Implementation of Physically Based Wake Simulation for Planning and Semi-Planning Yachts using Niagara VFX

Abstract

Amid the ongoing digital transformation, the demand for dynamic and interactive virtual yacht tours in the online marketplace is steadily increasing. A ship wake simulator serves as a core component of such applications. Rather than employing advanced numerical methods to compute wave resistance, this study concentrates on the real-time, physically based visualization of ship wake patterns. The final output will be an application with a user interface developed in Unreal Engine 5 (UE5). The selected theoretical models must be programmable and readily implementable in the chosen game engine. For far-field wake simulations, semi-analytical integrals are employed to reproduce Kelvin wake patterns. Michell’s thin-ship theory is applied to model ship waves along a straight course with explicit hull geometry, while Stoker’s influence-point model is adopted to handle arbitrary path curvatures. For near-field wake simulations, empirical equations are used to construct wake profiles through interpolation, and the Savitsky model is implemented to capture planing effects (e.g., rooster tails) at high speeds. Since both the theoretical models and the game engine impose inherent limitations, the raw results require post-processing and normalization to ensure visual consistency. For instance, normalization prevents excessive brightness in wake patterns when computed values exceed the [0, 1] interval. The processed wake data are then mapped onto texture layers to achieve the intended visualization. Benchmarking further demonstrates that the GPU-accelerated game engine achieves at least an order of magnitude higher performance than a Python implementation (e.g., Jupyter Notebook) in high-resolution simulations. Results for the base case scenario show good agreement with the classical Kelvin wake patterns. As the Froude number increases, the wake transitions from a transverse-wave-dominated to a divergent-wave-dominated pattern. A narrowing of the apparent wake angle is also observed. Results for more advanced cases (arbitrary path curvature, acceleration, interference, planing effect) also present the similar patterns from the other studies. Nevertheless, further investigation is needed to develop improved approaches that mitigate the singular behavior occurring at the wake arms. It is concluded that the developed program successfully achieves the primary objective of this study, namely, real-time rendering of ship wake patterns with preserved physical fidelity. Nonetheless, minor visual artifacts remain, and future work should incorporate more rigorous quantitative validation.