Aplicación de herramientas de Machine Learning al comportamiento en la mar

Abstract

The Artificial Intelligence (AI) has been developing in the last decades, focused on Smart writing, voice recognition systems, autonomous systems, and others. The application of AI in Naval Architecture field might become a relevant tool, especially those applied to assess the expected degree in compliance of a design. This work focuses on the application of Machine Learning to a specific field of Naval Architecture, the seakeeping. The complexity of ship-wave interaction makes common using mathematic simplifications, such as the platforms response is linear. And the Finite Element Methods or Boundary Element Methods are common numerical methodologies used to solve this type of problems. It is proposed a learning methodology based on an Artificial Neural Network (ANN) to evaluate the seakeeping performance parameters. And this methodology offers a time saving alternative when compared with other solutions. It is first introduced the state of art of IA applied to Naval Architecture and Marine Engineering. Then, it is explained the methodology and network training. Later, it is performed comparisons with different types of ships. And finally, it is shown relevant conclusions about the results obtained.