Demand-Side Management For PEV Charging In A Smart Grid

Plug-in electric vehicles (PEVs) are expected to become widespread in the near future because of environmental and economic benefits. However, high level penetration of PEVs may overload the distribution network, especially low-voltage transformers. In a smart grid, the charging of PEVs can be coordinated to reduce the peak load, known as demand-side management(DSM). This thesis focuses on the DSM for PEVs charging at low-voltage transformers in the residential area and a decentralized algorithm to coordinate PEVs charging is proposed. Main work of this thesis is as follows.1. The concept of smart grid and demand-side management, the history of electric vehicles and the negative effects of high level penetration of PEVs on power systems are introduced briefly. Besides, relevant works are reviewed and analysed.2. The DSM problem for PEV charging is formulated as a convex optimization problem, after the introduction of the PEVs dynamic charging model and distribution system structures. The objective is to flatten the load curve of the low-voltage transformer as much as possible, while satisfying each consumer’s requirement for their PEVs to be charged to the required level by the specified time.3. A decentralized algorithm based on the famous water-filling principle is proposed and pulse-width modulation(PWM) is utilized to handle the on-off type charger problem. The moving horizon idea is also introduced to alleviate the errors due to inaccurate forecast and quantiza-tion caused by PWM, as well as to tackle the random arrival time of PEVs. Moreover, each PEV only needs to calculate its own power allocation, and hence its implementation requires low computation and communication capacity and the algorithm is scalable.4. Theoretical proofs are given to prove the optimality of the proposed algorithm while nu-merical simulations are presented to demonstrate the effectiveness of the algorithm more manifestly.The value of the research is that the algorithm proposed in this thesis is simple and easy to implement, so it is suitable for engineering practice.