Battery Report - Apply Filters to Own Data
The Battery Degradation Report has a good idea of including outside temperature when determining which fleet values to compare to; but this encounters a quirk with cold climates and heated garages. Namely, the conditions when driving (and thus HVAC usage and other factors) can be wildly different from the conditions when charging; which causes the vehicle to be mapped against the incorrect fleet data set. Drive conditions have also been shown to affect estimated and rated ranges - with the chart effectively becoming a graph of temperature changes over the life of the vehicle.
For example, yesterday my average temperature while driving was 14F, and so the vehicle is reporting 221 rated miles at 90%. My garage, however, is heated, and as a result my charge when I got home is being recorded and compared to other charges at 41F. This results in a "range difference" versus the fleet of nearly 40 miles. During summer months, the difference is less than 10 miles.
Similarly, looking at just my vehicle's values, all 3 options (High vs Low, High vs Current, and Starting vs Current) show a similar skew. I.e. my "low" and "current" ranges right now are being based on today's frigid range, and comparing to my much more mild 60F "high" range.
Ideally, fleet data would be filtered by average drive temperature since the last charge, not ambient charge temperature. That might be difficult or time consuming with historical data though.
So, my proposal would be to add a "Filtered High Range vs Current Range" option under the trendline options. This would effectively work the same as "High Range vs Current Range", but with the "High Range" having the "completed charges over %" and "Only within xx degrees" filters applied to the vehicle's historical charges. This isn't quite ideal as it would still allow below-garage-temperature drives to blend together, but would at least cancel out the largest skew caused by summer vs winter weather.
Customer support service by UserEcho
Appears the "weather-based maximum range" was related to the BMS being wildly out of calibration. Basically, between data logging and Sentry Mode, the vehicle was only ever sleeping at 90% after a completed charge. Over time, the calibration drifted and the estimated range was just mapping weather patterns (low range in the winter, higher range in the summer).
Setting up schedules to give it 3 sleeps a day, with varying charge levels both throughout the day and throughout the week seems to have resolved it after a few weeks. My schedules start the week off with a maximum 90% charge Monday, and finish it with a 50-60% maximum charge Friday. I also set Sentry Mode schedules to sleep for 3.5-4 hours a day at work (during a slow time in the parking lot).
I use roughly 40% per day. I.e. Monday starts at 90% after waking up from a nightly sleep, arrives at work at ~70% and sleeps for ~3.5 hours, and arrives home at 50% and sleeps for 3-5 hours before charging starts.
That has me wondering if it makes sense to add a note to inform users that might be happening if their trends look unusual versus the fleet; or perhaps if a certain percentage of charges have been at the same charge level over a given time period?
40.5k miles is when I started using the staggered charge limit schedule and allowing it to sleep for 3-4 hours at work and before charging at home.