Designed Dynamic Reference with Model Predictive Control for Bidirectional EV Chargers

This paper presents a finite control set model predictive control (MPC) using a designed dynamic reference for bidirectional electric vehicle (EV) chargers. In the conventional MPC scheme, a PI controller is involved to generate an active power reference from the DC voltage reference. It is hard to find one fixed set of coefficients for all working conditions. In this paper, a designed dynamic reference based MPC strategy is proposed to replace the PI control loop. In the proposed method, a DC voltage dynamic reference is developed to formulate the inherent relationship between the DC voltage reference and the active power reference. Multi-objective control can be achieved in the proposed scheme, including controlling of the DC voltage, battery charging/discharging current, active power and reactive power, independently. Bidirectional power flow is operated effectively between the EV- and the grid-side. Experimental results are obtained from a laboratory three-phase two-stage bidirectional EV charger controlled by dSPACE DS1104. The results show that fast dynamic and good steady state performance of tracking the above objectives can be achieved with the proposed method. Compared with the system performance obtained by the conventional MPC method, the proposed method generates less active power ripples and produces a better grid current performance.