Using Machine Learning to Model Yacht Performance

Accurate modelling of the performance a yacht in varying environmental conditions can significantly improve yachts performance. However, racing is highly complex multi-physics system meaning that real-time prediction tools are always semi-empirical, leaving significant room for improvement. In this paper we first use unsupervised machine learning to analyse full-scale data. The widely documented ORC VPP (ORC, 2015) and commercial Windesign compared data across range wind conditions. then used train models. A number regression algorithms explored including Neural Networks, Random Forests Support Vector Machines improvements 82% obtained tools. physics- based models (Weymouth Yue, 2013) order reduce amount required achieve accurate predictions. It found outperform empirical even when predicting have not been supplied model as part training dataset.