f).

inducer, see Fig. 14(g-h).

toward the outlet at high flow rates.

**3.3.2 Numerical analysis** 

instabilities).

**3.3.2.1 Domain de control and grid generation** 

Fig. 14. Diverse cavitation forms obtained experimentally on a three-blade inducer with β*=16°* (Φ*=0.164*)

The industrial inducers loop consists mainly of the follow elements:


To capture the experimental structures of vapour through the transparent cover, it was used a digital camera under strobe lighting.

Fig. 14 shows the vapour formation at leading edge of a three blade inducer. The regions of vapour were mainly manifested in the form of three identical regions attached to each blade (σ*=0.280*). As the inlet pressure decreases, the cavitating structures suffer a growth phase, principally at tip leading edge of the blade (from σ*=0.230* to σ*=0.090*), which move to down to the hub until they block the flow channel (σ*=0.060* and σ*=0.055*). When the passage interblade is blocked by the vapour, the performance drop of the inducer occurs suddenly (σ*=0.050* and σ*=0.045*). The gradual vapour apparition generates noise and vibrations. In this figure, each image corresponds to a value of σ for a flow rate constant.

The pictures obtained during the cavitation tests are typical cavitation forms and, in general, they are consistent with those reported by (Offtinger et al., 1996), see Fig. 14. Representative pictures were obtained from different experimental tests on the three studied inducers at partial flow rate, nominal flow rate and over-flow rate. As an example, some pictures obtained on a three-blade industrial inducer with blade tip angle of 16° are commented:


The formation of various type of cavitation is caused by the kinematics of the flow in the inducer. The numerical and experimental analyses suggest that the high velocity regions (favourable zone to cavitation inception) are localized to the inlet at partial flow rates, and toward the outlet at high flow rates.

Fig. 15. Vapour behaviour for various cavitation conditions and its corresponding performance drop curve obtained experimentally on a three-blade inducer with β*=8°* (Φ*=0.057*)
