Collective Effects and Performance of Algorithmic Electric Vehicle Charging Strategies

TL;DR

This paper models and analyzes the collective behavior and performance of algorithmic electric vehicle charging strategies using a convex optimization framework integrated with real-time user preferences.

Contribution

It introduces a convex second order cone optimization model for EV charging control that incorporates individual driver preferences and evaluates collective effects through simulation.

Findings

Effective congestion control in distribution grids.

Impact of individual strategies on overall system performance.

Feasibility of real-time, preference-based EV charging optimization.

Abstract

We combine the power flow model with the proportionally fair optimization criterion to study the control of congestion within a distribution electric grid network. The form of the mathematical optimization problem is a convex second order cone that can be solved by modern non-linear interior point methods and constitutes the core of a dynamic simulation of electric vehicles (EV) joining and leaving the charging network. The preferences of EV drivers, represented by simple algorithmic strategies, are conveyed to the optimizing component by real-time adjustments to user-specific weighting parameters that are then directly incorporated into the objective function. The algorithmic strategies utilize a small number of parameters that characterize the user's budgets, expectations on the availability of vehicles and the charging process. We investigate the collective behaviour emerging from individual strategies and evaluate their performance by means of computer simulation.