When ships sailing at the ocean, six degrees of freedom of ship motions occur due to the sea waves. The wave-induced ship motions always bring adverse effects on carrying out maritime operations. Accurate real-time estimation of the deterministic ship motions is important for both ship motions prediction and assistant decisions making. Conventionally, the deterministic ship motions are calculated by using the ship motions RAOs and input wave spectrum. Where, the input spectrum is obtained by applying time-frequency transformation on the wave-excitation time sequential inputs. The conventional approach is linear. It fails to model the sea waves and its induced ship motions at severs states due to the nonlinearity. In this paper, a real-time estimation method for deterministic ship motions based on Long-Short-Term-Memory (LSTM) deep learning model was proposed. Numerical simulations are carried out for both validation and evaluation purposes. The nonlinear ship motions are simulated by a fully nonlinear computer program. The LSTM model is trained based on the simulated nonlinear ship motions data sets and the corresponding input wave data sets. The training process establishes the relationship between near-field wave-excitation input and ship motions. The trained LSTM model is applied to estimate the ship motions based on the wave-excitation inputs. Preliminary results show that the proposed method performs effectively in the real-time calculation of deterministic ship motions.
展开▼
