Project 13:
Optimal Charging Scheduler for Electric Vehicles on the Florida Turnpike

Objective

Develop the methodology for analyzing the roadway traffic patterns and expected penetration and timing of electric vehicles (EVs) on the Florida Turnpike. The work will determine the requirements for electric vehicle supply equipment at turnpike plazas, the options for equipment siting and the economics.

Brief Description

This project will develop a methodology to simulate and analyze roadway traffic patterns and expected penetration and timing of electric vehicles (EVs) with application directed toward the requirements for electric vehicle supply equipment (EVSE) siting and purchasing/leasing at turnpike plazas. The project will also develop a wireless-communication-based driver-assistance application that would optimize the location and timing of charging for individual drivers as well as efficiency of the overall charging network.

The first steps of the project were to develop the systematic methodology for analyzing expected penetration of electric vehicles (EVs) and their impacts on the overall transportation infrastructure. An analytical model consisting of three components was developed and applied. Furthermore, the stability and instability of the proposed model are also analyzed. The developed model was then applied to the Florida Turnpike network as a test system. As a result of three-year research efforts, three peer-refereed papers and three conference articles have been published. A journal article on scheduling of EVs on the Turnpike is currently under review.

Research Results

The project developed a systematic methodology for analyzing expected penetrations of electric vehicles (EVs) and their impacts on the overall transportation infrastructure. The analytical model consisted of three components:

• A dynamic model which, for a highway with a total of N nodes (entrances and exits), admits either instantaneous or average traffic flow passing through entrances/exits;
• A queuing model on the number of EVs waiting at a given service station for charging and their waiting time;
• A network level model that prescribes the decision making process of individual drivers as well as any coordination among services stations and individual drivers (using V2V and V2I communication).

Using the model, a network-level protocol was synthesized to optimize the performance (minimum waiting/charging time) of the overall system. In particular, a distributed scheduling algorithm was used for the overall charging network and a cooperative control algorithm for individual drivers to make their decisions. Furthermore, the stability and instability of the proposed distributed algorithm are also systematically analyzed.

The Florida Turnpike network has been used as a test system for this project. Two meetings were held between the research team and Florida Turnpike engineers, and real-world data were collected and compiled from Regional Integrated Transportation Information System (RITIS) database in collaboration with Project 7. Case studies via computer simulations are conducted using 24-hour Florida Turnpike roadway traffic data and the currently installed electric vehicle charging equipment data at the turnpike plazas. The charging equipment utilization and electric vehicle waiting time are compared for different EVs penetration rate and charging strategies. It is demonstrated that the proposed strategy results in a uniform utilization of the charging equipment at the service stations.

Discussion with Florida Turnpike Authority staff has determined that six Tesla charging stations have been installed at both the Drum (Yeehaw Junction area) and Turkey Lake (Orlando area) service plazas. These stations were installed and are owned by Tesla under a sub-lease agreement with the service plaza concession Areas. The Turnpike Authority expects to get yearly operational data from Tesla. There is also a level 2 charger installed by Charge Point at Snapper-Creek Plaza in Kendall, FL. The next charging station phase is the installation of 4 to 6 DC chargers at both the Drum and Turkey Lake Plazas by EVGO, again under sub-lease with Areas. Finally, the turnpike individuals have stated that EVGO is interested in installing more chargers at other service plazas and that UCF could help in this activity.

The 4-year project activities are:

• Year 1: Develop a mathematical model for analyzing and simulating EV charging needs based on roadway traffic patterns, and conduct preliminary simulation studies to demonstrate the model. (Status: Completed)
• Year 2: Initiate the effort of defining the available data, validate the model together with engineers at Florida Turnpike Authority (FTA), and use sample real-world traffic data to determine the requirement of EVSE at service plazas. (Status: completed, except for pending inputs from FTA)
• Year 3: Further analysis of stability and instability of the developed distributed scheduling algorithm. (Status: Completed)
• Year 4: Work with engineers at Florida Turnpike Authority to study the applicability of real-time scheduling, and develop a driver-assistance phone application that would optimize the location and timing of charging for individual drivers as well as efficiency of the overall charging network. (Status: pending inputs from FTA)

Impacts/Benefits

As a result of three-year research efforts, the following peer-refereed papers on the modeling has been presented/published or accepted:
Journal papers:

• Ali Maknouninejad and Zhihua Qu, "Realizing Unified Microgrid Voltage Profile and Loss Minimization: a Cooperative Distributed Optimization Approach," IEEE Transactions on Smart Grid, vol.5, no.4, pp.1621-1630, July 2014.
• Ali Maknouninejad, Zhihua Qu, Frank Lewis, and Ali Davoudi, "Optimal, Nonlinear, and Distributed Designs of Droop Controls for DC Microgrid," IEEE Transactions on Smart Grid, vol.5, no.5, pp.2508-2516, September 2014.
• H. Valizadeh Haghi, S. Lotfifard, and Z. Qu, “Multivariate Predictive Analytics of Wind Power Data for Robust Control of Energy Storage,” IEEE Transactions on Industrial Informatics, vol.12, no. 4, pp. 1350-1360, August 2016.

Conference articles:

• Azwirman Gusrialdi, Zhihua Qu, and Marwan A. Simaan, "Scheduling and Cooperative Control of Electric Vehicles" Charging at Highway Service Stations", in Proceedings ofthe 53rd IEEE Conference on Decision and Control, Los Angeles, California, USA, pp. 6465-6471, December 15-17, 2014.
• Hamed Valizadeh Haghi and Zhihua Qu, "Implementing smart grid optimal operation using model-based analytics," panel presentation, 2015 IEEE PES General Meeting, Denver, Colorado, USA, July 26-30, 2015.
• Azwirman Gusrialdi and Zhihua Qu, "Analysis of Cooperative Systems with Time Delay: Application to Transportation Systems," Multi-Conference on Systems and Control, Buenos Aires, Argentina, September 19-22, 2016.
• Youngjun Joo and Zhihua Qu, "Cooperative Control of Heterogeneous Multi-agent Systems in Sampled-data Setting", the 55th IEEE Conference on Decision and Control, ARIA Resort & Casino, Las Vegas, USA, December 2016. Project 13
• Roland Harvey and Zhihua Qu, ``Cooperative Control and Networked Operation of Passivity-Short Systems," in Control of Complex Systems: Theory and Applications, K. G. Vamvoudakis and S. Jagannathan (Eds.), pp. 499-518, Elsevier, Cambridge, MA, 2016. Project 13

Currently, a journal paper on scheduling of EVs on the turnpike is under review. In addition, a meeting with Florida Turnpike Authority is planned to discuss the results of the conducted case studies and to determine the next step of collaboration. The next steps depend more on inputs from engineers at Florida Turnpike Authority.

Reports

See publications above.