Challenges of numerical simulations of cavitation reactors for water treatment - An example of flow simulation inside a cavitating microchannel

Abstract

The research on the potential of cavitation exploitation is currently an extremely interesting topic. To reduce the costs and time of the cavitation reactor optimization, nowadays, experimental optimization is supplemented and even replaced using computational fluid dynamics (CFD). This is a very inviting opportunity for many developers, yet we find that all too often researchers with non-engineering background treat this "new" tool too simplistic, what leads to many misinterpretations and consequent poor engineering. The present paper serves as an example of how complex the flow features, even in the very simplest geometry, can be, and how much effort needs to be put into details of numerical simulation to set a good starting point for further optimization of cavitation reactors. Finally, it provides guidelines for the researchers, who are not experts in computational fluid dynamics, to obtain reliable and repeatable results of cavitation simulations.

Keywords: Cavitation; Computational fluid dynamics; Numerical simulation; Venturi.

Fig. 1

Experimental setup (top) and the Geometry of the Venturi microchannel (bottom).

Fig. 2

General observations of the phenomena, which are unique for developed cavitation in microchannels .

Fig. 3

Microchannel computational domain geometry and detail of the mesh in the throat region of the section.

Fig. 4

Measured and predicted pressure losses as a function of mass flow rate.

Fig. 5

Cavitation cloud shedding (Experiment: $\dot{\mathrm{m}}$ = 9.15 g/s, Δp = 4.00 bar, σ = 1.24, Simulation: $\dot{\mathrm{m}}$ = 9.03 g/s, Δp = 3.71 bar, σ = 1.26). The time difference between the images is Δt = 0.5 ms.

Fig. 6

The length of the attached cavity as a function of the cavitation number.

Fig. 7

Cavitation cloud shedding frequency as a function of the cavitation number.

Fig. 8

Strouhal number as a function of the cavitation number.

Fig. 9

Cavitation cloud separation and the formation of Kelvin-Helmholtz instability (Experiment: $\dot{\mathrm{m}}$ = 9.15 g/s, Δp = 4.00 bar, σ = 1.24, Simulation: $\dot{\mathrm{m}}$ = 9.03 g/s, Δp = 3.71 bar, σ = 1.26). The time difference between the images is Δt = 40 µs.