Session: 12-03-01 Deterministic Wave and Motion Prediction

Paper Number: 80042

80042 - Real-Time Ship Motion Prediction Using Artificial Neural Network 

Ship motions play an important role in the safety of ship, cargo and crew onboard the ship. An advanced prediction of ship motions from perhaps a few seconds to a few minutes would be useful for tasks like helicopter landing, crew transfer, offshore operations, and deployment of ROV. In these scenarios predicting short periods of low ship motions would provide valuable information required to complete such motion-limited tasks. Such predictions would also serve as a navigational aid to ship masters manoeuvring in rough seas.

This paper aims to present a framework for predicting vessel motions in advance using artificial neural networks (ANN). With technological improvements (especially phase-resolved wave-sensing radar) it is possible to measure the wavefield around the ship. This wave field together with a time series of ship motions over the past few minutes is used to predict ship motions a few seconds to a few minutes in advance.

In this proof-of-concept study, we have used potential flow simulations which are validated using full-scale measurements. The wave field information was obtained from wave radar mounted on the ship. The wavefield observed using the wave radar was recreated in Wasim software, which is a time domain potential flow code, to simulate ship motions. Simulated ship motions demonstrated a good match with motion measurements from gyroscopes onboard the ship.

The validated Wasim model would be further used to perform multiple simulations using waves of different period, amplitude, and direction to develop a database of ship motions in different wave conditions. Wave field around the ship and corresponding ship motions would be used to train an ANN models which can predict future ship motions. The advantage of the ANN model is that once trained, the prediction requires limited computing power such as a typical laptop computer, which should be available onboard. The model would be tested using a variety of different wave conditions to validate the accuracy and robustness of the model.

Presenting Author: Bhushan Taskar TCOMS

Authors:

Bhushan Taskar TCOMS
Kie Hian Chua TCOMS
Tatsuya Akamatsu MTI Co.,Ltd. Singapore Branch
Ryo Kakuta MTI Co., Ltd.
Song Wen Yeow TCOMS
Ryosuke Niki MTI Co., Ltd.
Keita Nishizawa MTI Co., Ltd.
Allan Magee TCOMS