Platform for computational battery design

In the global quest for decarbonisation of energy systems, batteries have emerged as crucial components. Not only do they promise to free transportation from fossil fuels, but they also offer the flexibility needed for energy systems that rely heavily on intermittent renewable sources. Amid this transition, advances in battery technology are crucial, especially in terms of cost-effectiveness and energy density. Enter SemiSolid lithium-ion battery (LIB) technology. By suspending electrodes in an electrolyte, similar to a clay-like substance, SemiSolid LIBs pave the way for thicker electrodes with higher capacity and a more material efficient cell design. To drive this technology forward, a collaboration between FREYR and NTNU is ongoing, which utilizes artificial intelligence (AI) and machine learning (ML). By developing a computational modeling framework, the project aims to create a digital twin, which facilitates extensive digital experimentation and simulation. This initiative targets to accelerate innovation while ensuring quality assurance and reducing production waste – a crucial step towards establishing a robust battery industry, not only in Norway, but internationally. Nevertheless, there are major challenges. Accurate simulation of LIB parameters remains daunting due to intricate internal processes. ML is promising, but the application of the technique requires careful attention. The project aims to overcome these obstacles by using advanced ML techniques on extensive datasets, both commercial and proprietary, to refine parameter estimation and optimize battery performance.

The SemiSolid lithium-ion battery (LIB) technology is a promising new approach to improve LIBs by bridging the gap between conventional LIB technologies and solid-state LIBs. At the core of the SemiSolid technology are electrodes that are co-suspended with the electrolyte, resulting in a clay-like composition. This allows for a much thinner separator membrane between the anode and cathode and results in batteries with more active material and a higher capacity. FREYR Battery Norway AS (FREYR) is an emerging producer of clean battery solutions based on the SemiSolid LIB technology. Currently, the company is considering building gigafactories in the US and Europe with annual production capacities of several tens of GWhs. Battery design and production represent a new frontier in artificial intelligence (AI) supported production with the potential to dramatically improve future battery cell performance, as well as testing and production processes. In this project, FREYR and NTNU will join forces to develop a software tool for SemiSolid LIB development in the form of a computational modelling framework using a state-of-the-art machine learning approach. The development of tools tailored specifically for SemiSolid LIBs will enable the creation of a battery digital twin, a virtual representation that facilitates comprehensive digital experimentation and simulation. The innovation will result in the following important benefits for FREYR: i. Exploring future innovations of the SemiSolid LIB technology with low effort and less time-consuming experimental work. ii. Enhance the quality assurance processes in battery production and reduce the scrap rate when upscaling the SemiSolid LIB process to gigafactory production volumes.