Performance Enhancement of an ACV in Varying Water Depth

This research focuses on the study of the behavior of a high speed vehicle and particularly an air-cushion vehicle (ACV) in varying bathymetry. An extensive data acquisition system is developed to gather data during the experiments. Four groups of experiments are conducted in a wave tank using a scale model surface effect ship to generate a database that is post processed to assess phenomena under various conditions. Group No1 experiments involved characterizing the wave motion in the tank in the absence of the vehicle as the waves transformed in response to variation in water depth. Based on these experimental datasets, the wave breaking type and position are predicted using a machine learning approach and, more specifically, a neural network of the multilayer perceptron type. Group No2 experiments are in support of a parametric study to evaluate the vehicle's performance under calm water conditions when the control inputs are varied. A system identific ation approach based on the experimental data is proposed to create a model that predicts the vehicles translational motion. In group No3 the experiments involve the vehicle travelling with a non-zero forward speed and encountering transforming head and following seas. Transient and non-linear phenomena and relations among parameters are observed Group No 4 experiments involve the vehicle maintaining a position in the "surf-zone" under manual control, encountering breaking waves that break on its bow skirt. Non-linear phenomena are discussed based on the experimental results.