Multi-port power electronic transformers (PET) are key parts of modern power grids, but their losses are difficult to model. DSIM simulations are useful for complex power electronic converters, especially PET simulations. This study analyzed the losses in a multi-port PET which showed the effectiveness of DSIM software for analyzing complex power electronic converter systems. This paper described the PET loss model which was then validated by simulations and experiments. The results show the optimum efficiency operating point of a single PET. When two PETs are available, one PET in operation is the most efficient for light loads and two PETs in operation are the most efficient for heavy loads. This loss model analysis provides a reference for optimizing the efficiency of PET operations.

A low-voltage ride-through (LVRT) control strategy for the multi-port power electronic transformer (PET) based on power co-regulation is proposed. During the sag and recovery of the grid-side voltage of the medium-voltage ac (MVac) port, the grid-connected active power of the low-voltage ac (LVac) port, rather than the power from external renewable energy sources (e.g., photovoltaic (PV)), is adjusted quickly to rebalance the power flowing across all ports, thereby preventing overcurrent and overvoltage. Moreover, a power-coordinate-frame-based LVRT mode classification is designed, and a total of six LVRT modes are classified to meet the LVRT requirements in all power configuration scenarios of the PET. In this way, the PET is endowed with the LVRT capability in both power-generation and power-consumption states, which is significantly different from traditional power generation systems such as PV or wind power. Furthermore, by optimizing the active power regulation path during LVRT transition, the overcurrent problem caused by the grid-voltage sag-depth detection delay is overcome. Finally, the effectiveness of the proposed control scheme is verified by experiments on a hardware-in-the-loop platform.