Temperature is the real issue, and kills LiFePO4 faster than anything else. You can be at 100% SoC and 0C and you will have far, far less impact than being at 50% SoC at 30C for the same amount of time. That's why I mention that SoC isn't a huge deal compared to temperature. Some datapoints from the paper:
Operating condition | Temperature (°C) | Capacity loss (%) | Resistance increase (%) |
---|
100% SOC, Calendar for 8 months | 30 | 1 | 9 |
100% SOC, Calendar for 8 months | 45 | 7 | 41 |
100% SOC, Calendar for 5 months | 60 | 26 | 58 |
30% SOC, Calendar for 5 months | 60 | 15 | 26 |
So at 30% SoC and 60C you see a 15% loss after just 5 months, compared to only 1% at 100%SoC and 30C after 8 months.
Similarly, another datapoint:
Operating condition | Temperature (°C) | Capacity loss (%) | Resistance increase (%) |
---|
100% SOC, Calendar for 12 months | 25 | 5 | 2 |
50% SOC, Calendar for 12 months | 25 | 3 | 5 |
100% SOC, Calendar for 12 months | 45 | 15 | 20 |
Just two percent difference for 100% vs 50% at 25C over 12 months. And that's assumed to have the battery sit at that temperature and state of charge for an entire year, not cycling at all.
This is very clear in this data point; the capacity loss between 100% and 50% over an entire year is practically negligible:
Operating condition | Temperature (°C) | Capacity loss (%) | Resistance increase (%) |
---|
100% SOC, Calendar for 12 months | 25 | 5 | 2 |
50% SOC, Calendar for 12 months | 25 | 3 | 5 |
100% SOC, Calendar for 12 months | 45 | 15 | 20 |
50% SOC, Calendar for 12 months | 45 | 12 | 21 |
So I still stand by my assessment that state of charge doesn't really matter if you allow the cell to settle and not keep it at a high voltage when the cell is in regular use. Temperature is the biggest thing to keep an eye on.