|
|
Publications
An, G.S., Xiang, G., Xiang, X.B. and Guedes Soares, C. (2025), Physics informed neural networks based identification modelling of ship maneuvering motion and associated optimal excitation design, Engineering Applications of Computational Fluid Mechanics, Vol. 19(1), 2566860.
Accurate and stable identification modelling of ship maneuvering motion is the key to achieving high-precision ship maneuvering control. The hybrid knowledge- and data-driven Physics Informed Neural Networks (PINN) provides a more nonlinear, expressive, and interpretable approach to identification modelling. This paper proposes an innovative optimal excitation design method for the identification modelling framework of ship maneuvering motion using PINN. Firstly, the mechanism modelling and numerical simulation considering different levels of noise interference of the 3-DOF ship maneuvering motion of the KVLCC2 self-propelled ship model is performed based on the MMGmodel. Secondly, the comparative study indicates that the identification modelling using PINN exhibits marked enhancements in terms of both robustness and generalizability when compared to the pure data-driven ANN. Furthermore, a comprehensive investigation has been conducted to evaluate the identification modelling performance of deep PINN for the ship maneuvering motion prediction. The results demonstrate that the identification modelling using PINN can enhance the generalizability and prediction accuracy while significantly reducing the data size and quality requirements, and while enhancing the interpretability of the identification algorithm. An optimal excitation design method using the D-optimal criterion and Gray Wolf optimization algorithm is proposed, which is verified to improve the identification modelling accuracy and stability, as well as maneuvering prediction accuracy by Monte Carlo simulation.
If you did not manage to obtain a copy of this paper: Request a copy of this article
For information about all CENTEC publications you can download: Download the Complete List of CENTEC Publications
|