The relation between reported consumption, remaining battery energy (% State of Charge, SoC), battery size, and eventually battery health (State of Health, SoH) may be more complex than what is initially apparent. For instance, a simple task such as reconciling reported consumption to how much SoC has been lost on a trip ends up requiring a good understanding of how much capacity is available to normal driving and in turn how much is set aside to buffers.

This write up is based on my own experience trying to understanding the battery state and health in our Skoda Enyaq iV 80x (pictured). It may also be useful for others with a similar quest.

Methods and Data

Data reported by the car at the dashboard(s) typically contain:

  • Current State of Charge (SoC) is the available battery energy for normal driving, as a percentage between 0% and 100%. This is often referred to as the “SoC Display”.
  • The average consumption, either as kWh/100 km or as km/kWh for some period of driving, typically since last charge or from some other defined point in time. The MEB platform reports kWh/100 km with 3 digits (i.e. 14.2 kWh/100 km) and km/kWh with 2 digits (i.e. 7.0 km/kWh). For this reason it may be preferable to have consumption reported as energy over distance, i.e. kWh/100 km. In my experience the consumption figures are quite accurate.
  • The range (mileage) covered for the same period(s) of driving.
  • The remaining range (mileage), which is based on your past 500 km driving pattern. This comes from what is sometimes called the Guess-O-Meter (GOM).

In addition, a Bluetooth dongle can be used to retrieve more detailed data directly from the OBD-II/EOBD interface in the car. I am using the Vgate iCar Pro “scanner” that can connect either using Bluetooth Classic or Bluetooth BLE (i.e. it works with both Android and iOS) with firmware version 4.1.10. Data is then read by the Car Scanner ELM OBD2 Android app that allows for presenting (and logging) in different ways.

Some of the additional key data points relevant here are at least the following:

  • SoC BMS is the State of Charge in % reported by the Battery Management System (BMS). It is different from the SoC Display (that is normally shown to the user) as it include buffers both at low and at high charge in addition to the energy available for normal driving and visible as the SoC Display.
  • BMS Energy Content or HV Battery Energy Content (EC) is the energy “useful for normal driving” (excluding buffers). On the MEB platform, the scale is offset to report 0 kWh when the SoC Display is also at 0% (even if SoC BMS % is then typically at ~5.75% but holds a buffer not directly visible to the user). Hence the readings can be negative even though the battery still holds (reserve) and “useable” energy in the buffer.
  • Maximum energy content of the traction battery (MEC) is an estimate of the useful energy available to the BMS, including some parts of the BMS buffers. It is not fully documented how this is derived; one claim has been that it is the SoC Display 0-100% capacity plus bottom buffer. My guess is that it is the full BMS range minus the unusable part of the bottom buffer (i.e. the range where only a crawl attempt is allowed), i.e. from around 3% to 100% SoC BMS. Nominally it should be equal to the net battery capacity (77 kWh).
  • Total accumulated charge is reported both in Ah and kWh. This is the accumulated total energy in to the battery, including recuperation(!), so it’s not all coming from your charging sessions.
  • Total accumulated discharge is also reported both in Ah and kWh. It is similarly the accumulated total energy drained from the battery.
  • In addition, the DC Battery voltage as well as Cell voltage and SoC for each individual cell, and a number of HV Battery temp and Battery inlet/outlet temp points are available to assess the overall condition of the battery – as well as a long list of other interesting data points including the estimated consumption for 7 different road categories, maximum charge voltage and the dynamic limit for charging that allow for estimating the current max charging effect etc.

Unlike a fossil car’s remaining fuel content, many of the battery metrics cannot be directly assessed, but are estimated from other metrics such as temperature, voltage, current and charge count, that all can be measured rather precisely, and perhaps also sometimes the past history of events.

Battery capacity and buffers

EVs are typically advertised with a gross battery capacity (mine is 82 kWh) and a net capacity (also called MEC; mine is nominally 77 kWh). The difference (here 5 kWh) is set aside for what could be called engineering buffers. These may partly account for manufacturing tolerances, to extend battery life, to prevent a total discharge (or even a cell-reversals), etc. How the 5 kWh is divided between low and high charge states is not publicly known; I’m showing it like 2.5 kWh in the bottom and top for illustration purposes as it fits well with the data available and at least 2.5 kWh needs to be available in the bottom. But it could also have been more in the bottom and more in the top.

Only the net capacity is “used” by the BMS (post, thread); the engineering buffers are not useable for (normal) driving, possibly apart from small amounts of “turtle mode” driving.

The role of the BMS is to protect the battery and make sure it works optimally. As part of this, an additional set of buffers are set aside. This is done by mapping the SoC Display so that only a range of the SoC BMS is seen by (available to) the user. In my case, 0% SoC Display corresponds to ~5.75% Soc BMS and 100% SoC Display corresponds to ~96.0% SoC BMS. This means ~90.25% of the allocated BMS capacity (~71.7 kWh nominal) is available for normal driving and the rest are BMS (and partially engineering) buffers.

Note that the relation between SoC BMS and SoC Display is very linear. On my car, the Pearson $R^2$ coefficient is above 99.99%, and similarly between EC and the SoC BMS as well as the SoC Display values.

The (4%) BMS buffer at the top probably further helps to protect the battery and would allow you to use recuperation even with a completely charged battery and perhaps also to take care of any temperature swings between charging and using the energy without a risk of overcharging.

The (5.75%) BMS buffer at the low end may likewise help to protect the battery, and is also what makes that car able to drive even when you reach 0% SoC Display. Some YouTubers excel at driving cars all the way down to become almost non-responsive; I would never want to do that on purpose, but am glad someone tested it so that the rest of us knows what to expect… A couple of examples are e.g. Bjørn Nyland testing an ID.Buzz and Rundt om Biler testing a Tesla Model 3.

On the vwIDTalk forum, ID.Furkan reported that his MEB based ID.4 car can be driven normally until SoC BMS is at 4% (showing -1.75% Soc Display) although it will say “CHARGE NOW!” in the display. Down to SoC BMS 3% it will beep frequently and at 3% SoC BMS (or around -2.5% SoC Display) it will stop and you then have two attempts to crawl the car really slowly, down to SoC BMS 2.5% – and then the car will no longer drive. So when you reach 0% SoC Display, you still have around 2.75% “useable” energy, which may give you anywhere between 6 and 12 km of additional range. Apparently the car will be completely “dead” at 2% SoC BMS. Note that SoC BMS 3% (where normal driving is no longer possible) is very close to 2.5 kWh, i.e. the size of the engineering buffer. Also note that the car reports a value called NV Energy Requirement that is, in my case, 2.55 kWh. It may be an estimate of the minimum energy needed to run normal operations in the car and if less than this available it will not run anywhere…

Despite not using the engineering buffers actively, the BMS system nevertheless appear to report its SoC BMS starting from a truly zero state and thus include the low engineering buffer within its reported range, i.e. SoC BMS is 0% at zero energy content (below the low engineering buffer). The top engineering buffer, however, is outside of the 0-100% range. It sounds a bit strange to me, but this is the way the available evidence and numbers reported are best explained.

This is all illustrated below.

The illustration above shows the energy available for normal driving (71.7 kWh, corresponding to 0% to 100% SoC Display) in green, as well as the Engineering buffers in red and the additional BMS buffers in yellow. The blue scale shows how the the SoC BMS is scaled from 0% at “the absolute 0” charge to 100% at 79.5 kWh above 0. The 77 kWh net energy is also referred to as the MEC, and excludes the engineering buffers completely. These numbers apply to the VW MEB platform used in e.g. Skoda Enyaq 80x model years 2021-2023 when using LG batteries (at least). VW ID.4 2023 and newer appear to have a somewhat different structure.

It is arguable whether all of marketed net capacity really is “available” to the user when it’s not useful in “normal” driving. The counter argument is that it’s useful in special conditions (recuperation at 100% SoC Display, i.e. if you charge fully and start driving downhill, or being able to maneuver then car when at 0% SoC Display), so I guess it can be claimed that it’s somehow “available” and can benefit the user.

Capacity Loss

The full battery capacity is not always available to the user.

Temperature impacts the chemical processes, and in turn also the amount of available energy. The relations are a bit complex (see e.g. the concrete example), but as a rule-of-thumb, the colder it gets, the less energy can be delivered by the battery. Higher temperatures can also cause the performance to drop, including a higher self-discharge rate and a shorter overall battery life. There is more info and background here, e.g.: Digging deeped into how temperature and speed impact EV range, Real Range for Electric Cars by Temperature and Weather, and How does ambient temperature affect EV batteries?

When the battery is cold, if can be heated by the car (using energy from the battery itself or from the power line, if the car is connected before driving). This should not happen too fast to avoid irreversible changes so even if some range can be recovered in cold weather by heating the battery, it comes with a cost in terms of lost energy. In addition, the range is also impacted in cold weather by higher air density or due to increased loss to wet road surfaces and higher use of heating.

Often the optimal battery range is reported to be 15-35 degree celcius (20-40), (15-45), with loss of capacity due to slower electrochemical reactions below 15 degree and faster degradation above 35 degree. See also Pesaran, A et al (2013): “Tools for Designing Thermal Management of Batteries in Electric Drive Vehicles”.

In addition, the battery will degrade over time. How much is left of the original designed-for capacity is referred to as the State of Health (SoH) and is specified in %. There are many attempts at testing and documenting this on the net, including this crowd-sources spreadsheet (and a corresponding German spreadsheet) that shows SoH over time and range for ID.4 cars.

Based on available info on the net and talking to VW technicians, apparently there is often an initial loss of capacity of 2-3% during the first months of use, followed by a slower degradation of 1-2% for a well maintained battery pr. year. So for e.g. the ID.4 cars, the model seems to indicate that after one year, the SoH will be around 95% and after 3 years it will be around 92.5%. However, these vary significantly with use and how you charge:

The SoH for a 215 cars based on self-reports by their users.

General good advices seems to be keeping the battery in the range of 20%-80% for 80% of the time (or even 70%-30%), only using fast DC charging when really needed, avoiding high drains (as in fast accelerations), drive shortly after a charge to more then 80% when needed, stay closer to the 40%-60% range for long term parking, etc. See also VW’s recommendations. If you want a really long lasting battery, NASA charging (perhaps 40%-60%) is reported to exceed 14.000 cycles before getting to a 30% degradation but the original source is difficult to find.

Charging Efficiency

When charging the car, there is a loss incurred in the conversion process. For AC charging, this happens inside the car, so not all of the energy delivered by the network is stored in the battery. In my car, the amount of energy reported by the BMS to be stored in the battery is 90.0% of the amount of energy delivered by the network, at least in temperatures around 10-20 degree celcius. This loss comes from a loss in the converter and running other electronics and pumps (in my case estimated to be around 92.5% efficiency) and from loss in the battery itself when being charged (in my case the efficiency is around 97.3%). These numbers are based on the energy delivered to the car by the charger, the energy attempted to be stored to the battery and the net increase in energy content in the battery.

For DC charging the amount of energy stored is around 90.6% of the delivered energy on average, but seems to vary some depending on the charger and charging state (I’ve seen numbers between 87.2% and 92.3%).

These numbers may be updated as I get better data.

Reported consumption

An often discussed number is the WLTP consumption, which includes charging losses as it is based on the energy supplied during the recharge that follows the consumption test.

This is often contrasted to the average consumption reported by the car. However, the car reports the net consumption (drain from the battery), excluding charging losses so it will typically be lower. Also, as the available energy for normal driving is not the full BMS MEC, and it could lead to an incorrect assumption that the battery has deteriorated more than it really is, or that the consumption reported is too low.

In my experience, the reported consumption by the car is quite accurate and fits well with the change in energy content in the battery. As an example, for a 282 km long drive the reported average energy use was 15.6 kWh/100 km. The energy used is therefore 44.0 kWh. The SoC Display changed from 80% to 16%, and with a nominal available energy of 71.7 kWh and a measured SoH around 95.5%, this amounts to 43.8 kWh, very well within the given precision of the measurements (the difference is only 0.5%).

In my car the WLTP consumption is listed as 17.0 kWh/100 km. This corresponds to approximately 15.3 kWh/100 km in “real” drain of the stored energy, and is also in line with what I see for conditions similar to the WLTP charging.

Interestingly the charging loss is approximately equal to the ratio of the available capacity for normal driving to the net capacity (71.7 kWh vs 77 kWh, or 93%) which can lead to some confusion as you can do an incorrect calculation and end up with almost correct numbers:

  • If you would fully charge a “nominal” car you would likely have to add around 79.7 kWh to fill up the battery at 90% charging efficiency, and the range would then be 468 km at the WLTP consumption of 17 kWh/100 km. This is correct.
  • If you look at the actual reported consumption from the battery, the range should be calculated from 71.7 kWh at 15.3 kWh/100 km, which ends up being also 468 km. This is the same figure, and also correct.
  • However, I even seen some (incorrectly) calculate range from the net capacity of 77 kWh and using the WLTP figures, and you would end up with 452 km which is close to the correct value but based on incorrect assumptions.
  • I’ve also seen others incorrectly report range based on net capacity and reported consumption, which then will overestimate the real range; in this case it would incorrectly lead to a range of 503 km.

Finally, the consumption is sum of the drain from the battery. When you brake using recuperation, you store some of the energy back into the battery (although perhaps with a small loss).

As an example, at battery content of 25.15 kWh the accumulated charge was 4414.12 kWh and the accumulated discharge was -4172.53 kWh in my car. After some additional driving, the energy content was down to 2.55 kWh and the accumulate charge was 4419.41 and the accumulated discharge was -4200.64 kWh.

The difference between the start and end stored energy was thus 22.6 kWh (25.15-2.55 kWh).

The total energy drain was 28.11 kWh (4200.64-4172.53 kWh) but the amount of energy re-charged in recuperation was then 5.29 kWh (4419.41-4414.12 kWh); together the loss of stored energy is then 22.82 kWh (28.11-5.29 kWh), which is almost identical to the difference between start and end stored energy of 22.6 kWh. Note that 18.8% of the used energy was recovered, so recuperation had a significant impact on my range.

Calculating State of Health

There is also sometimes confusion on how to calculate the SoH based on charging, reported consumption, loss of range, etc. Getting access to a BlueTooth dongle will give you better insights, so let’s walk through a simple example here to align all the numbers.

Often to estimate a cars SoH, it is recommended to discharge below 10% and to re-charge above 90%. This is due to the shape of the charge/discharge curve that become more steep below 10% and above 90%, giving the BMS a better ability to evaluate the battery state more precisely.

Having done this, one ID.4 owner reported a MEC of 73.35 kWh and a SoC Display of 90.72% with a corresponding EC at 62.1 kWh.

As the nominal MEC is 77.0 kWh, it means the car would be around 95.3% SoH (i.e. it has lost 4.7% battery capacity.

Taking the EC at 62.1 kWh and dividing by 90.72% you get an estimated EC of 68.45 kWh at 100% SoC Display. This corresponds to 96% SoC BMS. However, as 5.75% SoC BMS corresponds to 0 kWh EC it means the “real” energy content from 0 to 100% SoC BMS would be 68.45 kWh divided by 90.25% (96% - 5.75%), or 75.85 kWh. If the nominal value of the full battery at 100% would be 79.5 kWh, the SoH would then be around 95.4% (nearly the same figure).

The difference between the reported MEC and the estimated EC at 100% SoC BMS is, for this car, 2.50 kWh (or a nominal 2.62 kWh compensating for the SoH). Based on analysing data from 10 cars where the owners have reported both their EC at a high (and reported) SoC Display as well as the MEC, it appears the average nominal difference would be 2.5 kWh (S.D. 0.6 kWh) – which happens to be very close to the reported value of the NV Energy Requirement for some reason (perhaps it’s the size of the low “unusable” engineering buffer).

Finally, the nominal drivable range is then 90.25% of 79.5 kWh, or 71.7 kWh. In the case above, with an EC of 68.45 kWh at 100% SoC Display, the SoH based on comparing these numbers would be 95.4%, also very well in line with the previous estimates as it should be.

This all fits with the assumption that the low engineering buffer is around 2.5 kWh and is included in the SoC BMS numbers and that the upper buffer is on top of the 100% SoC BMS (i.e. up to 103.1%)

GOM Estimates

The GOM is not useful for calculating past consumption, but is an estimate of how far you can drive in the current conditions.

Skoda report that you need to drive at least 500 km to have an estimate that corresponds to your personal driving style. They also indicate that they tend to weigh the last 100 km driving in normal conditions, but as you get lower charge, it starts to use the last 10 km more, to e.g. help you in the typical scenario of trying to get to a charger and where your driving maybe changes to be more power efficient…

Comments

I don’t have any official comments section on this blog, but feel free to reach out to me via e.g. mail with comments or corrections.