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

Submission Number: 157540

A Graph Neural Network Based Surrogate Model for Flapping-Driven Ocean Energy Harvesting 

ABSTRACT

Cantilevered elastic foils or panels can undergo self-induced large amplitude flapping oscillations when submerged in a flowing fluid.  This canonical fluid-structure interaction phenomenon, commonly observed in nature, such as in fluttering leaves or fish fins, offers a promising mechanism for ocean energy harvesting. By leveraging the repetitive flapping motion, elastic foils can harness the abundant kinetic energy present in marine environments, including steady ocean currents and unsteady wave-driven flows.

The efficiency of energy harvested by these foils is dictated by the oscillation amplitude and frequency, which are influenced by interactions with the surrounding flow. Optimizing these parameters is crucial for maximizing energy conversion efficiency and designing robust systems. However, high-fidelity simulations required for optimization over a wide parameter space are computationally expensive. To address this challenge, we propose a graph neural network-based surrogate model (GNN-ROM) tailored for the inverted foil problem. This model effectively handles mesh-structured data from full-order simulations without modifications, enabling efficient and accurate fluid-structure interaction simulations.

In this study we consider a two-dimensional setup of an elastically mounted rigid foil undergoing single degree of freedom pitching motion about its trailing edge in uniform flow. The coupled fluid-structure system is simulated using a high-fidelity Petrov-Galerkin finite element approach, leveraging the arbitrary Lagrangian-Eulerian (ALE) formulation for precise interface tracking [1]. The proposed GNN framework utilizes these high-fidelity simulations as ground truth data for training and evaluation. Our deep learning model uses the rotation equivariant, quasi-monolithic GNN architecture proposed by [2]. This framework is based on the ALE formulation, wherein time series prediction of the coupled system state is made with two sub-networks. Essential coefficients describing mesh motion are extracted by proper orthogonal decomposition, which are predicted over time using a single multilayer perceptron. Simultaneously, the GNN-ROM evolves the flow field based on the system state. The structural state is implicitly modeled by the mesh movement on the solid-fluid interface. This approach effectively simulates the coupled FSI dynamics, offering a robust surrogate model for rapid parametric optimization and control of energy-harvesting devices based on inverted foil configurations.

REFERENCES

[1] Jaiman, R. K., Pillalamarri, N. R., & Guan, M. Z. (2016). A stable second-order partitioned iterative scheme for freely vibrating low-mass bluff bodies in a uniform flow. Computer Methods in Applied Mechanics and Engineering, 301, 187-215

[2] R. Gao and R. K. Jaiman, "Predicting fluid – structure interaction with graph neural networks," Physics of Fluids, 2023.

Presenting Author: Aarshana Parekh The University of British Columbia, Vancouver

Presenting Author Biography: I am a fourth-year PhD Candidate in Mechanical Engineering at the University of British Columbia, working under the supervision of Prof. Rajeev Jaiman. As a renewable energy enthusiast, I am deeply passionate about fluid-structure interactions, computational modeling, and designing innovative energy-harvesting systems. My research focuses on leveraging high-fidelity simulations and advanced machine learning techniques to optimize flow-induced vibrations in flexible structures for real-world renewable energy applications. Committed to advancing sustainable technologies, I aim to contribute to the transition toward clean and efficient energy solutions