High Speed Craft Peak Acceleration Prediction using Machine Learning Models

The ability to predict the peak impact of slams, which occur during high-speed motion of planing watercraft, could lead to significant opportunities for improving the safety of passengers and equipment aboard. We apply Random Forest and Convolutional Neural Networks to achieve this goal using peak acceleration, freefall duration, and velocity change from high-speed crafts to quantify the amplitude of wave impact. We analyze different characteristics separately in the freefall and impact regions for each wave. We report numerical results of both models, demonstrating the accuracy of these predictions. Keywords: high-speed craft, machine learning, random forest, convolutional, neural networks