UCLA

Increasing Electric Vehicle (EV) integration to power grid has created additional technical challenges by requiring additional power/energy for charging EVs. With the introduction of Vehicle-to-Grid (V2G), EVs can strategically feed power stored in the batteries back to the power grid for grid supports, shifting their roles from traditional loads to distributed generations.

Different than other types of storages in power grids, the availabilities and energy demands of EVs are subject to change. Traditional deterministic modeling and control methods for Demand Side Management (DSM) of stationary storages failed to capture the stochastic nature in EVs, thus making suboptimal control decisions. This dissertation proposes a stochastic optimization based DSM for energy management of V2G EVs a microgrid setting. In order to make the stochastic problem tractable, a model-free probability density estimation method is utilized together with sample average approximation. Real-life EV data is applied for demonstrating the correctness and merits of the proposed DSM.

As EV penetration continues to grow, the energy management problem becomes oversized to be efficiently solved in a centralized manner. Towards this concern, a distributed DSM for V2G integration into distribution grids is proposed to dispatch the centralized computational burden to distributed nodes. The proposed DSM accounts for both nodal and networked operational cost. To accelerate the computational speed and guarantee convergence, the proposed DSM is tightly relaxed to a convex form using second-order cone programming. Real-life EV data is used to test the proposed DSM in an IEEE benchmark test system.

Most of other existing researches on V2G integration has focused on numerical simulations. As an important area of research, this dissertation shows the design and implementation of two V2G prototypes which could be used as testbed for verification of various V2G applications. One platform is capable of both supplying DC loads and V2G, the other one is capable of bidirectional charging. The prototypes utilize a commercialized EV model with widely used CHAdeMO communication interface. Hardware and software design considerations are shown with test results showing the performances of the prototypes.