Patterns, no patterns, that is the question: Quantifying users' electric vehicle charging

Policymakers have structured legislation and set targets to increase the adoption of electric vehicles and the deployment of charging stations. As questions arise on how to expand the electric vehicle infrastructure, advanced computational models are attempting to quantify the future placement of charging stations. These models, however, do not necessarily incorporate the variability of user behavior as a subgroup of consumer behavior, as this information is not well-understood. This paper explores a state-wide network of 68 public charging stations' data representing 58,273 charging events to determine users' charging patterns. The methodological approach applied k-means clustering with Minkowski embedded distances to distinguish unique charging subgroups from 608 frequent users. The results indicate four user types: convenient (n = 319, 53%), gradual (n = 147, 24%), anxious (n = 104, 17%), and urgent (n = 38, 6%). Quantifying these users' unique patterns supports the human variability knowledge required for more comprehensive and holistic models of electric vehicle charging station placement. More broadly, these results support the user-centric design of public charging infrastructure, which assists in reducing the electric vehicle adoption threshold in the long run and facilitating the transition to a low-carbon, sustainable transportation system.

Automated summary

Policymakers have set goals to promote electric vehicle adoption and charging station deployment, and computational models are being used to determine the optimal placement of charging stations. This study analyzes data from a state-wide network of public charging stations to identify different patterns of user behavior. The results highlight four distinct user types and provide valuable insights for designing user-centric charging infrastructure.