The Application of Artificial Neural Network - MISO (Multi Input Single Output) in Catamaran Ship Roll Motion Prediction

Abstract Artificial intelligence (AI) technology has become increasingly prevalent in various fields within the shipbuilding and marine industries. Typical applications of AI include production management, voyage analysis, and motion prediction. The main objective of this study is to develop a ship roll motion prediction tool for operational conditions that is fast, accurate, and computationally efficient. To achieve this, an AI algorithm based on Artificial Neural Networks (ANNs) with a Multiple Input Single Output (MISO) architecture was developed. The algorithm requires input parameters such as ship speed, wave height, wave period, and heading. The methodology for developing the predictive algorithm is presented, along with the ship motion data used for training and validating the ANN model. The ship motion data were generated using a frequency domain analysis based on the Boundary Element Method (BEM). The AI predictions were compared against 20% of the hydrodynamic analysis results to validate the model’s performance. The results demonstrate that the developed AI tool effectively predicts the roll motion of a catamaran ship, capturing the non-linear behavior of ship dynamics under various operational conditions.