Electrification of vehicles and other mobile equipment relies on the widespread availability of rechargeable batteries with good performance, reliability and safety. Monitoring and checking the condition of the battery is therefore crucial. The development of the required battery management systems (BMS) benefits from Model-Based Design. That is why MathWorks, which jumped on the electrification bandwagon early on, continues to build industry-leading solutions for BMS.
Fully charged for the electrification trend
Checking and monitoring battery systems is a perfect application for MathWorks, say Javier Gazzarri, Mohamed Anas and Marcel Stakenborg. MathWorks’ comprehensive set of battery management systems solutions is the logical consequence of the way it has organised its R&D. ‘We maintain close relationships with OEMs and suppliers and recognise trends at an early stage. For instance, we spotted the emerging electrification of cars and other vehicles and appliances as early as ten years ago.’ Electrification of vehicles and other mobile equipment requires rechargeable batteries to store electrical energy and BMS systems to control these batteries. ‘Efficiency, costs and reliability are the most important themes, and they require businesses to innovate faster and faster. To innovate in this area, both battery manufacturers and battery users depend on two main activities: experimenting with new battery chemicals and developing performance improvement algorithms. Algorithm development is core business at MathWorks, so our focus on BMS has come very naturally. Many years ago, we saw huge potential in the market for modelling and simulation tools to support the electrification trend,’ says Mohamed Anas.
‘The range is increasing and is sufficient in most cases’
‘The biggest electrification challenges we encounter in Dutch industry, including mechanical engineering, automotive, railways, offshore and communications, revolve around cheaper electricity and storage. Engineers rely on the BMS to monitor and optimise the performance of their batteries, estimate internal health and predict how much energy is left. In an electric car, for example, you have the battery, the motor and the power electronics to connect them together. We provide software solutions for modelling and simulating these components individually and as part of a system, enabling the development and optimisation of closed loop algorithms.’
State of charge
One of the MathWorks experts in this field, employed since 2011 and based in the US, is Javier Gazzarri, a consulting application engineer for electrification and BMS. ‘We are targeting both battery manufacturers and users. A typical question from a battery user is: “I want to electrify this device or vehicle, it has to last so-and-so, with these weight and volume limitations. Now I have to make sure the battery delivers this kind of performance and that a certain level of performance is guaranteed after all these years.” This is a typical situation for the Model-Based Design of a battery and its management. We provide software to create models of battery packs and to simulate their use. Estimating the state of charge (SOC), for example, is much more complicated than measuring the fuel level in a conventional vehicle because you can’t measure the charge of a battery directly. You measure quantities such as voltage, current and temperature and combine them with a model of the battery to estimate the SOC. Models and simulations can be used to determine the size of the battery that meets the requirements and to determine which heat management system is needed in connection with the heat dissipation through the pack. The battery must behave well at a specific level of currents, voltages and temperatures. The information provided by modelling and simulation is crucial for BMS design.’
Performance, sustainability and safety are the main motives for technology development in this area, according to Gazzarri. ‘For electric vehicles, for example, everyone wants them to be at the same level as for conventional vehicles. Safety is of course of the utmost importance. To this end, the BMS must ensure that the battery is not overcharged or excessively discharged and that it maintains the battery temperature within the limits specified by the manufacturer. In addition, it distributes the charge over the battery pack and makes estimates of the state of charge and health. Lastly, the cost of the battery is a significant part of the total cost of the vehicle, so a lot of effort is put into cutting costs and working on materials science and manufacturing.’
Battery not a bottleneck
Previously, battery technology seemed to be a bottleneck in the electrification trend, with issues such as the ‘range anxiety’ of drivers who weren’t sure they could reach the next charging station before the battery ran out. That is no longer the case, says Gazzarri. ‘The range is increasing and is sufficient in most cases. However, one of the most difficult issues to address is battery degradation; a lot of effort is being made to understand this phenomenon.’ When asked about the barriers to electrification, customers mentioned two things, Anas adds. ‘First, there are supply chain issues for systems and components, which is in line with the general situation in the industry. And secondly, there is the necessary infrastructure, the possibility to connect to the electricity grid. There may be a mismatch with the old grid system; that’s a challenge for many customers.’
Most customers use Li-ion batteries, Gazzarri and Anas know. ‘But our approach is chemistry agnostic’, says Gazzarri. ‘We don’t directly model the chemical reactions and don’t require input of physical parameters; instead, we propose an electrothermal analogy that behaves like the battery in terms of voltage, current, and temperature: those are the only measurements you can make in a real system.’ Data on these quantities is used to develop models in terms of equivalent circuit descriptions with parameters such as open circuit voltage, internal resistance and capacity. ‘This gives us some knowledge about the battery in question – we call it fingerprinting the battery. We would therefore not call our model a black box, but rather classify it as a grey box. In addition, we also need to keep track of how these parameters change with the number of cycles and the age of the battery.’ Anas: ‘A next step in the further development of the models is the introduction of the electrochemical effects, for example, through reduced-order modelling from electrochemical battery models.’
Over the years, MathWorks has demonstrated the development of BMSs using Model-Based Design. MATLAB and Simulink modelling and simulation capabilities, combined with Stateflow® and Simscape functionality, enable BMS development, including single-cell equivalent circuit formulation and parameterisation, electronic circuit design, control logic, and automatic code generation. To complete the end-to-end range, they also facilitate verification and validation by running real-time simulations for hardware-in-the-loop testing. Based on this comprehensive range, MathWorks customer NXP, a leading semiconductor supplier, has built a set of toolboxes with BMS design samples that customers can use to simply simulate, generate code from MATLAB, Simulink and other MathWorks tools and focus on NXP processors. For the Dutch automotive start-up Lightyear, which has developed a solar-powered electric car with rooftop solar panels, MathWorks software was used to develop a BMS that, among other things, regulates the solar energy to the battery. The Belgian start-up Ellio used the same possibilities to develop a two-wheel drive e-bike. The extensive modelling of the electric motor control system took into account external factors such as aerodynamic drag and road gradient.
In the current digitisation trend, MathWorks has added artificial intelligence (AI), digital twins, and cloud reference architectures to Model-Based Design. This also applies to the field of BMS development, says Anas. ‘In a battery, many processes take place simultaneously and their effects on performance and degradation are difficult to understand from first principles. This is where measurements from a real battery can be used to train an AI algorithm that predicts states. In this way, the AI algorithm is a promising means of filling the gaps in knowledge’, explain Anas and Gazzarri. ‘A good example of this is understanding how operating and environmental conditions influence the rate of ageing.’
How to get started
So there is much to be gained from using Model-Based Design in BMS development. ‘When we talk to our customers in the Benelux, we can clearly see that they are looking for ways to keep up with the electrification trend’, Stakenborg and Anas conclude. ‘At the same time, they realise that modelling and simulation are critical to them. That’s why we’ve planned an event in the fourth quarter of this year, entitled Accelerating Electrification with Model-Based Design.’
MathWorks is a leading developer of data analysis and Model-Based Design software. MathWorks supports digital engineering through its platforms, the best known of which are MATLAB®, Simulink® and Simscape™. They form the basis for building a digital twin, which MathWorks defines as an ‘evolving digital representation of a product, process or system that calculates its status and generates information’. This way, through integrated and systematic use of data and models, decision-making for design and operation can be supported throughout the life cycle.