Uncertainty estimation of a CFD-methodology for the performance analysis of a collective and cyclic pitch propeller

摘要

Estimation and analysis of the uncertainty introduced by using a numerical model for the investigation and study of any type of flow problem have become common industry practice. Through understanding and evaluation of the uncertainty introduced by a numerical model, the accuracy and applicability of the model itself are evaluated. In this paper, the numerical uncertainty of a CFD-methodology developed to analyse the hydrodynamic performance of a collective and cyclic pitch propeller (CCPP) is estimated and analysed. The CCPP is a novel propulsion and manoeuvring concept for autonomous underwater vehicles, aimed to generate both propulsion and manoeuvring forces through advanced control of the propeller's blade pitch. The numerical uncertainty is established for three performance parameters, the generated propulsive force, the side-force magnitude, and the side-force orientation, by conducting a grid and time-step refinement study over three operational conditions. Additionally, the influence of the oscillatory uncertainty, introduced by the periodic nature of the problem, is investigated although shown to have a minimal effect when properly monitored. Based on a least-squares regression analysis of the refined simulation results, the numerical uncertainty is proven to be dominated by the introduced discretisation errors. In the case of the propulsive and side-force magnitude, the total uncertainty is dictated by the time discretisation uncertainty under bollard pull conditions, while the total uncertainty of the captive cases is mainly a result of the spatial discretisation uncertainty. The total uncertainty in the side-force orientation is observed to be primarily a consequence of the time discretisation uncertainty for all simulated cases. Overall, the total uncertainty for captive cases can be considered satisfactory for all three performance parameters, while further work is needed to reduce the observed uncertainty of the simulations under bollard pull conditions.