The conventional finite-control-set model predictive control (FCS-MPC) approach for controlling three-phase three-level inverters is plagued by significant current ripple. To minimize the root mean square error (RMSE) of the current control, this paper proposes a more straightforward and efficient solution using dual vectors. Firstly, a dual-vectors algebraic preselection method is presented to reduce computational complexity. Secondly, the continuous current reference is adopted for the first time instead of the discretized value to improve the accuracy of the current control error calculation, which leads to a more precise RMSE calculation. Thirdly, the impact of the sequence and duty cycles of the vectors on the RMSE is quantified for the first time, and the optimal solution is proposed to minimize RMSE. Compared to existing dual-vectors optimization algorithms, the proposed algorithm reduces the RMSE and total harmonic distortion of output current by more than 25% and 10%, respectively. Experimental results are presented to validate the advantages of the proposed algorithm.

The computational complexity of the three-level space vector pulse width modulation (SVPWM) limits its application in high-switching-frequency converters. To address this issue, this paper proposes a simplified SVPWM algorithm based on virtual chopping. Firstly, the reference voltage vector (VV) is adjusted to avoid excessive line-to-line voltage jumps. Secondly, boundary VVs in the space vector diagram, with simple switching sequences, are chosen as standards. Thirdly, the switching sequence of the reference VV is determined by virtual chopping the corresponding standard VV’s. Furthermore, to generate it in the microcontroller, the corresponding modulation waves are obtained. Finally, a method to balance DC side capacitor voltages is proposed. Consequently, the proposed algorithm simplifies the process, requiring only four equations. The sector determination, dwell time calculation and allocation are bypassed. Compared with the existing algorithms, the execution time and design complexity of the proposed algorithm in the microcontroller is reduced. Experimental results are presented to verify its effectiveness.

This paper clearly demonstrates the architecture of time synchronization network of experimental advanced superconducting tokamak (EAST) poloidal field (PF) power supply control system and analyzes characteristics of the IEEE 1588 in detail at the same time. IEEE 1588 has unique advantages for this paper to utilize as protocol in order to improve the accuracy of the time synchronization network (TSN). Consequently, we design a time synchronization system based on PCI eXtensions for Instrumentation-6683H. This paper utilizes one master node to synchronize all slave nodes on the time network in each second. The result shows that the synchronization accuracy can be achieved within 50 ns. According to results of experiments, this system can perfectly meet the requirements of EAST PF power supply control system.

In the high-power superconducting fusion converter, the single-phase neutral point clamped (NPC) three-level converter is applied to replace the traditional thyristor converter using an integrated gate-commutated thyristor (IGCT) as a power device. The operating frequency of IGCT is usually lower, only a few hundred Hz. An optimized hybrid space vector pulsewidth modulation (SVPWM)-8H is proposed to solve the problems of discontinuous output average voltage, low dc bus voltage utilization rate, dead time compensation, and neutral point potential balance. The new sectors are inserted to make the range of output average voltage of adjacent sectors intersect, which improves the continuity of output average voltage. Furthermore, some sectors are composed of four vectors to reduce the minimum working time of redundant small vectors, improving the dc bus voltage utilization ratio and the converter’s performance. Moreover, by analyzing the influence of dead time on redundant small vectors, redundant vectors are rearranged to make redundant small vectors’ working time synchronously in sector and sector switching, which reduces the influence of dead time on neutral point potential balance. Finally, by analyzing the influence of dead time on the output average voltage, dead time compensation is applied, which reduces the difficulty of compensation implementation. Simulations and experiments verify the rationality of the optimized hybrid SVPWM-8H.

For three-level three-phase inverters, a novel finite-state machine-based model predictive control (FSM-MPC) is proposed. The state transition diagram of FSM is regarded as the operation guideline to avoid excessive voltage jumps (Δ V ). No more than five voltage vectors (VVs) are selected as FSM-MPC's candidates VVs, based on the VV reference and the previous optimal VV. And the cost function is simplified in terms of time duration. Thus, evaluating a candidate VV requires a very low computation cost. Without losing control performance, the proposed FSM-MPC's execution time is decreased to 50% of the MPC algorithm, which enumerates all basic VVs to get global optimal VV and performance. Compared with the existing simplified algorithms, the proposed FSM-MPC makes current harmonics lower, and its average switching frequency is 33% less, which means much less switching loss. Furthermore, the proposed algorithm is robust when the electric-circuit parameters are mismatched in the control system. Experimental results are provided to validate the advantages of the proposed algorithm.

With high real-time performance and a good filtering effect, the wavelet transform-based real-time filter is more commonly used in the fusion magnet power supply. Considering the signal characteristics of the fusion magnet power supply, the sliding time window method is widely used to satisfy the requirements of filtering efficiency. In this paper, an updated half-edge extension method is proposed to improve the boundary effect of sliding window data. Moreover, the hardware-based wavelet loopback structure is proposed to avoid the double computation of long sliding window data, thereby improving real-time performance and satisfying the demands of fusion magnet power supply filtering. The proposed half-edge extension method and the wavelet loopback structure are implemented by designed FPGA and experimented on the Experimental and Advanced Superconducting Tokamak (EAST) fusion magnet power supply system to verify the filtering efficiency. According to the experiment’s results, the proposed half-edge extension method has a better filtering effect, and the proposed wavelet loopback structure decreases the time by 67.4%, which can satisfy the requirements of the power supply system.