EV Charger Power

Aggregating information from multiple charging sessions and potentially multiple vehicles to determine the charging power behavior for EV chargers

As discussed in batteryCapacity, most vehicles do not report batteryCapacity so the battery capacities for specific make, model, year, and trim for some specific Teslas are given in this file:

8KB
battery_capacity_for_some_teslas.csv
Open
File containing manually researched battery capacity for Teslas

The code to calculate full charging sessions in the Code Appendix can be used to generate information about charging sessions. For example, an example level three charging session is given here

bat_capacity = 75
counties = gpd.read_file('tl_2022_us_county.zip')
states = gpd.read_file('tl_2022_us_state.zip')
urban = gpd.read_file('tl_2022_us_uac10.zip')

out_full = calculate_charging_sessions(df, urban, states, counties, bat_capacity)
Example charging session

The charging session data can be aggregated based on the locations as in the below code snippet. Note that the charging sessions are aggregated based on the first two digits of the latitude and longitude of the GPS locations at which the charge started.

Using two digits can mean that some charging locations are combined into a single coordinate pair, but using three digits sometimes means a single charger is split into multiple coordinate pairs. While three decimal places generally gives about 111 meter accuracy (at the equator), the error in the gps signal itself (i.e. gps drift) means the location itself is imperfect. Two decimal places generally gives about 1110 meter accuracy, which is fairly large, but typically public charging stations will not be this close together.

session_df contains the concatenated results of calculate_charging_sessions for multiple vehicles.

At the end gb contains:

  1. rounded_point_str: the location for these charges

  2. kw_est_mean: the average estimated rate of charge (in kilowatts) based on the change in state of charge and the battery capacity (in kilowatt-hours)

  3. kw_first_mean: the average first charger_power observation for charging sessions at this location

  4. kw_last_mean: the average last charger_power observation for charging sessions at this location

  5. subject_count: the number of charges to occur at this location

  6. subject_nunique: the number of unique vehicles that charged at this location

The following code will combine the list of soc values and charger_power values to plot the charger_power versus soc for all observed values of soc

And here is an example graph where it can be seen that the charger can delivery about 150 kilowatts, but the vehicle's battery management system will decrease the power as the soc approaches 100%.

Some aggregated charging session data can be found on this map: https://felt.com/map/Charger-Power-Data-NCYlp5WxRVmW0NQ9CVrHO7A?loc=46.12,-73.74,3z&share=1 .

Additionally, to be able to compare power delivered during different phases of charging sessions across many vehicles when charger_power is not available, charging sub-sessions can be calculated as well.

Code Appendix

PreviousEV Compatibility Score - Level 3 ChargersNextVehicle Depreciation

Last updated