Well, the reasons why we sometimes see the displayed SOC drop after a longer drive (or increase) is comming from the inability to correctly measure the SOC during the drive.
-The BMS having an overestimate of the capacity leads to overestimated SOC, and a drop. And vice versa for a underestimate.
-The BMS having an overestimate of the capacity leads to overshoot in the charging, and vive versa.
This is a drive I have posted about before seberal times, A 513km drive on one charge to find the real capacity.
Ypu can see parking 10 a’clock at 52% (51.9% according to the BMS).
This drive was a ~90 kph drive on autopilot to enable the possibility to drive back home without charging. No stops, but a rather constant load for 3 hours.
After only 30 minutes the SOC was updated to 53% (I did not stay at the car so I have no BMS readout for that).
After the car slept for ~ 7 hours, the car displayed 54% SOC (53.9% in the BMS)
During this time my BMS had a gross underestimation of the capacity.
Doing the math of the estimated capacity (which leads to) - nominal remaining will show that the resulting estimated SOC when parking is a exact reflection of the portion of the nominal remaining minus the delivered energy.
Nominal full pack was 77.5kWh, so 74.0kWh usable. The theoretical used energy is 48.05%, rendering the SOC at parking to be 51.95%.
The resulting real capacity after the 100-0% drive was done (also equals the capacity estimate from about one month later when the BMS had adjusted the estimated capacity back) matches the SOC that the car showed after the sleep.
So, by knowing the capacity before this it was actually possible to know that the real SOC after the drive would be around 54% instead of 52%.
When I got my Plaid, and mounted the SMT it showed 95.7 kWh capacity out of 99.4 kWh full pack when new.
To test the capacity I noted SOC before and after the first couple of long drives.
The change in SOC reflected a capacity of 98-98.3kWh capacity, which the BMS ended up showing 1-2months later.
I also found that the end state SOC always reflected the used soc / estimated capacity, and that it would recocer SOC after each longer drive.
During the time when the BMS recovered the recovery got smaller and at the time when the BMS estimate was correct, the estimated SOC was correct and there was no change to the SOC after the drive.
