Adjoint optimisation of hydrofoils for multiple conditions
Adjoint optimisation of hydrofoils for multiple conditions
Hydrofoils in yacht racing are operated across a wide range of conditions. Between these conditions the state and the trim of the yacht changes significantly. This must be reflected in the optimisation of the hydrofoils. Previously, we have performed hydrofoil optimisations for a single condition. These were conducted using an adjoint velocity prediction program (VPP) in a gradient based optimisation routine. The method allows accurate optimisation with hundreds of design variables at a small computational cost. This framework is now extended to optimise foils for multiple conditions. Therefore, the VPP is solved for different conditions and returns the respective velocities made good (VMG) and gradients. The different VMG and the according gradients are averaged to enable the multi-condition optimisation within a single objective. The framework is used to optimise the hydrofoil of an America´s Cup yacht for six conditions derived from historical weather data. The optimal foil is perfectly smooth and exhibits the expected features. It is 7.2% faster than the initial foil in terms of average VMG. This amounts to an 81.9 second timesaving over a whole race. The multi-condition optimisation can be performed at no additional cost and demonstrates the immense capabilities of adjoint VPP-based hydrofoil optimisation.
Adjoint Method, Algorithmic Differentiation, Hydrofoil, Multi-Condition, Optimisation
Society of Naval Architects and Marine Engineers
Tannenberg, Rafael Andreas Maximilian
bd497a87-ea59-4798-a28e-804d3fcb713b
Hochkirch, Karsten
11d869b5-8f02-46a8-8851-abe2105048fd
Turnock, Stephen
d6442f5c-d9af-4fdb-8406-7c79a92b26ce
Boyd, Stephen
bcbdefe0-5acf-4d6a-8a16-f4abf7c78b10
12 March 2025
Tannenberg, Rafael Andreas Maximilian
bd497a87-ea59-4798-a28e-804d3fcb713b
Hochkirch, Karsten
11d869b5-8f02-46a8-8851-abe2105048fd
Turnock, Stephen
d6442f5c-d9af-4fdb-8406-7c79a92b26ce
Boyd, Stephen
bcbdefe0-5acf-4d6a-8a16-f4abf7c78b10
Tannenberg, Rafael Andreas Maximilian, Hochkirch, Karsten, Turnock, Stephen and Boyd, Stephen
(2025)
Adjoint optimisation of hydrofoils for multiple conditions.
In SNAME 25th Chesapeake Sailing Yacht Symposium, CSYS 2025.
Society of Naval Architects and Marine Engineers..
(doi:10.5957/CSYS-2025-005).
Record type:
Conference or Workshop Item
(Paper)
Abstract
Hydrofoils in yacht racing are operated across a wide range of conditions. Between these conditions the state and the trim of the yacht changes significantly. This must be reflected in the optimisation of the hydrofoils. Previously, we have performed hydrofoil optimisations for a single condition. These were conducted using an adjoint velocity prediction program (VPP) in a gradient based optimisation routine. The method allows accurate optimisation with hundreds of design variables at a small computational cost. This framework is now extended to optimise foils for multiple conditions. Therefore, the VPP is solved for different conditions and returns the respective velocities made good (VMG) and gradients. The different VMG and the according gradients are averaged to enable the multi-condition optimisation within a single objective. The framework is used to optimise the hydrofoil of an America´s Cup yacht for six conditions derived from historical weather data. The optimal foil is perfectly smooth and exhibits the expected features. It is 7.2% faster than the initial foil in terms of average VMG. This amounts to an 81.9 second timesaving over a whole race. The multi-condition optimisation can be performed at no additional cost and demonstrates the immense capabilities of adjoint VPP-based hydrofoil optimisation.
This record has no associated files available for download.
More information
Published date: 12 March 2025
Keywords:
Adjoint Method, Algorithmic Differentiation, Hydrofoil, Multi-Condition, Optimisation
Identifiers
Local EPrints ID: 503122
URI: http://eprints.soton.ac.uk/id/eprint/503122
PURE UUID: 808aea76-f0ab-4402-84b9-6aa2b45a5e5d
Catalogue record
Date deposited: 22 Jul 2025 16:33
Last modified: 23 Jul 2025 02:07
Export record
Altmetrics
Contributors
Author:
Rafael Andreas Maximilian Tannenberg
Author:
Karsten Hochkirch
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics