EAGER: CET: Investigating Orbital Motion Dynamics in Offshore Wind Turbines During Installation: A Novel Digital Twin Approach Based on Generative Artificial Intelligence (AI)

University Of Maine, Orono ME

Investigators

Abstract

This project is jointly funded by the Established Program to Stimulate Competitive Research (EPSCoR), and the NSF-wide Clean Energy Initiative supporting DCL 23-109. This EArly-concept Grants for Exploratory Research (EAGER) award is made in response to Dear Colleague Letter 23-109. The US plans to deploy 30 GW of offshore wind energy by 2030, increasing to 110 GW by 2050. Such deployment is projected to power 10 million homes, create 77,000 new jobs, and reduce carbon emissions by 78 million metric tons. However, significant knowledge gaps exist in understanding the unique dynamics of offshore wind turbines (OWTs) during their installation phases, posing major challenges to their safe and cost-efficient deployment. This high-risk, high-reward EAGER project, led by an interdisciplinary team from the University of Maine, will address a critical challenge related to installing OWT blades, which impedes the adoption and development of offshore wind energy. The team will conduct original research on the orbital motion dynamics in the nacelle of a partially installed bottom-fixed OWT configuration through wave tank experiments and numerical modeling. They will then use a digital twin framework based on generative artificial intelligence to predict the orbital motion dynamics in real-time. The project also aims to advance offshore wind education and workforce training through activities such as integrating new knowledge into existing undergraduate and graduate courses at the university, expanding collaboration with a local high school’s diverse STEM academy by offering K-12 research internships to students from underrepresented communities; preparing an open-source graduate textbook on OWT installation, and mentoring graduate students in partnership with the National Renewable Energy Lab (NREL), Equinor - a global leader in offshore wind farm development, and Kartorium - an Alaska-based digital twin company. This project explores the root causes of the intricate oscillation orbits in OWTs during the blade installation phase. The current lack of understanding in this area including the absence of a predictive tool compromises the safe and cost-efficient deployment of OWT. The interdisciplinary team from the University of Maine will examine the orbital motion dynamics in the nacelle of a bottom-fixed, monopile-mounted OWT in its hammerhead configuration and identify correlations with variables such as varying sea states, wave headings, and hammerhead configurations. A Froude-scaled model experiment in a wave tank will be performed, developing and employing a groundbreaking methodology for designing a flexible scaled model of an OWT in hammerhead configurations, suitable for wave tank testing. The experiments will be complemented by a high-fidelity numerical model for analyzing these orbital motions and trajectories. The project will use the results from the model scale experiments and the high-fidelity numerical simulations to design, verify, and evaluate a digital twin of the dynamic system based on generative AI algorithms based on a novel physics-informed latent diffusion model to make real-time predictions of orbital motions. Here, domain-specific physics-based knowledge will be integrated into the architecture and training process of the modern diffusion model, enhancing the AI model’s learning capabilities, interpretability, and generalizability. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.