Facilities location optimization for electric vehicle charging infrastructure

• Main Author: Hong, Jun Chew
• Other Authors: Lee Bu Sung, Francis
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/156560
• Institution: Nanyang Technological University

With a growing acceptance of Electric Vehicles (EV) from both consumers and policymakers [1], demand for EV charging stations has been projected to increase 35 times to a total of 60,000 EV charging points by 2030 in Singapore [2]. Additionally, the availability of location-tracking devices has allowed for movement patterns analysis with a high degree of spatial and temporal grain using this Global Positioning System (GPS) data. Hence, we propose a data-driven methodology for Facility-Location-Optimization (FLO) of EV charging infrastructure. We start by approximating charging station demand from GPS data into discrete geographical grids, after which we can estimate regional demand. With that, we will then formulate an optimization problem equivalent to the well-known set cover problem [3]. After which, we look to apply the proposed methodology to optimize EV grid layout in Singapore, with the objective of minimizing the average distance travelled by drivers and total number of charging stations. We explored various methodologies such as classical machine learning methods such as DBSCAN. This report also includes the investigation into the setup of a distributed system to accelerate the research process. By storing and processing data on a file system such as Hadoop File System (HDFS) [4] on the cloud, we can take advantage of parallel processing provided by map-reduce [5] and the scalability of cloud providers such as Google Cloud [6]. Lastly, together with A.Prof Markus Schläpfer (Adjunct) Professor, Prin cipal Investigator at the ETH Future Cities Laboratory (ETH-FCL), A.Prof Francis Lee Bu Sung and Dr Seanglidet Yean. We built a GPS dataset and distributed infrastructure to investigate the Vehicle to Grid (V2G) technology, with a paper currently in preprint [7] on arXiv with plans for conference submission.