Transportation is a key contributor of greenhouse gas emissions, recognized as one of the main factors contributing to climate change. For California to reach 100 percent zero-emission new passenger cars and trucks sales by 2035, per Executive Order N-79-20, understanding the effects of increased adoption of electric vehicles on the electrical system and the infrastructure required to support them is paramount to ensuring a smooth transition from internal combustion engine vehicles to electric vehicles. This project researched, developed, and demonstrated vehicle-grid integration in non-residential facilities; quantified the effects of electric vehicle charging on the grid, including its flexibility and revenue streams; and developed strategies to manage electric vehicle load to minimize the impact on the distribution system while minimizing customer utility costs. Finally, the project developed a method to evaluate emission impacts of charging electric vehicles given the current, and future, mix of bulk generation resources.
The project focused its analysis on passenger vehicles at workplaces and on a university campus electric bus fleet. At the workplace sites selected for the project and using historical data from charging sessions, the project demonstrated that applying smart charging strategies can achieve a 25 percent cost reduction and a 76 percent reduction in peak electricity demand, as well as add 66 percent more electric vehicles to the grid without requiring infrastructure upgrades while simultaneously guaranteeing that electric vehicle energy requirements are satisfied.
The algorithm used for real-time operation, which did not have information about the future, delivered 80 percent of the theoretical total energy required by electric vehicles, with a total savings of 30 percent, and a reduction in the maximum demand of 40 percent. For the electric bus fleet, smart charging obtained up to 88 percent in cost savings, with a 57 percent peak demand reduction and 52 percent emissions reduction. Finally, a transformer thermal datadriven model was developed and resulted in a 9.8 percent root mean squared error of the mean value.
