12/02/2026
The Alan Turing Institute / Olesinski Collaboration.
Turing Research Fellow Zack Xuereb Conti, from the Alan Turing Institute recently completed a 14-month secondment with Olesinski Naval Architects as part of a wider agreement between Olesinski and the UK’s national institute for data science and AI. Zack worked closely with Bill Edwards, Olesinski’s Head of R&D.
Challenge Addressed -
High-fidelity Computational Fluid Dynamics (CFD) simulations for a single hull geometry can take several hours or days to complete, depending directly on the fidelity required. This severely limits design exploration and slows down hull form optimization cycles. We partnered with The Alan Turing Institute to develop a machine learning based solution that could predict the hull flow rapidly, accurately and data efficiently. The strict time-budgets inherent to bespoke yacht engineering demand a parsimonious data-driven solution that does not depend on large simulation budgets for training.
Solution -
Zack developed a lightweight, data-efficient machine-learning pipeline capable of predicting hull surface pressure and wall shear stress orders of magnitude faster than traditional CFD, while retaining meaningful physical accuracy.
What was delivered?
An end-to-end software pipeline providing:
• Field data pre-processing for machine learning with fully internal mesh handling, eliminating reliance on external software for batch-processing
• Feature extraction
• Training and validation of reduced-order surrogate models for rapid, uncertainty-aware flow prediction to support safer decision-making
• Customised visualisation workflows
We would like to acknowledge TotalSim Ltd. and Professor Gabriel Weymouth (Delft University of Technology / TU Delft ) for their respective expert insight throughout the project.
