BIO
Dr Jessica Guichard is a post-doctoral researcher at the University of Plymouth, specialising in the integration of offshore renewable energy with hydrogen production. Her research combines experimental and numerical approaches, including comparisons between experimental data and CFD simulations for various wind generation system prototypes, simulations of flow around an outboard motor, and the design of a wind turbine blade for scaled testing of floating wind turbine platforms. She has also conducted experimental investigations of sloshing in a container on a barge, as well as the motions of a flexible riser for hydrogen transport from a floating wind platform to the seabed. Her work further includes cost optimisation and energy system modelling of offshore renewable energy systems coupled with hydrogen production using Python for Power System Analysis. She is currently investigating experimentally the impact of different baffle configurations on sloshing in a cylindrical container.
PROPOSAL
Robust prediction of wave-structure interaction (WSI) is pivotal for offshore renewableenergy systems yet remains computationally demanding when relying solely on high-fidelity CFD models. CCP-WSI convenes the UK’s CFD and structural mechanics communities to develop advanced coupled WSI tools, benchmark cases, and shared methodologies, while HEC-WSI provides inclusive access to national HPC resources that enable large-scalesimulations and sustainable community software development.
Dr Jessica Guichard is an ORE Integration Researcher at the University of Plymouth withexpertise spanning energy-system optimisation, CFD modelling, and experimental hydrodynamics. Her work includes optimisation of rSOC (reversible solid oxide cells)-offshore wind integration using PyPSA (Python for Power System Analysis), experimental investigations of wave structure interaction phenomena, hydrodynamic testing of flexible risers, floating wind turbine platforms, and sloshing in barge-mounted containers. She has also undertaken comparative studies between laboratory experiments and CFD for developing a laboratory wind generator suited to floating wind hydrodynamic testing. This blended background in system-level energy modelling, experimental design, and CFD-relevant hydrodynamics, underpins her aspiration to apply AI tools to create physics-integrated surrogates and reusable models that can accelerate WSI prediction while reducing the cost and turnaround time associated with traditional CFD.
Research at the University of Plymouth has generated a rich set of WSI datasets ideally suited for AI-enabled analysis: (i) COAST Laboratory experiments, including floating platform tests, and empty-basin hydrodynamics with motion, load, and free-surface measurements; (ii) CCP-WSI blind-test cases, such as sloshing experiments, which can be augmented with analytical solutions to form synthetic training sets; (iii) flexible riser WSI experiments relevant to emerging hydrogen demonstrator systems; and (iv) numerical datasets generated through CCP-WSI and HEC-WSI CFD workflows under combined wind, wave and current conditions.
The project will focus on a selected dataset, prioritising coverage of key WSI physics -nonlinear wave loading, hydro-elastic response, and mooring/riser coupling – while ensuringdata quality in terms of metadata completeness and temporal/spatial resolution. The objectiveis to learn, explore and develop suitable AI models capable of accelerating CFD and other numerical WSI tools. Within the Learn AI stream, the project will establish robust baselines using temporal architectures such as CNNs, RNNs and Transformers for prediction of time-series loads and motions, supported by structured data-processing pipelines fornormalisation, cross-validation and benchmarking.
To progress within the NPRAISE Learn AI stream, the project will trial physics-informed neural networks (PINNs) and operator-learning surrogates to map wave conditions to pressure fields, global loads and 6-DOF responses. To directly accelerate numerical models,the project will explore hybrid AI-numerical loops where AI-derived initial conditionsreduce computational cost or replace selected sub-steps with fast surrogate inference. Model credibility will be established using Blind-Test-style validation with holdout experimental cases, and uncertainty quantification and error-budget analyses will be reported alongside speed-up metrics. Trained models, scripts and adapters will be packaged for CCP/HEC-WSI communities, supporting reproducible integration into emerging AI-WSI workflows.
Once validated, the hybrid AI-numerical approach can be extended to additional datasets, enabling scalable development of AI-accelerated WSI models. All outputs will be FAIR (Findable, Accessible, Interoperable, Reusable)-aligned and sustained via CCP-WSI channels, including versioned datasets, numerical models, reports and documentation foracademic and industrial stakeholders.
The anticipated impact is to lower entry barriers, standardise validation practices and accelerate design cycles across WSI, supporting the UK’s ambitions in offshore renewables and hydrogen production with more rapid, trustworthy predictive tools.