In the last decade, both the U.S. and Chinese governments have spurred efforts to boost the utilization of transportation electrification technologies. For example, in China’s ‘The Twelfth Five-year Plan’, ‘Energy saving and new energy vehicles industry development plan (2012 – 2020)’, the production capacity of electric vehicles and plug-in hybrid electric vehicles is expected to reach 2 million per year, which is close to 10 percent of the current annual car sales.
This emerging demand because of unanticipated peaks and variability can pose a significant risk to the stability of the grid. For example, the U.S. has a generation capacity of approximately 1,000 GW. Assuming that each PEV has a storage capacity of 20KWh, with just a 10 percentadoption rate, the total energy stored in PEVs will be as much as 500GWh, a significant fraction of the current grid capacity. The problem is likely to become worse as adoption rates increase. According to one recent study by Bloomberg New Energy Finance, with big reductions in battery prices ahead, 35 percent of vehicles could be electric by 2040.
This project is aimed at turning the challenge into an opportunity by bringing together a team of researchers with expertise in algorithms for electronic commerce and for managing variable demand, distributed systems, and security/privacy. The key accomplishment will include laying the foundation for leadership in the area of a data-driven approach for models, algorithms, and mechanism design to incentivize users to charge vehicles at appropriate times and locations, leading to better load management, a more reliable grid, and even time and cost savings for end-users especially under high penetration of distributed renewable energy.
Furthermore, the research aims to make the mechanisms protect user privacy and be robust against manipulation or adversarial use. As a further opportunity, assuming the battery in vehicles (and in-home batteries like Tesla’s PowerWall permit) can potentially even become a distributed energy storage device for the utility in times of real high stress, with cash flow as incentive to end-users. Clearly, designing these mechanisms needs to factor in end-user’s driving and charging patterns, current and predictive grid load, and locations and charging rate of nearby charging devices.
The research objectives are as follows:
- Establish a multi-stage security-constrained framework to address uncertainties in EV behaviors; and maintain safety of the grid;
- Compile and preprocess available datasets from multiple sources related to EV charging and grid behaviors;
- Establish accurate statistical forecast models of EV charging and renewable energy generation
- Apply data mining techniques to the preprocessed datasets to help define charging models;
- Develop mechanism designs and algorithms to guide EVs’ charging/discharging behaviors under high penetration of renewable energy leveraging the model from the previous step. Ideally, mechanism should optimize the combination of perceived benefit to end-users (e.g., lower the weighted sum of charging costs, charging times, inconvenience due to unavailability of the car when it is charging), while ensuring the stability of the grid with a desired threshold of probability.
- Evaluate the mechanisms from the perspective of data that would need to be gathered and to the extent that the users’ privacy can be protected by anonymization or randomization;
- Implement a proof-of-concept testbed for evaluating the mechanisms and measuring their effectiveness.