Session: 01-01-01 Offshore Platforms-I

Paper Number: 131710

131710 - Efficiency and Innovation in Offshore Renewable Energy: A Surrogate Modeling Approach to Mooring System Optimization 

As the offshore renewable energy industry progresses, it has become significantly important to ensure that designs can effectively generate the necessary energy, endure the demanding environmental conditions throughout their operational life, and maintain cost-efficiency. To tackle these conflicting priorities, design processes now incorporate optimization techniques. These methodologies not only facilitate the discovery of innovative design ideas but also aid system designers in making well-informed strategic choices.

With a rising number of offshore renewable energy devices adopting floating solutions, mooring systems have become a critical subsystem impacting both device survivability and costs (Thomsen et al. 2018). However, due to the significant computational time required for mooring system simulations, the use of optimization algorithms in the design cycle is not yet widespread. Without numerical optimization techniques, mooring system design relies on an iterative engineering approach based on experience and engineering judgment. This often results in innovative mooring designs being overlooked and suboptimal mooring solutions being deployed (Johanning, Smith, and Wolfram 2006).

To implement optimization methods effectively in complex engineering design problems, surrogate modelling has emerged as a vital technique. It involves the use of simplified, low-fidelity models that approximate high-fidelity results at a reduced computational cost (Won and Ray 2005; Voutchkov and Keane 2006; Jin 2011).

In mooring system design, it is common practice to prioritize designs that minimize costs or excursions while adhering to constraints related to line tension. Given the intricacies of these design considerations, adopting an optimization approach, particularly multi-objective optimization, becomes essential. This approach helps in quantifying the trade-offs among competing design objectives and enhances the decision-making process.

This study introduces a comprehensive framework for the multi-objective optimization of mooring line systems in the context of offshore renewable energy applications. By integrating a ridge regressor-based surrogate model with the Non-dominated Sorting Genetic Algorithm II (NSGA-II), our approach offers a novel and effective means of addressing the complex challenges associated with the design and operation of mooring systems. This framework is exemplified through the optimization of mooring lines specifically tailored for a wave energy converter, showcasing its versatility and applicability to real-world problems encountered in the offshore renewable energy sector.

The optimization framework begins with the utilization of validated numerical models of the mooring system. These models serve as the basis for training a ridge regressor-based surrogate model, which, in turn, facilitates a computationally efficient optimization routine. The genetic algorithm is employed to systematically navigate the search space, allowing for a thorough exploration of design possibilities. This integration of advanced computational techniques ensures the reliability and effectiveness of the optimization process.

The primary objective is to optimize the design of a mooring line, considering the dual goals of minimizing manufacturing costs and maximizing the x-direction offset motion of the platform while ensuring adherence to all specified constraints throughout the optimization process.

The surrogate model is trained on a comprehensive dataset of simulations, encompassing various decision variables employed in the optimization process as input features. The output targets for the model comprise the tension and x-offset motion of the platform. Consequently, the surrogate model can predict targets for any given set of decision variable values.

During the optimization process, the surrogate model serves to validate the population in each generation. This obviates the necessity to execute simulations in OrcaFlex, which is a widely recognized software tool for analysing offshore structures in real-time scenarios, for validation purposes, resulting in a notable acceleration of the entire optimization process in terms of computational efficiency and time, while upholding its efficacy and accuracy.

Our study yields a spectrum of optimal solutions that characterize the intricate relationship between design modifications and the trade-off between cost and x-direction motion. By systematically varying parameters within the mooring system, we identify key factors influencing performance and cost considerations. The results provide not only optimal solutions but also a nuanced understanding of the design space, enabling informed decision-making in the context of offshore renewable energy projects.

Presenting Author: Zohreh Moradinia Queen's University Belfast

Presenting Author Biography: I am Zohreh Moradinia, currently pursuing a PhD at Queen's University Belfast with a specialization in Electrical Engineering and Computer Science. My research centres on the application of transprecise computing in multiphysics simulations. My academic journey includes extensive research in electromagnetics waves and electrical engineering, underscoring my commitment to scholarly pursuits. I eagerly anticipate participating in discussions surrounding technological advancements and research breakthroughs at the upcoming conference.

Authors:

Zohreh Moradinia Queen's University Belfast
Hans Vandierendonck Queen's University Belfast
Adrian Murphy Queen's University Belfast