Search

This paper proposes a real-time joint regulating reserve deployment (RRD) model of electric vehicles (EVs) and coal-fired generators (CGs) considering EV battery degradation. The superiority of the proposed model lies in the integration of the EVs' fast regulation speed and CGs' large regulation capacity. The whole optimization model is formulated under a two-stage Markov Decision Process (MDP) to correspond to the real-time RRD process. The user behavior characteristic and EV battery degradation are taken into account to stimulate the users to participate in RRD. Based on the formulated MDP, a piece-wise linear (PWL) function based scalable approximate dynamic programming (ADP) algorithm for different EV clusters is constructed to solve the real-time RRD model under uncertainties. A separable slope update method is proposed to update the slopes of PLF for different EV clusters with different user behaviors. The proposed ADP based real-time RRD algorithm (ADP-RTRRD) provides the approximate optimal real-time RRD policy with the empirical knowledge embedded through off-line training. Numerical simulation on a modified IEEE-39 bus system and Henan power grid in China verify the superiority of the proposed joint RRD model and the advantages of the proposed ADP-RTRRD algorithm. • Proposing the joint RRD model of EVs and CGs considering EV battery degradation. • Taking advantage of EVs' fast regulation speed and CGs' large regulation capacity. • Applying EV aggregation methods to simplify optimization complexity. • Reformulating the real-time RRD process under a two-stage MDP. • Proposing a scalable ADP algorithm to obtain approximate optimal solution. [ABSTRACT FROM AUTHOR]