Space-Vector-Based Hybrid PWM for Zero-Sequence-Circulating-Current RMS and Common Mode Voltage Reduction in Two Parallel Interleaved Two-Level Converters

This paper aims at reducing the root mean square (RMS) values of the zero-sequence circulating current (ZSCC). The analysis reveals that the ZSCC is determined by the duty ratios of the medium voltage vector and zero vector. Whereas the reference voltage fixes the duty ratio of the medium vector, the duty ratio of the zero-voltage vector depends on the distribution of the small and larger vectors, which can be used to optimize the ZSCC. As such, we propose a generalized PWM architecture, where the distribution of the active vectors is parameterized by a coefficient k. Based on this, we derive regions for k that attain the same minimal ZSCC peak. Within these regions, we further optimize k to minimize the ZSCC RMS. Depending on the reference voltage, the method selects different coefficient k for the ZSCC RMS optimizations. The proposed modulation strategy is therefore a HBSVM due to its adaptive feature in the distribution of the active vectors. The proposed HBSVM uniformly applies to the entire vector plane and is computationally affordable for mainstream microcontrollers. Finally, the experimental results validate the merits of the proposed method.