Abstract
We will develop a decision-making tool to facilitate EV fleet management at scale. Our proposed framework consumes various data sets such as individual EV’s historical driving patterns and energy needs, energy profiles of buildings that are hosting charging infrastructure, and weather and traffic predictions. This data will be mapped onto electric and transportation graphs and will used to continuously determine charging and discharging decisions at fleet and individual EV levels.
Description
The Biden administration recently announced ambitious plans to achieve “50% Electric Vehicle Sales Share in 2030". In addition, the newly signed infrastructure bill is expected to direct substantial financial investments towards building EV charging infrastructure in the US. Electrifying transportation systems has the potential to shape the future of transportation sector as well as energy supply chain. In light of this, it is expected that major transportation and delivery companies will increasingly “electrify" their fleets and po- tentially enter into the business of delivering electric energy to buildings (i.e., charge and discharge energy) in addition to transporting goods and humans. Please see Figure 1.
With these new opportunities come also new challenges in coordinating and optimization of EV operations (i.e., routing, charging and discharging) across a vehicle fleet, as well as operations of energy transfer between vehicles and buildings. The underlying coordi- nation can be done to achieve any combinations of the following objectives: (i) maximizing sourcing of EV energy needs from clean sources (ii) providing optimized energy services at different locations in the power networks (i.e., transporting electricity), and (iii) charge management to transport goods and human. It is important to note that, although adopting electrified mobility reduces climate footprint of our transportation sector, electrification can overburden the power grid and turn on pollutant generators.
In this project, we will develop a decision-making tool to facilitate EV fleet management at scale. Our proposed framework consumes various data sets such as (i) individual EV’s historical driving patterns and energy needs, (ii) energy profiles of buildings that are hosting charging infrastructure, and (iii) weather and traffic predictions. This data will be mapped onto electric and transportation graphs and will used to continuously determine charging and discharging decisions at fleet and individual EV levels. The following scenarios show- cases three realistic decision making examples:
• Optimized multi-objective decision-making for electric vehicle fleets in conjunction with building energy management systems. Decision making variables include, when to charge, how much to charge each vehicle in conjunction with renewable sources, energy level of the battery, and anticipated needs of the fleet and the building.
• Feasibility analysis and guidelines for operation of EV fleets. This analysis can help answer questions like how many vehicles are needed to guarantee availability of n vehicles at any given time.
• Contingency planning for worst-case scenarios including power outages, demand disruptions, and charge point failures.
Timeline
07/2022 -- 12/2022: Setting-up of the decision-making framework for EV fleet management including charging times and places.
10/2022-12/2022: Analyze data from deployment partners and scaling up data
01/2023-06/2023: Develop an optimized multi-objective decision-making system for EV fleets in conjunction with building energy management systems. During this time period, we will continue to collect data from the partners and fine tune our decision making tool.
Strategic Description / RD&T
Deployment Plan
Our main deployment product, in collaboration with our deployment partner, will be a software tool for multi-objective decision-making for managing electric vehicle fleets in conjunction with building energy management systems. The software tool will output decisions on when and how much to charge each vehicle while taking as input the constraints of the EV fleet and the building. We aim to provide an extensive analysis of our decision-making tool and asses its performance in terms of total energy usage, fleet size, etc. We also plan to utilize this tool to conduct a detailed analysis of EV fleet feasibility and obtain an understanding on the minimum number of EVs needed in the fleet to satisfy given requirements.
Expected deliverables include
-- An open source software tool that outputs decisions on when and how much to charge individuals EVs in a fleet to satisfy given constraints on the fleet and building energy usage.
-- A report demonstrating the performance of the resulting decision-making tool.
-- A report on the feasibility of EV fleets and lessons learned about EV fleet operation and management, which we expect to be useful community stakeholders and policy makers.
In addition to these products, we plan to publish our findings as journal and conference papers (e.g., IEEE Transactions on Control of Network Systems, IEEE Transactions on Networking. Preprints of these publications will be publicly released online.
Expected Outcomes/Impacts
Expected accomplishments include
-- Methods for optimized multi-objective decision-making for electric vehicle fleets in conjunction with building energy management systems. Decision making variables include, when to charge, how much to charge each vehicle in conjunction with renewable sources, energy level of the battery, and anticipated needs of the fleet and the building.
-- Feasibility analysis and guidelines for operation of EV fleets.
-- Contingency planning for worst-case scenarios including power outages, demand disruptions, and charge point failures.
Expected Outputs
TRID
Individuals Involved
| Email |
Name |
Affiliation |
Role |
Position |
| mkoeske@cmu.edu |
Koeske, Matt |
ECE |
Other |
Other |
| TBN@andrew.cmu.edu |
TBN, PhD Student TBN |
ECE |
Other |
Student - PhD |
| oyagan@ece.cmu.edu |
Yagan, Osman |
ECE |
PI |
Faculty - Untenured, Tenure Track |
Budget
Amount of UTC Funds Awarded
$100000.00
Total Project Budget (from all funding sources)
$282679.00
Documents
| Type |
Name |
Uploaded |
| Data Management Plan |
Managing_EV_Fleets_to_deliver_humans_goods_and_electricity.pdf |
Feb. 11, 2022, 8:05 a.m. |
| Progress Report |
388_Progress_Report_2022-09-30 |
Sept. 27, 2022, 11:43 a.m. |
| Progress Report |
YaganReportMarch2023.pdf |
March 21, 2023, 6:22 a.m. |
| Progress Report |
388_Progress_Report_2023-03-30 |
March 28, 2023, 11:54 a.m. |
| Publication |
Dynamic coupling strategy for interdependent network systems against cascading failures |
April 10, 2023, 8:41 p.m. |
| Publication |
Spreading processes with population heterogeneity over multi-layer networks |
April 10, 2023, 8:42 p.m. |
| Publication |
Correlated combinatorial bandits for online resource allocation |
Oct. 18, 2023, 7:26 a.m. |
| Publication |
Evaluating Resilience in Mixed-Autonomy Transportation Systems |
April 10, 2023, 8:44 p.m. |
| Publication |
Robustness of Random K-out Graphs |
Oct. 18, 2023, 7:26 a.m. |
| Publication |
On the connectivity and giant component size of random k-out graphs under randomly deleted nodes |
April 10, 2023, 8:46 p.m. |
| Publication |
The effects of evolutionary adaptations on spreading processes in complex networks |
Oct. 18, 2023, 7:26 a.m. |
| Final Report |
Final_Report_-_Yagan_388.pdf |
July 31, 2023, 4:49 a.m. |
| Presentation |
Evaluating the Optimality of Dynamic Coupling Strategies in Interdependent Network Systems |
Oct. 18, 2023, 7:26 a.m. |
| Progress Report |
388_Progress_Report_2023-09-30 |
Oct. 18, 2023, 7:26 a.m. |
Match Sources
No match sources!
Partners
| Name |
Type |
| University of Texas - Austin |
Deployment Partner Deployment Partner |
| Allegheny County |
Deployment Partner Deployment Partner |
| Port Authority of Allegheny County |
Deployment Partner Deployment Partner |
