A probability transition matrix based decentralized electric vehicle charging method for load valley filling

• A decentralized method is proposed for electric vehicles to fill the load valley.
• A probability transition matrix is used as the control signal.
• Charging schedules for valley filling are created in only several iterations.
• The method is suitable for controlling a heterogeneous population of electric vehicles.

This paper presents a decentralized control method to schedule EV (electric vehicle) charging loads to fill the overnight load valley while meeting customers’ charging requirements. A PTM (probability transition matrix) is calculated at the aggregator side as the control signal to guide EV charging processes based on submitted EV charging schedules. Elements of the PTM represent the transition probabilities of moving a charging load from one time period to another. At the EV side, each EV individually updates its charging schedule according to its charging requirements and the PTM. Then the updated schedules are sent back to the aggregator. This process is repeated iteratively until convergence. In this method, no optimal control problems need to be solved locally so that its implementation on the EV side requires low computation capability. Simulation results show that the proposed method can create desired EV charging schedules for load valley filling within only several iterations, making it suitable for real-time implementation.