Numerical uncertainty due to parallelization in an unsteady cloud cavitation

Cavitation is the formation of vapor bubbles in a liquid when the pressure drops below the vapor pressure of the liquid. This can happen when a liquid is accelerated to high speeds, such as in the impeller of a pump or the propeller of a boat. Experimental studies of cavitation can provide valuable insights into the flow dynamics, but they can be expensive and time-consuming. Numerical simulations are a more economical way to study cavitation, and they can be used to obtain preliminary designs for devices that are susceptible to cavitation damage. In this study, cloud cavitation was studied in a venturi geometry using the incompressible solver from OpenFOAM. Due to the presence of different bubble sizes and time scales, the homogeneous transport equation model was used to reduce the computational effort. The Merkle model was used to model the phase transition due to cavitation. The flow is very complex in the presence of cloud cavitation, and there are many uncertainties involved in the calculations. One source of uncertainty is unavoidable due to parallelization, which is needed to decompose the domain to reduce simulation time. It was observed that the shedding behavior varies when different numbers of processors are used. This study shows that the uncertainty coming from parallelization can be reduced by increasing the sampling rate.