Author | Reverberi, A., Lloyd, T. and Vaz, G. |

Title | Towards cavitation modelling accounting for transition effects |

Conference/Journal | 19th Numerical Towing Tank Symposium (NuTTS), St. Pierre d'OlĂ©ron, France |

Month | October |

Year | 2016 |

Pages | 120-125 |

Abstract

Cavitation is an important phenomenon which should be taken into account when designing propellers in order to minimise noise radiation, vibration and erosion. The occurence of cavitation is however highly sensitive to a number of factors which influence its inception; these include water quality, surface roughness and transition from laminar to turbulent flow. Since cavitation observations of propeller designs are typically made at model scale, it is important to understand and control for differences in the flow conditions between model and full scale (Arndt, 1981). At MARIN, work has previously been carried out to examine cavitation inception (Kuiper, 1981) and the effects of viscosity on cavitation (Van Oossanen, 1974) for marine propellers. One method for reducing the scale effect on cavitation inception is to apply roughness, thereby also ensuring a fully turbulent boundary layer which more closely resembles that seen at full scale. However, in cavitation tests where roughness is not applied, the observed cavitation patterns are highly dependent on transition to turbulence (Figure 1 shows typical model scale flow regimes). In this case, cavitation inception is driven by intense pressure fluctuations resulting from laminar separation or transition to turbulence (Arndt, 1981).

From a modelling perspective it therefore makes sense to account for transition in the prediction of cavitation. Typical two-equation RANS models do not predict transition correctly, although this behaviour can be adjusted by combining them with a transition model (Eca et al., 2016). Most commonly used homogeneous mixture-based cavitation models only use a simplified inception criterion, that is p < pv, where p and pv are the local and vapour pressures, which does not account for the effects of turbulence. Figure 2 shows an example of overprediction of cavitation in a region of the blade where laminar flow may be expected (Vaz et al., 2015). A parameter representative of the turbulence level can be included, such as the turbulence kinetic energy (Singhal et al., 2002), yet this only serves to increase vapour production. Asnaghi (2015) modified the model of Sauer and Schnerr (2001) to account for the effect of strain rate in the cavitation inception criterion and model vapourisation rate. In this paper, we present a preliminary investigation into how application of a transition model influences the flow prediction for a cavitating marine propeller (without modifying the cavitation model), as well as identifying potential areas for further model development.

Cavitation is an important phenomenon which should be taken into account when designing propellers in order to minimise noise radiation, vibration and erosion. The occurence of cavitation is however highly sensitive to a number of factors which influence its inception; these include water quality, surface roughness and transition from laminar to turbulent flow. Since cavitation observations of propeller designs are typically made at model scale, it is important to understand and control for differences in the flow conditions between model and full scale (Arndt, 1981). At MARIN, work has previously been carried out to examine cavitation inception (Kuiper, 1981) and the effects of viscosity on cavitation (Van Oossanen, 1974) for marine propellers. One method for reducing the scale effect on cavitation inception is to apply roughness, thereby also ensuring a fully turbulent boundary layer which more closely resembles that seen at full scale. However, in cavitation tests where roughness is not applied, the observed cavitation patterns are highly dependent on transition to turbulence (Figure 1 shows typical model scale flow regimes). In this case, cavitation inception is driven by intense pressure fluctuations resulting from laminar separation or transition to turbulence (Arndt, 1981).

From a modelling perspective it therefore makes sense to account for transition in the prediction of cavitation. Typical two-equation RANS models do not predict transition correctly, although this behaviour can be adjusted by combining them with a transition model (Eca et al., 2016). Most commonly used homogeneous mixture-based cavitation models only use a simplified inception criterion, that is p < pv, where p and pv are the local and vapour pressures, which does not account for the effects of turbulence. Figure 2 shows an example of overprediction of cavitation in a region of the blade where laminar flow may be expected (Vaz et al., 2015). A parameter representative of the turbulence level can be included, such as the turbulence kinetic energy (Singhal et al., 2002), yet this only serves to increase vapour production. Asnaghi (2015) modified the model of Sauer and Schnerr (2001) to account for the effect of strain rate in the cavitation inception criterion and model vapourisation rate. In this paper, we present a preliminary investigation into how application of a transition model influences the flow prediction for a cavitating marine propeller (without modifying the cavitation model), as well as identifying potential areas for further model development.