The first open standard for physics-based lithium-ion battery models has been launched by the Faraday Institution. It has come up with Battery Parameter eXchange (BPX), which it argues is designed to make it easier for manufacturers to access highly accurate physics-based models as part of their development processes, potentially reducing the cost and time-to-market of new batteries.
Businesses in the sector are being approached to join a new standards body, which will own and drive the development roadmap for BPX and facilitate its adoption across the industry.
BPX defines the battery parameters, the equations that use the parameters, and the reporting of experimental measurements used to validate the reported parameters.
The Faraday Institution, the UK’s independent institute for electrochemical energy storage research, skills development, market analysis, and early-stage commercialisation, believes BPX will make it easier for physics-based models to be deployed in a broad range of scenarios in industry, cutting costs and stimulating innovation.
Importance of battery models
When it comes to developing new batteries, creating and testing physical prototypes is an expensive business. This, says Robert Timms, a postdoctoral research assistant at the Mathematical Institute at the University of Oxford, drives the need to create accurate software models for virtual prototyping, where promising batteries can be simulated to understand their performance and how they behave in certain scenarios, as well as how cells might degrade as they get older.
Timms explains, the most popular type of battery model currently in use in industry is the equivalent circuit model.
“This type of model uses traditional electrical components like resistors and capacitors to explain how the battery is behaving overall,” he said. “Typically you cycle your cell, apply some current and measure the voltage and temperature. You then use that cycling data and match it to the parameters of an equivalent circuit and say ‘this is how I think my battery will behave’.”
Such models are simple to apply, Timms said, making them popular in industry, but they are not without their drawbacks. “They don’t really tell you why something is happening, they just describe what’s probably going on.”
With the number of use cases for lithium-ion batteries growing by the day, highly accurate models are required which enable new batteries and battery-based systems to be developed more quickly, and which deliver a more thorough understanding of the performance of a battery over its entire lifecycle. This is particularly important when looking at ways to optimise a cell’s performance as it degrades.
Write down equations
This is where physics-based models come in. “With a physics-based model, you write down equations for all the things that you think are going on inside the cell to build up a picture of how it will behave,” Timms said. “Things like charge and mass and heat being transported around the cell. The benefit of these models is they give you much more insight into why things are happening; you can start asking why a battery is behaving in a certain way and try to understand that. But they’re also much more complicated and difficult to parameterise.”
Battery tear down
Indeed, gathering data for a physics-based model often involves taking a battery to pieces and measuring the performance of individual components, a time-consuming and costly process known as “tear down”. Not only does this require resources beyond those available to the average SME or start-up, but it also depends on a large amount of computer power being available to crunch all the numbers and accurately model a battery’s performance.
The sheer number of physics-based models – and varieties of individual models – further complicates matters. “There’s a very popular model, the Doyle Fuller Newman model, which has been around since the 1970s,” Timms said. “But even if you say to someone ‘I’m using this model’, there are a bunch of different formulations, and people call things by different names which mean the same thing.”
Standardisation has the potential to solve these problems, making highly accurate, ready-to-use models for businesses that aren’t able to develop their own. This in turn has the potential to create a more diverse battery supply chain, providing more choice for battery manufacturers and end-users, and creating opportunities for companies, particularly advanced engineering consultancies and battery component developers.
Quite apart from the time and money saved, the technical benefits for those working on new batteries are substantial according to Timms. “It removes ambiguity. If you’re relying on another organisation to parameterise a cell, you want to make sure you’re using those parameters in the same way.
“It’s also the case that there are a lot of tools out there, so even within one organisation you’ll have teams working on different parts of the battery lifecycle who will be using different tools that suit their particular job. But they need to be able to talk about the same cell or the same model, so they need the ability to take a parameter set and know they’re going to get the same result whatever tool they load it into.”
Enter BPX
Timms is part of the Faraday Institution Multi-Scale Modelling Project, a multi-disciplinary research team led by Imperial College London with eight other academic research and 14 industry partners, developing better battery models.
As part of this project, two tools have been created – Python Battery Mathematical Modelling, better known as PyBaMM, and DandeLiion. “PyBaMM is an open-source battery modelling with a focus on extensibility for including new physical mechanisms,” Timms said. “It has a big community of thousands of people involved with it from around the world and lots of people from academia contribute models to it. I’m part of the core development team for PyBaMM and we want to make sure it’s as easy as possible for industry to use.
“DandeLiion is slightly different in that it’s focused on one particular model and highly optimised for large-scale simulations of battery packs. So these are two very different tools, and with BPX we wanted to create a standard that could be understood by both.”
Making the two tools interoperable is key to fulfilling their potential in industry, said Pete Keevill, BPX Project Lead at the Faraday Institution. A successful telecoms entrepreneur by background, he was brought on board to aid the commercialisation of the outputs of the Multi-scale Modelling Project, and quickly spotted the requirement for a standard like BPX.
Whole greater than the sum
“When I looked at PyBaMM and DandeLiion, it was clear to me they fit into different places in the battery development cycle, and I was convinced that the whole would be greater than the sum of the parts,” he said. “By presenting them as a combined tool set, it seemed like they would be more powerful commercially, but then the problem becomes how do you find a common format to describe the battery cell between the two tools?”
Beyond PyBaMM and DandeLiion, it soon became clear to the Faraday team that there was a desire in the industry for a standard physics-based model to be developed. “We were talking to potential industrial partners and we were hearing that the availability of models themselves was a big barrier to the use of physics-based models,” said Keevill. “They’re difficult and expensive to build for businesses, but with the skills available within the Faraday Institution, coming up with a common format was something we were able to do relatively easily.”
And so BPX was born. To develop the standard, the researchers used a definition of the Doyle Fuller Newman model set out in a recent review article published by Faraday Institution researchers who were looking at a broad continuum of physics-based battery models. They then decided on the JSON human-readable file format for the parameters and created a software package in the Python coding language which can parse a BPX.json file and validate it to check all the correct information is contained.
About:Energy, a spinout from Imperial College London and the University of Birmingham involving members of the Multi-scale Modelling Project, helped to author the standard and developed the first two example models provided in the standard download pack.
BPX is intended as a starting point for using physics-based models, providing a standardised basis on which manufacturers can innovate to suit their specific use cases. Timms said the goal is to make the standard compatible with the many commercial software packages on the market used for battery modelling.
“We’d like to see the other players in the simulation space adopt BPX so that the same file can be passed to Comsol, or GT-AutoLion, or the other packages which are used by battery manufacturers,” he said. “So far the industry has been supportive of the idea, because they can see the potential to speed things up and make life easier. If the software companies believe there’s an appetite there, it will make them much more likely to commit.”
Building a standards forum
Indeed, getting buy-in from the wider ecosystem of automotive companies, battery OEMs and software developers involved in the battery supply chain will be vital to BPX’s success. To encourage this wider adoption, the Faraday Institution plans to launch the Battery Modelling Standards Forum, an industry-led organisation that will own and maintain the standard and build a clear technical, commercially informed, roadmap for its development.
Keevill is leading on the development of the Standards Forum, and said: “Once we’d built BPX, it made sense to make it publicly available, because it’s something that’s missing in the battery industry. But for that to work there needs to be a standards body behind it made up of users who are interested in its development and want to amplify the impact it can have.”
Physics-based models
BPX launched with two webinars, one making the case for physics-based models, and another which took a deep dive into the technical details of the standard. Keevill said the task now is to encourage interested parties to sign up to the Standards Forum, which could launch later this year.
For interested businesses, joining the Battery Modelling Standards Forum is an opportunity to keep abreast of developments in the fast-moving electric vehicle (EV) battery technology space, to discuss challenges and potential solutions with peers, and to get a better handle on the evolving requirements of customers and end users.
Keevill believes that, with the backing of the Faraday Institution’s deep knowledge of battery models, the Standards Forum could be well placed to support the growth of BPX, as well as other emerging battery standards.
“We think the Forum is the best way to ensure BPX remains relevant and useful and to create other battery modelling standards as necessary in a coordinated way to ensure they remain coherent in the way they operate,” he said.
“Coming into the battery world from telecoms, I was surprised something like BPX didn’t already exist. The scale of the industry, and the speed at which it’s moving, means more standards are likely to develop for other battery modelling applications.
“BPX is the immediate focus, but if you look across cell design, system design, optimisation of deployed systems and then add in all the second-life applications and recycling, there’s a lot of different battery modelling applications and quite possibly a need for standardised frameworks to deliver overall efficiency of the industry.
“It makes sense to try and keep them all open and consistent, which is why a body like the Battery Modelling Standards Forum is sorely needed.”
Timms also expects other standards to develop alongside BPX, and believes the Faraday Institution is well placed to play a leading role in how these are defined.
“Battery models have been slow to the game, in part because of the complexity of the models,” Timms said. “For those working in research and industry, it’s often difficult to agree what the correct model is and that has stopped us moving forward on standards. What we want to do with BPX is to lead that conversation about standardisation.”
