Abstract
Electrical power systems with high solar generation experience a phenomena called "duck curve" which require conventional power generators to quickly ramp-up their output, thus resulting in financial losses. In this paper, we propose an online model (OLM) for scheduling the charging of electric vehicles (EV) located at park-and-ride facilities for flattening solar "duck curves". This model provides a significant improvement to existing ones for similar systems in the sense that the availability of information is related to the time period for which the optimization is done. In addition, a procedure for finding the schedules for EV charging that significantly decreases the ramping requirements is introduced. Proposed procedure includes a combination of a heuristic function and a neural network (NN) to make a decision on which EVs will be charged at each time period. The training of the NN is done based on optimal solutions for problem instances corresponding to the full information model (FIM). The computational experiments have been performed for instances reflecting different levels of solar generation and EV adoptions and prove highly promising. They show that the OLM manages to find schedules of similar quality as the FIM, while having some more desirable properties.
Original language | English |
---|---|
Title of host publication | 2020 International Joint Conference on Neural Networks (IJCNN) |
Place of Publication | Piscataway, NJ. |
Publisher | IEEE |
Number of pages | 8 |
ISBN (Print) | 9781728169279 |
DOIs | |
Publication status | Published - 28 Sep 2020 |
Event | IEEE World Congress on Computational Intelligence 2020 - Glasgow, United Kingdom Duration: 19 Jul 2020 → 24 Jul 2020 https://wcci2020.org/ |
Conference
Conference | IEEE World Congress on Computational Intelligence 2020 |
---|---|
Abbreviated title | WCCI |
Country | United Kingdom |
City | Glasgow |
Period | 19/07/20 → 24/07/20 |
Internet address |
Keywords
- electric vehicle charging
- solar duck curves
- online model