Search

The modular multilevel matrix converter (M3C) is a promising topology for high-voltage high-power applications. Recent researches have proved its significant advantages for adjustable-speed motor drives compared with the back-to-back modular multilevel converter (MMC). However, the branch energy balancing in the M3C presents great challenge especially at critical-frequency points where the output frequency is close to zero or grid-side frequency. Generally, this balancing control depends on the appropriate injection of inner circulating currents and the common-mode voltage (CMV) whereas their values are hard to determine and optimize. In this paper, an optimization based predictive control method is proposed to calculate the required circulating currents and the CMV. The proposed method features a broad-frequency range balancing of capacitor-voltages and no reactive power in the grid side. For operation at critical-frequency points, there is no increase on branch voltage stresses and limited increase on branch current stresses. A downscaled M3C system with 27 cells is designed and experiment results with the R-L load and induction motor load are presented to verify the proposed control method.

The modular multilevel matrix converter (M3C) is a promising topology for high-voltage high-power applications. Recent researches have proved its significant advantages for adjustable-speed motor drives compared with the back-to-back modular multilevel converter (MMC). However, the branch energy balancing in the M3C presents great challenge especially at critical-frequency points where the output frequency is close to zero or grid-side frequency. Generally, this balancing control depends on the appropriate injection of inner circulating currents and the common-mode voltage (CMV) whereas their values are hard to determine and optimize. In this paper, an optimization based predictive control method is proposed to calculate the required circulating currents and the CMV. The proposed method features a broad-frequency range balancing of capacitor-voltages and no reactive power in the grid side. For operation at critical-frequency points, there is no increase on branch voltage stresses and limited increase on branch current stresses. A downscaled M3C system with 27 cells is designed and experiment results with the R-L load and induction motor load are presented to verify the proposed control method.

The modular multilevel matrix converter (M3C) is a promising topology for high-voltage high-power applications. Recent researches have proved its significant advantages for adjustable-speed motor drives compared with the back-to-back modular multilevel converter (MMC). However, the branch energy balancing in the M3C presents great challenge especially at critical-frequency points where the output frequency is close to zero or grid-side frequency. Generally, this balancing control depends on the appropriate injection of inner circulating currents and the common-mode voltage (CMV) whereas their values are hard to determine and optimize. In this paper, an optimization based predictive control method is proposed to calculate the required circulating currents and the CMV. The proposed method features a broad-frequency range balancing of capacitor-voltages and no reactive power in the grid side. For operation at critical-frequency points, there is no increase on branch voltage stresses and limited increase on branch current stresses. A downscaled M3C system with 27 cells is designed and experiment results with the R-L load and induction motor load are presented to verify the proposed control method.