Scholarlink

<p>This study applies artificial neural networks (ANNs) to predict the velocity made good (VMG) of the America’s cup class version 5 (ACCV5) sailboat model. Geometric and rig dimensions of the ACCV5 are analyzed using Maxsurf VPP software to generate velocity polar plots and VMG data under varying true wind conditions. Validation against towing tank experiments showed a maximum error margin of 6%, demonstrating the software’s reliability. An ANN model is developed with 18,000 samples, generated by systematically varying sail dimensions from 25 to 175% of their original size. Using 21 input features, the model achieved high accuracy with an architecture comprising two hidden layers and 84 neurons, trained over 1000 epochs. The ANN successfully generalized VMG predictions across diverse scenarios, facilitating the derivation of a predictive equation. To enable practical use, the equation is integrated into a graphical user interface, named the ANN-VMG prediction tool. Finally, the results underscore the ANN-VMG prediction tool’s effectiveness as a valuable resource for optimizing sailing performance across a wide range of condition and offer a solid foundation for future research and practical applications in the field of marine engineering, specifically in the optimization of sailboat design and performance using AI-based models.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicting velocity made good for the ACCV5 sailboat using artificial neural networks

  • 期刊论文

  • 发表日期：

    2025年5月5日

查看全文
  • Hussien M. Hassan,
  • M. M. Moustafa

摘要

This study applies artificial neural networks (ANNs) to predict the velocity made good (VMG) of the America’s cup class version 5 (ACCV5) sailboat model. Geometric and rig dimensions of the ACCV5 are analyzed using Maxsurf VPP software to generate velocity polar plots and VMG data under varying true wind conditions. Validation against towing tank experiments showed a maximum error margin of 6%, demonstrating the software’s reliability. An ANN model is developed with 18,000 samples, generated by systematically varying sail dimensions from 25 to 175% of their original size. Using 21 input features, the model achieved high accuracy with an architecture comprising two hidden layers and 84 neurons, trained over 1000 epochs. The ANN successfully generalized VMG predictions across diverse scenarios, facilitating the derivation of a predictive equation. To enable practical use, the equation is integrated into a graphical user interface, named the ANN-VMG prediction tool. Finally, the results underscore the ANN-VMG prediction tool’s effectiveness as a valuable resource for optimizing sailing performance across a wide range of condition and offer a solid foundation for future research and practical applications in the field of marine engineering, specifically in the optimization of sailboat design and performance using AI-based models.