Numerical Prediction of Thruster-Thruster Interaction
Many vessels deploying offshore activities nowadays are dynamically positioned by multiple azimuth thrusters instead of anchors. The multiple propulsor set up, gives a considerable flexibility to work fast and accurate. Due to the fact that the thrusters are positioned relative close to one another their performance is influenced. Normally to quantify this influence and take into account in the DP control algorithm, elaborate experiments have to be performed. To optimize the results a robust numerical flow solver is developed to predict the interaction effects. The program is used to optimize the effort put into these experiments. The developed propeller interaction model is a first order potential based panel method, which uses zero order doublets and sources panel elements. This method is selected to prove the main objective of this research that; Although the slipstream of a thruster has a very turbulent character the interaction can be modeled without taking the viscosity into account as long as an accurate distorted flow field behind a propeller can be predicted. At the 2nd thruster the distorted flow field due to the 1st thruster is modeled by means of two wake field models; a linear potential wake model and an empirical turbulent jet model. Due to the intersection of wake and body panels at the 2nd thruster, numerical instabilities occur at the collocation points. These instabilities are removed by applying a realistic vortex model instead of the analytic vortex model which has infinite velocities in the core. The second problem is to capture the divergent and subsiding character of a propeller wake field by means of a linear potential wake model. This problem is resolved by validating the region for which the results are still accurate. From the results it is concluded that the thruster interaction propeller model coupled to the turbulent jet wake field yield accurate thruster interaction results. For the linear potential wake field results are promising but adaptations are needed to improve the prediction of the divergent and subsiding character of the physical wake field.