Assessment of Experimental, Computational, and Combined EFD/CFD Methods for Ship Performance Prediction

Sammanfattning: In today’s highly competitive market, alongside increasingly stringent regulatory requirements, the precise prediction of ship performance has assumed paramount importance for both design verification and operational evaluations. This thesis addresses the need for a comprehensive assessment of Experimental Fluid Dynamics (EFD), Computational Fluid Dynamics (CFD), and their combination to enhance the accuracy of performance predictions. Moreover, it explores the potential of combined EFD/CFD methods in improving power predictions by either replacing or complementing certain aspects of the existing methodology, while also introducing novel methods. The investigation identifies the Prohaska method as a prominent source of uncertainty in the ITTC-78 method. As an alternative, the CFD-based form factor method is meticulously examined, employing various codes and numerical approaches. The findings robustly establish the applicability and accuracy of the CFD-based form factor method, even when subjected to diverse numerical approaches and computational grids. Furthermore, best practice guidelines are derived for double-body RANS computations, ensuring compatibility with experimental form factors. Another debated issue within the ITTC-78 method is the very concept of form factor. This study conclusively affirms the Reynolds number dependence in form factors when the ITTC-57 line is employed. However, the numerical friction lines derived in this research, effectively eliminates these scale effects. Additionally, this study addresses conditions with flow separation, which renders the conventional form factor approach inadequate. A two form factor method (2−k method) is proposed to address instances of separated flow, complemented by an empirical correction formula for vessels with deep transom submergence and wetted transom flow. Furthermore, this thesis delves into the exploration of direct full-scale CFD computations for ship performance prediction. Extensive validation studies, encompassing numerous test cases and sea trials, are conducted to compare the accuracy of full-scale CFD computations with EFD based, and combined EFD/CFD methods. This thesis quantifies, for the first time in the literature, the difference in accuracy between fully computational and extrapolation-based methods using a large number of test cases and sea trials. The results indicate that while the prediction accuracy of full-scale CFD computations for power and RPM is lower than the other methods, the discrepancy is not substantial. Conversely, the investigations underscore that the combined EFD/CFD methods stand as the most accurate prediction method. Consequently, this thesis recommends incorporating combined EFD/CFD methods into the recommended procedures, as it offers immediate improvements to the existing ship performance prediction methods.