Battery

Most AI assistants are generalists. Ask them about Butler-Volmer kinetics, the difference between SEI growth mechanisms in NMC versus LFP, or how to interpret a GITT pulse sequence - and they will give you something plausible, occasionally correct, and impossible to trust without cross-checking. That is a fundamental problem for engineering work. A tool you cannot trust at the domain boundary is a liability, not an asset. The Battery Expert AI was built to close that gap - a domain-specific assistant that understands battery electrochemistry deeply enough to be genuinely useful, and runs entirely on-premise so that proprietary cell data and OEM specifications never leave the building. ...

Most battery models fail not because the equations are wrong - but because the assumptions smuggled in with them are wrong. Fifteen years working with battery models across the stack: from ECM spreadsheets that needed to survive a BMS real-time loop, to full electrochemical models trying to capture what actually happens inside the cell during a fast charge. The gap between “runs on paper” and “works in a vehicle” is where most of the interesting problems live. ...

GitHub Repository CFD-based battery simulation is powerful and routinely impractical. OpenFOAM can model the coupled thermal-flow behavior of a battery pack in detail that lumped models can’t touch. It can resolve cell-to-cell temperature gradients, coolant channel flow distribution, and transient heat propagation through complex pack geometries. It can also take an experienced CFD engineer a day to set up a single case, another day to run it, and significant effort to extract anything useful from the output. ...

Most battery simulation work happens inside large OEM simulation frameworks - tools with hundreds of subsystems, long setup times, and infrastructure built for full-vehicle analysis. Useful when you need the full vehicle. Counterproductive when you need to iterate quickly on battery-specific questions. At Volvo Trucks, the problem was clear: the battery and BMS teams needed a simulation environment they could actually use - fast iteration, battery-focused, not dragging a full vehicle model along for every run. ...

The EV marketing spec says “0 to 80% in 22 minutes.” It rarely says how long 80 to 100% takes. The omission is deliberate - and it reveals something fundamental about how lithium-ion cells actually work. The CC-CV Profile and Why the Transition Matters Lithium-ion charging follows a two-phase profile, not because someone decided it was a good idea, but because cell physics demands it. Constant Current (CC) phase: You push current in at a fixed rate. The cell voltage rises as SoC increases. This phase is efficient - you’re putting energy into the cell at the maximum rate the chemistry allows without exceeding the voltage limit or driving electrochemical side reactions. ...

The Situation Nobody Talks About Here is a situation that happens more often than you would think in battery development. An EV company has a working cell from Supplier A - tested, characterised, reliable. But for their next programme, they need a different cell: larger capacity, different form factor, different thermal profile. They ask the supplier: “Can you make us a 21700 version with 4000 mAh from the same chemistry?” ...

A battery’s safe operating envelope is not a fixed number. It’s a surface - varying continuously with temperature, state of charge, state of health, and operating history. Getting it wrong in the conservative direction costs performance and range. Getting it wrong in the aggressive direction costs cell life, and at the extreme, safety. The Current Limits Generator was built to define that surface correctly - from physics, not from conservative blanket rules. ...

The Quiet Problem Inside Every Battery If you own a smartphone or an electric vehicle, you have experienced battery degradation. That slow, frustrating decline where a full charge used to last all day and now barely makes it to dinner. The science behind this is actually fascinating, even if the experience is annoying. Inside every lithium-ion battery, a protective layer called the Solid Electrolyte Interphase (SEI) forms during the first charge. This layer is essential - without it, the battery would destroy itself immediately. But here is the cruel part: the SEI never stops growing. It creeps thicker over time, consuming active lithium that could otherwise be used to power your device. ...

Most thermal modelling projects in automotive spend the majority of their time not solving physics - but configuring it. Setting up geometry representations, mapping cell sensor signals to Simulink ports, assigning material properties, wiring the cooling boundary conditions from CFD into the thermal solver. The same steps, repeated for every new battery variant, every new pack geometry, every new cooling configuration. The Battery Thermal Model Configurator was built to eliminate that overhead - and to produce something better than what the manual process was generating. ...

What This Was This wasn’t a project with a start date and a delivery date. It was three to four years of incremental work on a living codebase that had to keep working while being extended. Mercedes-Benz R&D, in collaboration with other European OEMs, built an in-house full-vehicle simulation framework on top of CarSim — initially scoped for IC engine vehicles. When I joined in 2015, the task was to extend that framework for EV, HEV, and PHEV architectures, improve the existing model libraries, and build the component models that didn’t yet exist. ...

An EV thermal management system is not one problem. It’s eight problems that happen to share coolant. The battery wants to stay between 15 and 35°C. The power electronics want active cooling at high load. The cabin wants heat in winter and cooling in summer, and it’s competing for the same refrigerant circuit the battery uses. The motor generates heat during aggressive driving. The DC-DC converter has its own overtemperature failure mode. All of these are connected - through coolant loops, refrigerant circuits, and control logic - and all of them interact dynamically. ...

After spending years working in battery simulation across Mercedes-Benz, Caterpillar, Volvo Trucks and later A123 Systems, I slowly realised something that was not obvious to me in the beginning. Most battery engineers think they are building battery models. In reality, they are usually building battery infrastructure. The distinction sounds subtle. It isn’t. A battery model answers a question. A battery simulation framework enables an organisation to ask thousands of questions. Those are completely different engineering problems. ...

After spending years working in battery simulation across Mercedes-Benz, Caterpillar, Volvo Trucks and later A123 Systems, I slowly realised something that was not obvious to me in the beginning. Most battery engineers think they are building battery models. In reality, they are usually building battery infrastructure. The distinction sounds subtle. It isn’t. A battery model answers a question. A battery simulation framework enables an organisation to ask thousands of questions. Those are completely different engineering problems. ...