
A new V2GSim case study demonstrates the ability of V2GSim to predict EV charging loads in a stochastic manner. The case study demonstrates how users can provide statistical inputs to V2GSim to describe how drivers use their cars to predict EV charging loads for any number of vehicles. In the case study, two scenarios are demonstrated for a collection of 1000 EVs. In the first case, grid loads are predicted for a case where vehicles charge predominantly in the evenings. In the second case, grid loads are predicted when charging is allowed during the day and evening.
All results are resolved on a secondbysecond basis enabling users to see the up and down spikes in grid loads that arise from individual vehicles plugging in and unplugging. With this statistical inputs approach, 
V2GSim predicts both the most likely charging load, but also provides an uncertainty estimate on grid loads arising from individual drivers taking unexpected trips, or taking longer time to complete an individual trip. In order for users to calibrate input statistics for a given region or fleet, automated methods have been built into V2GSim to derive input statistics from a deterministic data source. For instance, a user can derive the input statistics from the San Francisco Bay Area segment of the National Household Travel Survey and then conduct scenario analysis on grid impacts by systematically changing input statistics between each V2GSim run. 