Session: 08-06-01 Integrated Design & Environmental Assessment of Floating Offshore Wind Systems

Submission Number: 180523

Combined Global-To-Structural Optimization Methodology for Material Efficient Floating Offshore Wind Substructures 

The industry is targeting an LCOE of $45 per MWh by 2035 for floating offshore wind compared to more than $90 today. This ambitious goal can only be reached by cutting costs wherever possible. With the floating structure accounting for up to 30% of the CAPEX, maximizing its material efficiency is an indispensable step in Floating Offshore Wind Turbine (FOWT) design. 

Current workflows in floating platform design, however, are limited in computational tractability and representativeness of the physical system as a whole. Typically, analysis threads are run independently: the coupled analysis will be carried out in one environment and transferred to a FEA model through fragile format conversion scripts. For every design iteration, time consuming hand operations are needed which hinder the automatization and scaling of the global-to-local design loop. As a result design choices are made sequentially, addressing the global performance first and structural performance after without integrating their inter-dependencies. This issue is compounded by the high cost of conventional FEA methods. When we consider the many query nature of optimization workflows in addition to the already large set of load cases, this means that the known benefits of in-depth optimal design exploration and optimization are inaccessible to FOWT design. 

Here we present a novel framework that unlocks combined global-structural design optimization for floating offshore wind platforms. The approach is underpinned by the RB-FEA methodology for structural analysis which generates a 1000x speed-up over standard FEA, thus addressing the tractability issue. The technology stack is made up of (i) the Integrated Load and Structural Analysis (ILSA) methodology for FOWT platforms, (ii) a multi-fidelity optimization approach that addresses global floater shape to local member dimensions. The ILSA toolchain integrates together the Capytaine BEM solver for hydrodynamics analysis, OpenFAST for global FOWT analysis, Akselos’ RB-FEA and load mapping methodology. This delivers a seamless fully integrated cloud-based analysis pipeline that generates holistic, high fidelity and high speed combined global-to-structural FOWT assessments, and where crucially solver-to-solver transfers are handled automatically. The multi-fidelity optimization framework integrates the ILSA combined solver into a 2-level architecture for efficient design space exploration. The first fidelity level optimizes the global floater shape using a machine learned surrogate model of the design space; the second level optimizes the floater’s internal member dimensions using geometric parameters embedded in the RB-FEA reduced basis. 

In this paper, we demonstrate the optimization workflow for the open-source OC4 DeepCWind semi-submersible platform. Design variables cover the global floater shape through column height, diameter, inter-column spacing among others, as well as internal element dimensions such as shell and stiffener thicknesses and positioning. The performance attributes contain metrics linked to both global performance such as energy production, or nacelle acceleration, and structural performance, such as strength Utilization Factor (UF). Analysis conditions correspond to DLC 1.6 for FOWTs in operation under severe sea state which is known to impact structural integrity. We run the optimization with the objective of minimizing structural mass under constraint on energy production, nacelle acceleration and strnegth UF. 

We show that using our methodology generates a substantial >10% reduction in structural mass without downgrading FOWT performance. Not only that, we demonstrate how using this automated design optimization approach speeds up engineering time by accelerating end-to-end design from months to days while using fully standard compliant numerical tools.

Presenting Author: Jeremy Nahon Akselos

Presenting Author Biography: Jeremy Nahon is a Software Research and Development Engineer specialised in computational modelling of physical phenomena. He has accumulated 10 years of experience in CFD and FEA analysis applied to design. At Akselos, he is in charge of developing features tailored to fixed and floating offshore wind technologies.

Authors:

Jeremy Nahon Akselos
Romain Pinguet Akselos
Mathieu Ancellin Mews Labs
Thang Do Akselos
Daria Mesbah Mews Labs
Sylvain Vallaghe Akselos
David Knezevic Akselos
Daniel Chauveheid Mews Labs