<<< back

Electrical Vehicle Charging


The paper Optimal scheduling of electrical vehicle charging under two types of steering signals presents a linearithic time EV scheduling algorithm for a quadratic objective. Below a short description and a visualization of the algorithm is given. The publication can be cited as:

- T. van der Klauw, Marco E. T. Gerards, G. J. M. Smit, and J. L. Hurink. Optimal scheduling of electrical vehicle charging under two types of steering signals. In IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe), Istanbul, Turkey, pages 122:1--122:6, USA, October 2014. IEEE Power & Energy Society. [ DOI | http ]

Problem statement

Electrical vehicles impose a high load on the electricity grid. Although nowadays most neighborhoods have few electrical vehicles and typically no problems occur, a future scenario wherein many vehicles are charged simultaniously may result in an overloaded grid. In a recent field test such future scenario was studied by charging multiple vehicles and using ovens simultaneously in a Dutch neighborhood, which resulted in power quality issues and a black out on a part of the local network (Dutch media coverage). Our research makes it possible to shape the charging profile, such that it can be used as a subroutine in demand-side management approaches (such as profile steering) to prevent such overloading.

Our solution

The paper presents an algorithm that finds the optimal charging profile in O(N log N) time, where N is the number of charging intervals. Both prices, distance to the target profile and linear combinations of these can be used as an objective for optimization. Below, only the minimization of the euclidean distance toward the target profile is demonstrated, while the theory in the paper can be applied more generally. Due to the low complexity, the algorithm is scalable and can be easily implemented in an embedded system.

Demonstration of the algorithm

The demonstration below shows the profiles (using 15-minute intervals) that result from our algorithm. The slider below the plot configures the allowed charging window, e.g., the arrival and departure time of the EV. The baseload of the considered house, the capacity of the EV, the initial state of charge, the maximum charging power, and the target profile can be configured at the left of the figure.

Baseload of the house

Capacity of the EV

SoC before charging

Max. power:

Target profile

Last updated: 2015-11-27