Battery Analysis and Cell Balancing Monitoring Feature Update



Limitations of the Previous Degradation Prediction Method and the Direction of Improvement

The default degradation model in Dr.EV was originally developed with the goal of accurately reflecting the initial rapid degradation phase, aiming to represent the physical characteristics of real-world batteries as closely as possible.

However, we received frequent feedback that this approach felt overly conservative compared to users' actual experiences. To address this, we have improved the algorithm so that early-stage degradation is expressed more moderately.

In particular, for newer vehicles, using the previous model as the basis for estimating SOH (State of Health) or RUL (Remaining Useful Lifetime) often resulted in unexpectedly fast degradation predictions, reducing the reliability of long-term forecasts.

To resolve this issue, we adjusted the early degradation zone, resulting in a model that offers more trustworthy long-term predictions.

Improvement of Initial Capacity (BOL) Settings

Previously, the initial capacity (BOL: Beginning of Life) was defined based on the rated capacity, in an effort to account for pack tolerances and initial conditions when calculating SOH.

However, since actual vehicle BMSs also consider manufacturing tolerances, using a fixed rated capacity often caused the SOH to appear lower than it should be.

To address this, we adjusted the initial capacity values for some vehicles, resulting in more reasonable and accurate SOH values.

Additionally, during our internal review, we found inconsistencies in BOL values across some vehicles and manually corrected these discrepancies.

※ If you believe there is an issue with your vehicle's initial capacity setting, feel free to contact us anytime at info@battermachine.com.

Loss of Variability Expression and Fun Element

Due to these updates, one of the enjoyable aspects we initially intended—namely, expressing cell- and pack-level manufacturing tolerances and capacity variation caused by SEI layer formation—can no longer be visualized as originally planned.

Reference: Wang, A., Kadam, S., Li, H., Shi, S. & Qi, Y. Review on modeling of the anode solid electrolyte interphase (SEI) for lithium-ion batteries. npj Comput Mater 4, 15 (2018).



Manufacturers tend to be conservative about showing variability, because managing production tolerances is considered a key aspect of their engineering capabilities.

However, at Dr.EV, we believe that such variability can actually be an interesting and engaging element for users.

Just like a kind of "luck of the draw", when a vehicle's actual initial capacity exceeds the BOL (Beginning of Life) that we have set, we consider it a “lucky pack”.

We're currently considering adding a feature to display a congratulatory message within the app when this happens. That said, because this kind of assessment is only meaningful for new vehicles, this feature would likely only be available to users with recently manufactured cars.

Below is an explanation of how Dr.EV interprets “default” and “adjusted” degradation values.

Some users mistakenly believe that battery degradation doesn't occur if charging and discharging do not take place. In reality, natural degradation progresses even when the battery is not in use.

If a battery is continuously maintained at 50% charge and the temperature is kept below 25°C, then natural degradation would be extremely slow. However, in practice, it’s quite difficult to maintain such ideal conditions. Most vehicles are left either fully charged, in a high state of charge, or sometimes even deeply discharged.

In particular, leaving the battery fully charged during hot summer temperatures can greatly accelerate degradation. Similarly, deep discharge also hastens battery wear.

Therefore, even if a vehicle hasn't been driven much, improper storage conditions or poor battery management can result in significantly more degradation than expected for that mileage.



1. Default Algorithm (Standard Display Value)

This algorithm reflects real-world degradation trends based on actual measured data.

Once a sufficient amount of charging history is accumulated, it provides the most accurate estimation of actual battery degradation.

It particularly incorporates the typical characteristics of lithium-ion batteries, which show rapid initial degradation followed by a gradual decline.

2. Alternative Algorithm (Trend-Based Value)

For the alternative algorithm, Dr.EV applies a linear degradation model at the early stage, resulting in a smoother and more gradual decline curve.

Currently, all analysis and metric calculations in Dr.EV are based on this alternative algorithm for consistency across the platform.

🔧 Cell Balancing Feature Update

For our members who actively use the cell balancing feature,we have expanded the cell balancing monitoring levels from 3 to 5 stages.

This allows for more precise condition monitoring, and when the monitoring result reaches level 5, Dr.EV considers the corresponding cell to be at a "defective level."

We will continue to improve our system to provide more accurate and reliable battery diagnostics moving forward.