Project description
Revolutionising renewable energy storage
Renewable energy storage and production depend on electrochemistry, and reversible solid oxide fuel cell/solid oxide electrolysis cell (SOFC/SOEC) technologies are crucial for this purpose. These technologies involve various electrochemical processes, such as surface reactions, ionic diffusion, charge collection and conduction, which occur within a limited region and require characterisation at the appropriate place, time frame, and operating conditions to understand the key parameters. To make time-saving predictions, multiscale modelling is necessary to transform experimental datasets into a genuine scientific description. The EU-funded KNOWSKITE-X project will develop a knowledge based methodological approach to finding new electrode materials based on perovskites for reversible SOFC/SOEC technologies. It will use machine learning, deep learning, and AI-enabled tools to create a new generation of materials.
Objective
We target a knowledge-based methodological entry to the finding of new generation electrode materials based on perovskites for reversible SOFC/SOEC technologies. The latter are archetypal complex systems: the physico-chemical processes at play involve surface electrochemical reactions, ionic diffusion, charge collection and conduction, which all occur timely within a very limited region. Hence, true in-depth understanding of the key parameters requires characterisation at the right place, at the right time frame and under the proper operating conditions. The price to pay for achieving this multiply-relevant characterisation is the involvement of non-trivial, advanced characterisation techniques. Multi-scale modelling will contribute to turn experimental datasets into a genuine scientific description and make time-saving predictions. In KNOWSKITE-X, the coupling between theoretical and experimental activities is made real by the choice of partners, who are all active in genuinely articulate theory and practice to understand active systems. To provide unifying concepts and to widen the project’s outcomes, intensive collaboration with knowledge discovery using machine-learning and deep learning methods is planned and AI-enabled tools will be used to compensate the smallness of relevant datasets. Such efforts are intended in view of building strong correlations capable of establishing robust composition-structure-activity-performance relations and hence, lead the way to knowledge-based predictions. By doing this, we also target the implementation of simplified testing protocols and tools operable by industrial stakeholders, which results can be augmented thanks to the knowledge-based pivotal correlations implemented during the project. To this end, dedicated efforts will be made in certifying the interoperability and usability of the project’s advances in the form of harmonised documentation and open science sharing.
Fields of science
Programme(s)
Topic(s)
Funding Scheme
HORIZON-RIA - HORIZON Research and Innovation ActionsCoordinator
75794 Paris
France