Session: 08-05-01 Deep Learning for Predictive Modeling in Marine and Offshore Systems

Submission Number: 156540

Efficient Prediction of Mooring Forces of a Floating Offshore Wind Turbine Using LSTM-Based Deep Learning 

The relationship between motion and mooring forces in floating offshore structures, such as floating offshore wind turbines (FOWTs), is inherently nonlinear and complex. Traditional physical experimental and computational methods used to determine this relationship are often time-consuming and costly, particularly in the context of space exploration and large-scale simulations. To address these challenges, we propose an efficient approach using Long Short-Term Memory (LSTM) networks, a type of deep learning model, to predict the unsteady mooring forces on FOWTs. In our method, we employ techniques like dropout, leaky linear rectification, and Swish activation layers to effectively approximate the mapping between motion time series and mooring force time series. The model simulates FOWT behavior subjected to irregular waves, generated based on the JONSWAP spectrum, which induces surge motion. This motion is then input into the deep neural network for training. The LSTM model is trained using a stochastic gradient descent approach to forecast the time series of mooring forces. When compared to traditional simulation data, our deep learning model delivers predictions that exceed the accuracy of full-order simulations by nearly two orders of magnitude, all while requiring significantly fewer computational resources. Additionally, we introduce a novel method for achieving real-time predictions without modifying the LSTM network itself. This LSTM-based approximation technique shows great promise for parametric design and digital twinning applications in the FOWT industry.

Presenting Author: Tharindu Miyanawala Technology Center for Offshore and Marine, Singapore

Presenting Author Biography: Dr. Tharindu Miyanawala is a Scientist at TCOMS Singapore. His research is focused on building digital twins for offshore assets. Dr. Miyanawala obtained his PhD from the National University of Singapore in 2019.