A statistical approach for price-optimal plugin electric vehicle charging

Authors
Delak, Serge A.
ORCID
Loading...
Thumbnail Image
Other Contributors
Kar, Koushik
Chow, J. H. (Joe H.), 1951-
Wang, Meng
Issue Date
2017-05
Keywords
Computer Systems engineering
Degree
MS
Terms of Use
Attribution-NonCommercial-NoDerivs 3.0 United States
This electronic version is a licensed copy owned by Rensselaer Polytechnic Institute, Troy, NY. Copyright of original work retained by author.
Full Citation
Abstract
Most of the STAT2 algorithm’s cost savings potential comes from the fact that it operates as a “bang-bang” controller operating on a volatile price signal. However, current RTM price signals do not reflect instantaneous load conditions, so this behavior improves cost savings while exacerbating, rather than reducing, load swings. These swings must be counterbalanced with additional regulation capacity. This thesis attempts to quantify the combined cost impact of using the STAT2 algorithm and the resulting increased need for regulation. Depending on the regulation cost model assumed, the cost savings potential of the STAT2 algorithm is either reduced or reversed.
The STAT2 algorithm was iteratively developed by using historical data to simulate charging a fleet of Tesla PEVs in the New York City price environment for a given developmental algorithm, assessing algorithm performance by comparison to the performance of simple and ideal (theoretical best) baseline algorithms, and then revising the developmental algorithm. Compared to the simple method of charging a PEV immediately after midnight, the STAT2 algorithm reduces costs by 16%, compared to an ideal cost savings of 23%. These cost savings are reduced as the combination of network constraints and increased PEV penetration reduces opportunities for deferring charging. However, the algorithm’s ability to successfully meet nightly PEV charging goals closely matches the best performance possible, even as charging capacity limitations become significant.
The STAT2 algorithm is designed around statistical price-energy distributions which use forecast data and historical statistics to anticipate prices and charging capacity through an entire charging period. Such a model is necessary because these parameters are not known with certainty until real time, yet anticipating them for the entire charging period is critical for successful optimization. The statistical price-energy distributions are used to generate vehicle-specific price setpoints which are compared, in real time, to the current RTM price to make a charging decision. The distributions are continually updated based on the latest charging and network conditions so that vehicle charging goals can be achieved adaptively at minimal cost.
Systems that consume energy and have energy storage capacity can defer purchasing energy during high price periods and use their stored energy to “ride through” those periods without impacting their ability to carry-out their main objective. Utilizing this practice can save money for consumers and, depending on the degree to which supply network loading is accurately reflected in prices, has the potential to smooth that network’s demand with time. This thesis investigates practical methodologies for best utilizing this practice for achieving cost savings by considering the specific application of Plugin Electric Vehicle (PEV) charging. Specifically, this thesis documents the evolutionary development of a practical algorithm for charging a fleet of PEVs within a Real Time Market (RTM) price environment. The final algorithm, termed “STAT2,” is also designed to limit charging rates to avoid violating network constraints, and to accommodate the unique charging needs of multiple independent PEVs.
Description
May 2017
School of Engineering
Department
Dept. of Electrical, Computer, and Systems Engineering
Publisher
Rensselaer Polytechnic Institute, Troy, NY
Relationships
Rensselaer Theses and Dissertations Online Collection
Access
CC BY-NC-ND. Users may download and share copies with attribution in accordance with a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License. No commercial use or derivatives are permitted without the explicit approval of the author.