archi 2018


SEVA Model charging for the future

Knowledge about how, where and when people charge makes you better prepared for the increasing demand for charging. The University of Amsterdam’s SEVA Model provides this. It used 4.5 million charging transactions from 2014 to 2017 to create a training set with different variables to make the most accurate possible predictions. It examines:

1. Activity patterns
2. Geographic clustering and destinations
3. Separate choice models for distance, costs and speed

Tested scenarios: an increase in car sharing; switch from plug-in hybrid to full electric; strategies to increase the number of electric vehicles. You can also test a random selection of locations, an increase in unique users, an increase in peak use or how much more vulnerable the grid becomes with an increase in demand. The end result is an optimal charging infrastructure, where users are persuaded to switch to electric vehicles. Presentation The University of Amsterdam

Go back