Ideally, the inner (the upper or lower arm) current of a modular multilevel converter (MMC) is ideally assumed to be the sum of a dc component and an ac component of the fundamental frequency. However, as ac current flows through the submodule (SM) capacitors, the capacitor voltages fluctuate with time. Consequently, the inner current is usually distorted and the peak/RMS value of it is increased compared with the theoretical value. The increased currents will increase power losses and may threaten the safe operation of the power devices and capacitors. This paper proposes a closed-loop method for suppression of the inner current in an MMC. This method is very simple and is implemented in a stationary frame, and no harmonic extraction algorithm is needed. Hence, it can be applied to single-phase or three-phase MMCs. Besides, this method does not influence the balancing of the SM capacitor voltages. Simulation and experimental results show that the proposed method can suppress the peak and RMS values of the inner currents dramatically.

In this paper, an improved pulse width modulation (PWM) method for chopper-cell (or half-bridge)based modular multilevel converters (MMCs) is proposed. This method can generate an output voltage with maximally 2N+1 (where N is the number of submodules in the upper or lower arm of MMC) levels, which is as great as that of the carrier-phase-shifted PWM (CPSPWM) method. However, no phase-shifted carrier is needed. Compared with the existing submodule unified pulse width modulated (SUPWM) method, the level number of the output voltage is almost doubled and the height of the step in the staircase voltage is reduced by 50%. Meanwhile, the equivalent switching frequency in the output voltage is twice that of the conventional SUPWM method. All these features lead to much reduced harmonic content in the output voltage. What is more, the voltages of the submodule capacitors can be well balanced without any close-loop voltage balancing controllers which are mandatory in the CPSPWM schemes. Simulation and experimental results on a MMC-based inverter show validity of the proposed method.

In this paper, the influence of one-sample delay for sampling and computation in digital control on the bandwidth of the inner current loop of a 400-Hz ground power unit (GPU) is analyzed first. The results show that it is difficult and even impossible for high-power 400-Hz GPUs to maintain low total harmonic distortion content in the output voltage with the conventional proportional-integral-based double-loop control. To improve the performance, resonant controllers with parallel structure which are widely used in active power filters are applied to the single-loop control of the 400-Hz GPU. The parameter design criterion for the parallel resonant controllers is discussed in the discrete time domain. Meanwhile, adoption of proportional gain in the single-loop control is investigated. The results show that it can improve the performance little and may cause instability problems. Comparisons between different control methods for the 400-Hz GPU are also made, and the single-loop control method in this paper seems to be the most suitable one in terms of simplicity and performance. Experiments on a 16-b fixed-point DSP-controlled 90-kVA 400-Hz GPU prototype show satisfactory results of the single-loop method feeding linear/nonlinear and balanced/unbalanced loads.

In the last decade, many methods for control of three-phase boost-type pulsewidth modulation (PWM) rectifier under unbalanced input voltage conditions have been studied and presented. These methods use the sequential components of input voltages, pole voltages, and input currents of the rectifier to analyze and control the instantaneous powers, of which dual (positive-sequence and negative-sequence rotating) frame control is the most common structure. Anyway, this paper analyzes the instantaneous powers of the PWM rectifier in two-phase stationary frame. Based on this analysis, the input-power-control, input-output-power-control, and output-power-control methods for PWM rectifier under unbalanced voltage conditions are proposed in single stationary frame. Compared with the existing methods, simplicity may be the major advantage of the method in this paper. Rotating transformation and phase detection of the input voltages are both eliminated. Moreover, sequential component extraction of the control variables is also not necessary, leading to much reduced time delay to the control system. Experimental results on a 9-kVA PWM rectifier show validity and effectiveness of the proposed methods.

This letter focuses on the power module (PM) capacitor voltage balancing of a $\pm 350$ -kV/1000-MW modular multilevel converter (MMC). For this MMC, the current flowing through the IGBTs will exceed the nominal current of the adopted commercial high-voltage (3.3 kV or higher) IGBT products, i.e., 1500 A. As for the PM voltage balancing of this MMC, the switching frequency must be kept low to reduce the losses and the accurate calculation of the PM switching frequency is of great importance to the design of the cooling system. This letter presented a low-switching frequency PM balancing method by introducing the balancing adjusting number (BAN). The PM switching frequency calculation method is also presented with different BANs. Computer simulation and the real-time control hardware in the loop test results based on RT-LAB show the effectiveness of the proposed PM balancing method and the accuracy of the switching frequency calculation method under the rated power.

For high power voltage source converter based high voltage direct current transmission systems, modular multilevel converter (MMC) is the most popular circuit topology nowadays. It is known that the existence of the second-order harmonic circulating current (SHCC) component in the arm current of MMC increases the loss of the converter. Conventionally, the SHCC component eliminating methods are usually adopted to reduce the loss. The relationship between the loss of MMC and the SHCC component is analyzed first in this paper. The analysis shows that the conventional eliminating methods are not optimal in terms of the minimum loss of MMC. A loss optimization control scheme is also proposed by injecting the optimal SHCC component into the arm current of MMC, rather than eliminating the SHCC component of the arm current. The simulation and experimental results of the ±350 kV/1000 MW MMC and the down-scaled MMC setup verify the correctness and effectiveness of the proposed loss optimization method under different operation conditions.