While vehicle manufacturers (OEMs) can go to great lengths to try and explain the benefits of a particular type of electric motor architecture, the design of the windings found in the stator, the type of battery cells they’ve developed and the type of Lithium-ion chemistry within them – all of which are important technical features – what customers really want to know is: what’s the range, and how fast does it charge?
While, in general, range has been steadily increasing, fast chargers have improved rapidly. The first 100kW DC charger in the UK was installed in 2018; back then 50kW was the benchmark. Now, 350kW chargers are commonplace in motorway service stations. But regardless of the maximum charging power and charging times quoted by OEMs, customers can be very disappointed if their EV doesn’t achieve them.
Even under ideal conditions, in most cases the battery won’t charge at its maximum rate for long. This is deliberate – not a fault as customers sometimes believe – and is done for very good reasons: maximising battery safety and durability. But is it possible to go closer to the limits without compromising those two fundamental attributes?
Most EV batteries are controlled using software that relies on look-up tables that are mapped out against battery temperature, state of charge and state of health. It’s a method very similar to the way in which ECUs control combustion engines – and have done for decades. But when applied to batteries it results in a big difference between actual battery performance and the technical limits of the battery cells themselves. The main reason is that OEMs want to prevent harmful degradation processes that can take place within the cells during fast charging, notably lithium plating.
Lithium plating occurs during fast charging because the high current causes the lithium to build up on the surface of the anode, which can reduce energy storage capacity, increase internal resistance and reduce battery life. Plating can also occur when fast charging at very low ambient temperatures. To prevent this, the battery is pre-heated, again using pre-set look-up table values, but this reduces overall energy efficiency. Whenever plating becomes severe, lithium dendrites can form and, if they grow large enough, can pierce the separator and cause a short circuit within the cell.
Instead of look-up tables, Breathe takes a totally new approach. We developed a physics-based model that runs in real-time and uses closed-loop control to accurately simulate the electrochemical states within the battery cells, primarily electrode potentials. We’re still using the same input parameters, but the model’s accuracy and real-time capabilities enable us to push the cells much harder and get closer to the technical limits while avoiding lithium plating and other harmful reactions.
As well as a performance benefit, we can save on development timescales too. Look-up tables can take years of testing across different cell sample stages before being complete. Working with some sample cells from the OEM, and in partnership with their battery management software (BMS) team and battery system design team, we’re able to parameterise and calibrate our model in approximately 6-8 weeks. And because it’s a flexible platform, we can easily adapt it to suit different cell formats and chemistries while the fundamental physics remain the same.
From there we can provide a drop-in embedded software solution tuned to a specific battery that’s ready to go into the application layer within the BMS. After the integration engineering is complete, we work with the OEM to define a validation strategy, and then help with the simulations, rig testing, and whole-vehicle prototype testing and validation, including cold and hot weather testing.
Ideally, we’ll be part of a vehicle programme from the start so that we can help influence the design of the battery system to enable the maximum gains. We have competitors, of course, but their smart charging software is still focused on state-of-charge and accurate state-of-health estimation. These will always be very important, and give insight into how the battery is doing, but they’re not actively modelling or controlling anything that improves the charging experience, and that’s what the OEMs are looking for.
We know we need to stay ahead, so we’re continuously improving the model to make it more accurate, and we do this through our collaborative research with OEMs and cell suppliers. We’re also looking at other aspects we can control, and degradation mechanisms other than lithium plating, for example. Learning from the consumer electronics industry – where our technology has also been applied – transfers back to automotive because the development timescales and product lifecycles are so much shorter. We see advances in cell chemistries happening much faster there, so what’s new in EV batteries may have been in phones for a while. Innovations on the software side also happen much faster too, and we apply those innovations as well.
As the automotive industry begins the new era of the software-defined-vehicle, we believe such vehicles need software-defined batteries.
Dr Christian Korte is Head of Software and Engineering at Breathe Battery Technologies.
Study calls for changes to public toilet design
Rather bad new for Dyson then. Compared to the traditional mushroom head push button + nozzle type hot air hand dryer that dried your hands with,...