Electric vehicle (EV) batteries are on track to becoming a US$111bn global industry by 2032—a 64% increase on 2024, according to Fortune Business Insights. An essential component of electrified mobility, including resurgently popular hybrid vehicles, batteries are expected to become cheaper, smaller, and more powerful in the second half of the decade.
While exploring new, potentially game-changing chemistries has value, many automakers and suppliers want to discover how existing platforms can be optimised. However, the paucity of data within the industry concerning how EV batteries operate in the real world makes this challenging. Finding a way to unlock that information and leverage it for future R&D could have significant implications for battery durability, performance, and safety.
Mining company Fortescue believes its intelligent software platform, Elysia, can help the automotive industry finally understand batteries. The solution started out as Williams Advanced Engineering Technologies, which sought to optimise the Williams racing team’s EVs during the intense rigours of the track. The company and its technology were acquired by Fortescue in 2023 as part of its wider push into the green energy space.
“We were at the cutting edge of what was possible in terms of energy and power density,” says James Wallace, Co-Founder and Head of Product at Elysia. Now, the team wants to optimise lithium-ion batteries across every vehicle segment.
Performance: today and tomorrow
Wallace believes all EV manufacturers are grappling with the same three challenges: making batteries smaller to reduce costs and weight, extracting the best performance possible, and ensuring system safety at all times. As brands implement net zero targets and strategies for achieving fully electric product line-ups, harmonising these goals will be crucial for convincing customers to make the switch from internal combustion engines (ICE).
In May 2024, Jaguar Land Rover (JLR) signed a multi-year deal to incorporate Elysia across its entire luxury EV stable, starting with the Range Rover Electric. This is part of the automaker’s Reimagine strategy, which aims to redefine modern luxury by combining EVs with sustainability and social initiatives. With that transition comes a lot of data: “JLR generates 86 billion diagnostic data points every month across all its EVs,” Wallace tells Automotive World. But how can this voluminous quantity of information be parsed to improve batteries?
“It’s key to understand how a battery will perform tomorrow, not just today,” he states. During a project to optimise electric mining trucks, the Elysia team observed the “perfect recipe” for ruining a battery: prolonged exposure to extremes of heat in harsh terrain and a constant need for fast charging to maintain uptime. “Our thinking was the same as on the racetrack: push the batteries to their limits and measure whether the packs would need repairing, replacing, or upgrading during their operation lifetime.” The knowledge it gained subsequently helped Fortescue understand how to maximise battery tech for passenger EVs.
Extending battery life
Elysia emphasises not only discovering how components integrate but also modelling and monitoring. “A lot of commercial and academic battery models are difficult to use: you need to be a real expert to use them,” says Wallace. Furthermore, they often utilise lab findings instead of real-world data, limiting their usefulness for solving OEMs’ product issues. “Battery lifespans are difficult to predict because they’re determined by use. The health of batteries in two identical EVs could be totally different depending on how they’re charged and the conditions in which they operate.”
Through physics-based models, Elysia helps break down batteries into collections of separate but interconnected systems. By combining real-time time data from batteries actively used in vehicles with digital twins of the same cell types, its software can run prognostics to detect potential hardware faults before they occur. Elysia is offered as both a streamlined cloud platform as well as embedded battery management system algorithms. Fortescue estimates that its solution can extend battery life by up to 30%.
Although Elysia uses probabilistic machine learning, Wallace states that background data and experience of how batteries operate in the field is the vital ingredient for using artificial intelligence (AI) tools effectively. Automakers often have narrow testing timeframes, and a scattergun approach using “black box AI algorithms” is unlikely to yield productive results, particularly when factoring in complex variables like battery composition and regional market needs. While some posit that only quantum computers could accurately simulate the miniscule interactions of electrons, Fortescue claims Elysia can provide lifetime insights on battery degradation without getting “bogged down” in difficult electrochemical theory.
Eliminating uncertainty
Beyond its use as an optimisation tool, intelligent software platforms like Elysia could be an essential part of safe maintenance as ICE is phased out. Until recently, technicians in service centres accumulated the technology and training necessary for dismantling, diagnosing, and remediating engine issues over a long period of time. Combined with long-term supplier relationships that helped build intimate familiarity with complete ICE systems, automakers had a solid grasp of their powertrain’s capabilities. Wallace believes this is changing in the era of electrification. “Batteries are much more difficult to take apart and test in situ, and OEMs don’t have the decades-long supplier relationships they’re used to.”
If a system doesn’t have enough data on its battery, it can’t accurately predict its state of health
The recent controversy surrounding an EV fire in a South Korean residential building highlights the importance of ensuring battery system integrity. At the same time, the success of bringing EVs to the mass market hinges on achieving range and recharging capabilities comparable to ICE as soon as possible. In Wallace’s view, using data-driven software analytics platforms is the best way to give automakers immediate clarity on their vehicles’ battery health.
The goal, he continues, is to help automakers determine through software how existing battery designs can be optimised without altering the hardware or compromising safety. If OEMs are informed that batteries are degrading for a specific reason, engineers could then make appropriate adjustments to the EV platforms in a straightforward manner. Importantly, they will be assured by Elysia’s holistic data model that the change won’t produce undesirable outcomes. “That could extend the life of a battery enough that warranty claims are no longer an issue.”
Wallace concedes there are always more insights that a platform like Elysia could include, and overcoming those limitations will depend on greater collaboration on data between Fortescue and automakers. However, he believes it can be a leading offering in the fight to eliminate a significant and existential problem in e-mobility today: uncertainty. “Most people have experienced their phone stating it has 20% charge left, only for it to die two seconds later during a critical call. It’s the same problem with EVs: if a system doesn’t have enough data on its battery, it can’t accurately predict its state of health.”