How the estimate works
Each estimate runs an electrochemical battery model built with PyBaMM, an open-source battery modelling library. It simulates your first year of driving in full, then projects it forward.
Your day becomes a test cycle
Daily distance, vehicle efficiency and charger power turn into one day of use: driving, parking, then charging back to your chosen level. Everything is scaled to pack size, so the same model works for a hatchback or a city bus.
A full year is simulated
The model runs 365 of those days, with the temperature changing month by month for your climate. If the battery is heated and cooled, it stays between 15 and 30 °C while driving and charging, but sits at the outside temperature while parked.
Two kinds of wear are tracked
A surface film (the SEI) slowly grows on the graphite electrode and traps lithium. It grows faster in heat and at a high state of charge. Fast or cold charging can also plate metallic lithium onto the electrode, and part of it is lost for good.
The trend is projected forward
The lithium lost in the first year is fitted with a square-root-of-time term, which is typical of film growth, plus a straight-line term for wear from each cycle. That fit is extended to later years and scaled by a calibration factor for wear the model leaves out, such as electrode cracking.
Limitations
- The parameters come from one NMC811 cylindrical cell (OKane2022). LFP batteries usually age differently.
- Manufacturer buffers and battery-management software aren't modelled.
- Absolute numbers are still being calibrated against real fleet data, so they lean optimistic.