The Modular Multilevel Matrix Converter is a direct AC/AC topology and promising in the medium-voltage high-power application. However, the current M3C prototypes mainly adopt the centralized control structure which may hinder its modularity and flexibility. Based on current distributed control architecture, a new one and corresponding design principles for M3C are proposed to realize both the distribution of control tasks and improve the expandability. And in order to reduce the dependency of real-time communication network, a time-dimension decoupling strategy based on voltage decoupling and current coupling control algorithm is employed, in which, the voltage loop are divided into the input side, output side and inner circulating decoupling models which are executed in the system controller, while the real-time current loop is established on the coupling abc coordinate frame in each local controller. In order to reduce the influence of communication delay, a novel scheme to replace direct AC current reference with DC power value and double synchronization phase information is adopted. The controller parameters of current loop are also designed carefully after frequency decomposition. Finally, an experimental prototype is constructed to verify the feasibility of proposed control scheme after a series tests of steady-state and dynamic operations.

The modular multilevel matrix converter (M3C) is introduced and applied to a single stage AC/DC solid state transformer (SST), which uses only one concentrated medium frequency transformer (MFT) for isolation. As a result, the transformer insulation design is simplified significantly and higher power density could be achieved. Meanwhile, a novel decoupling transformation based on two internal three-phase circulating systems is proposed to improve previous method. Through this transformation the unbalanced capacitor voltage components can be divided into four independent dimensions, which correspond to four kind of capacitor voltage ripple frequency. And a DC-side vector model is then established to calculate the required circulating current reference for capacitor balancing. Besides, under the application of SST, the capacitor voltage ripple suppression strategy via injection of extra circulating current is proposed. The amplitude and frequency index are optimized to obtain the best effect. Therefore, the submodule capacitor volume can be reduced to achieve higher power density. Simulation and experimental results are also provided to verify the theoretical analysis.

with the widespread growth in the use of electric vehicles (EVs), there is an urgent demand for fast-charging battery charger so as to alleviate the range anxiety for car users. A novel multi-pulse flexible-topology thyristor-based rectifier (mPFTTR) is proposed in this paper, which can be a suitable candidate for high-power fast-charging battery charger of the EVs. The proposed architecture has five operating modes: series mode with 24 pulses (SM24P), parallel mode with 24 pulse (PM24P), hybrid modes with 24 pulses (HM24P), series mode with 12 pulses (SM12P) and parallel mode with 12 pulses (PM12P). To meet the practical requirement of fast charging, the predictive current control strategy combined with multi-step constant current charging (MSCCC) is also proposed, and followed by the detailed implementation and analysis. Simulation and experimental verifications are carried out, and both results show that the predictive current control strategy combined with the MSCCC method can charge the EVs with wide-range input voltage, and fast charging is also achieved. Finally, efficiency evaluation and comparisons of total harmonic distortions (THDs) of the input current are developed between the proposed mPFTTR which operates in HM24P and the conventional thyristor-based rectifier which operates in SM24P.

Since the voltage stress on the power switches is only half of the input voltage, three-level (TL) converters have been widely used in high voltage applications. To ensure normal operation of TL converters, the voltages on the input divided capacitors and the blocking capacitor must be essentially balanced. In practice, asymmetry of the main circuit and the drive circuit may result in unbalance of these voltages, leading to higher voltage stress on the power switches and other problems. This paper proposes a novel Pulse-Width Modulation (PWM) method with better voltage-balancing ability, which is implemented by interleaving two traditional PWM methods. First, the mechanism of voltage unbalance is analyzed. The new PWM method decouples the blocking capacitor voltage from the voltages of the input divided capacitors, and ensures that the former voltage is always kept around its ideal value, which can greatly simplify the sampling circuit and the voltage-balancing control scheme. Then, a voltage-balancing strategy aiming to adjust the phase shift of the drive signals only is given. Finally, experimental results are presented to verify the effectiveness of the proposed method and voltage-balancing strategy.

Vienna rectifiers are widely used as front-end rectifiers in many applications because of their high efficiency and high power density. However, this topology has a problem that its current is always distorted near zero-crossing points, especially when the converter-side current ripples are large. This paper reveals and models the zero-crossing current distortion phenomenon, which is caused by non-shared vectors of adjacent sectors when the sector detection error occurs. Near the zero-crossing points, the inevitable sector detection errors will occur due to the presence of the switching current ripples and other factors such as the control delay. The errors can be divided into two types, namely the lag error and the lead error. The effects of the two types of errors are equivalently modeled in the dq frame in this paper. When the detection error occurs, it is equivalent to applying additional interference terms to the control output. Based on the equivalent error model, the paper proposes two compensation methods to eliminate the zero-crossing effect caused by the sector error. The proposed methods are easy-to-implement and effective. Finally, simulation and experimental results validate the theoretical analysis.