Dynamic DC-Link Current Minimization Control to Improve Current-Source Solid-State Transformer Efficiency

Abstract

This paper presents a fast predictive control method to minimize the dc-link current in current-source solid-state transformers (SST). The motivation for dc-link current reduction is to improve efficiency of the SST and reduce conduction loss and transformer winding loss, which are among the major loss mechanisms of the current source solid-state transformers. The proposed cycle-by-cycle dc-link current minimization strategy is realized by a predictive control method based on the converter model. The dynamic dc-link current reference is calculated each switching cycle by adding up the input and output phase current references with form factors. Therefore, it is adaptive to different loading conditions. Moreover, the control does not require high computational burden and is suitable for implementation in regular DSP for online calculation. In this paper, the improved soft-switching solid-state transformer (S4T) with reduced conduction loss is studied as an example of the current-source SST. The dc-link current minimization control method is verified in simulations and experiments at 2 kV under both steady-state and dynamic conditions. With the proposed control, dc-link current can be minimized to achieve approximately 18% loss reduction of the SST at the operating point under study.