Project
THOR has five key impacts: industrial (cost reduction), scientific (improved understanding of battery mechanisms and materials), economic (increased business opportunities), and societal (motivation for younger generations to study electrochemistry and take part in battery R&D), regulatory (acceptance of virtual tests as part of the certification process).
Objectives
Develop a highly predictive model
for assessing performance at the cell, module, and pack levels
This model operates with a minimal set of parameters, thus reducing the need for extensive experiments and a large sample size. Additionally, establish the methods:
-
To acquire the necessary input data for the model
-
To replicate the model across various chemical compositions
Create a sophisticated physics-based aging model
at the cell, module, and pack levels
It will use a smart parameterisation method for accurate updates, even when parameters are interconnected and affected by chemistry and operational conditions, which can lead to non-linear degradation.
Develop a two-dimensional model
at the cell, module, and pack levels
It can predict the thermal (heat release) and toxic (gas emissions) hazards resulting from the thermal runaway of a battery. The model will cover various failure scenarios, both thermal and electrical, and will be based on cell parameters derived from a limited set of tests, enabling quick adaptation to different chemical compositions.
Construct a real-time multi-scale Digital Twin
using artificial intelligence-driven surrogate models (Reduced Order Models – ROMs)
These should rely on prediction capabilities for performance. The Digital Twin should feature a user-friendly graphical user interface (GUI) that allows for easy visualization of relevant results, such as number of cycles and temperature distribution.
Develop intelligent design of experiments (DoE)
and methodologies to identify the most influential parameters for each model.
Also, establish processes for characterising batteries, including collecting input parameters for the models and the Digital Twin. These processes are efficient, enabling rapid and cost-effective assessments.
Innovations
THOR will enhance the battery industry by improving efficiency, reducing costs, enhancing safety, and promoting sustainability.
A Digital Twin
This is a virtual representation of a physical battery system. The Digital Twin will be used for experimenting surrogate models powered by artificial intelligence and will be licensed to early adopters.
-
It will automatically generate ideal battery designs based on expected characteristics, facilitating faster and more cost-effective battery development, reducing the need for physical prototypes and minimizing material waste. This improved design increases the battery lifetimes, reliability and safety.
-
Digital Twin will enable maintenance programs based on performance and safety predictions, reducing operational costs, outages, and scheduled battery replacements.
3 physics-based models
These represent the performance, ageing and safety behaviour of a battery system at different scales: cell, module and pack.
Safety and risk reduction
By providing a better understanding of safety hazards and aging mechanisms, the project aims to improve battery safety. This, in turn, can reduce insurance and warranty costs and increase consumer confidence.
Better energy management
The project will also contribute to further reducing operational costs and optimising recharging based on real discharge profiles. It also facilitates the reuse of retired batteries, reducing investment costs for stationary applications.
Outcomes
-
Closing of the loop with the possibility of predicting the best set of parameters for a given behaviour.
-
Democratisation of virtual testing: performing virtual tests will become more common. In 10 years, 80% of industries will use a digital twin during the battery development phase.
-
New standards for battery data: harmonisation of the data related to battery testing. The possibility to share data among projects will concurr the development of open-source databases.