For modular multilevel converter (MMC)-based traction substation (TSS) in 24kV medium voltage direct current (MVDC) railway power supply system (RPSS), fast dynamic response is core requirement for its DC-link voltage control. In this paper, a fast-dynamic control strategy by a combination of dynamic reference generation (DRG) concept and model predictive control (MPC) is proposed to achieve fast dynamic response of MMC-based TSS’s DC-link voltage control. In outer DC-link voltage control loop, an improved DRG strategy based on extended state observer (ESO) (remarked as ESO-based IDGR strategy) is proposed to regulate DC-link voltage, in which an ESO is utilized to estimate system total disturbance (sum of TSS’s output power and power loss), and ac side current references are generated based on discrete-time model of TSS’s DC-link voltage. In inner current control loop, a model predictive current control (MPCC) strategy based on continuous-control-set MPC (CCS-MPC) principle is developed to simultaneously regulate MMC’s ac side current and circulating current with fast dynamic response. With the proposed control scheme, good disturbance rejection and fast voltage reference tracking abilities, as well as satisfactory steady-state performance can be achieved for MMCbased TSS. Finally, several simulation results are provided to demonstrate effectiveness and feasibility of the proposed control strategy.

Flexible traction substations (FTSSs) integrating power converters, solar photovoltaics (PVs), and energy storage systems (ESSs) are beneficial for electric railways (ERs) evolving toward a sustainable and efficient system. However, traction load and PV generation are uncertain and fluctuating, unfavorable to efficient PV energy consumption in ERs. Thus, this paper proposes a power flow control strategy for FTSSs. First, a real-time power flow optimization method is proposed to calculate the output power references of the back-to-back converter (BTBC) and ESS. It aims for efficient regenerative braking energy and PV energy utilization and negative sequence current suppression. Then, a model predictive power control (MPPC) scheme is designed for BTBC to track output power and DC-bus voltage references quickly and accurately. It includes a revised dynamic reference generation method and a model-based double-line frequency ripple (DLFR) filter. Meanwhile, an MPPC scheme for a bidirectional DC-DC converter is designed to regulate output power of ESS, which realizes accurate power reference tracking and smooth operational mode transition. Finally, the validity of the proposed strategy is confirmed through simulations and experiments.