Charge Cycle Efficiency and Grid Power Quality Enhancement of an Off-Board Battery Charger Using Model Predictive Duty Cycle Control

The conventional finite control set model predictive control (FCS-MPC) applied to battery chargers exhibits a variable switching frequency operation. Subsequently, it becomes difficult to select appropriate semiconductor devices, design and tune passive components, and voltage/current sensing filters due to non-uniform thermal stress on semiconductor devices and unclean/noisy frequency spectrum, especially in grid-connected systems. To address the aforementioned concerns, this article introduces a model predictive duty cycle control (MPDCC) strategy for a two-stage (ac-dc and dc-dc) off-board battery charger with fixed switching frequency operation. This method generates duty cycles using optimal sector determination, voltage vector selection, and cost function minimization. In addition, a new time-weighted average efficiency (TWAE) metric is introduced to quantify charger performance under constant current-constant voltage (CC-CV) and constant power-constant voltage (CP-CV) charging profiles. To validate the efficacy of the proposed MPDCC method, a two-stage off-board charger system is developed and experimental studies are conducted on a 3-kW laboratory charger prototype with a dSPACE MicroLabBox-DS1202 control platform to charge the 2.4 kWh, 400 V battery pack under different charging modes. Finally, the charger's performance using the proposed MPDCC approach is compared with existing control methods in terms of grid power quality, battery voltage/current ripples, average switching frequency, peak efficiency, and the TWAE efficiency to show its superiority.