Condition: Reynolds number = 1.0 × 10 7 with double model assumption.
Commentary
Simulation of viscous flow with a high Reynolds number is time-consuming, taking several days per case, and hinders interactive and multidisciplinary design. Developing a surrogate model of viscous flow simulation in the entire design space of the ship is one of the promising ways to realize the interactive and multidisciplinary design of the ship.
We provide fast flow simulation around the ship's hull surface, making a few hours of simulation fast and interactive. We presume a double model in the current model to simplify the code and dataset, but simulation conditions such as Reynolds number, Froude number, and incident waves can be considered by setting those in the machine learning model input.
Reference
- Ichinose, Y., Gaspar, Henrique M., 2023. Interactive Ship Flow Simulation Enhanced By Neural Network Model In A Web Environment, ECMS,DOI:https://doi.org/10.7148/2023-0155
- Ichinose, Y., Taniguchi, T., 2022. A curved surface representation method for convolutional neural network of wake field prediction. J Mar Sci Technol 27, pp.637 647.
- Ichinose, Y., 2022. Method involving shape-morphing of multiple hull forms aimed at organizing and visualizing the propulsive performance of optimal ship designs. Ocean Engineering 263, 112355.
- MOERI KVLCC2