The capacitor-current feedback active damping is widely applied to attenuate the LCL resonance. Due to the control delay, the virtual impedance equivalence of active damping may exhibit a negative-damping effect, challenging the system stability. To address this issue, the feedback function should be well selected and designed according to the LCL resonance frequency. On the basis, a general graphical evaluation method is proposed, which can intuitively identify the positive-damping frequency ranges of different feedback functions and thus provide guidance for selecting a reasonable active damping feedback function. Using the proposed method, a phase-lag compensator with strong robustness against the fluctuation of LCL resonant frequency is put forward. Experiments were carried out on a 6 kW single-phase LCL grid-connected inverter, and results verified the theoretical analysis.

Owing to their advantages such as high efficiency and simple control, Swiss rectifiers show promise in the field of electric vehicle charging. Under conditions of unbalanced power grid, the use of traditional control methods will lead to the generation of even-order harmonics on the DC side and odd-order harmonics on the AC side of a Swiss rectifier, thereby affecting the system operation. On the basis, two control methods for the Swiss rectifier under unbalanced power grid conditions are introduced, i.e., constant output voltage control (COVC) and sinusoidal and symmetrical current control (SSCC). With COVC, the DC voltage remains stable at the cost of negative-sequence current emerging in the input current, which results in a phase difference between the grid voltage and current of the Swiss rectifier. Due to constraints on the topology, this phase difference may lead to grid current distortion. SSCC ensures a symmetrical and sinusoidal current, but significant fluctuations in the DC voltage may exist. To solve these problems, a coordinated control method of suppressing current distortion control (SCDC) is proposed, which can realize sinusoidal input current throughout the entire grid voltage drop range while minimizing the fluctuations of active power. The effectiveness of the SCDC method was verified by experimental results.

To address the challenges of a modular multilevel converter (MMC) without redundant submodules when using the traditional PI control method for fault-tolerant control, which include excessive control parameters, complex structure and slow dynamic responses, a model with a fixed number of calculations for implementing the fault-tolerant control by fault prediction is proposed. First, the AC current is optimized. Then, multi-objective optimization on the circulating current and the total voltage of submodule capacitors is performed, and the balance of submodule voltage is realized based on the direction and level difference of bridge arm current. With only eight times of rolling optimization, the optimal solution can be found under this strategy, significantly reducing the computational requirements. In addition, this algorithm not only can ensure that the MMC operates stably and has high-quality output under normal conditions, but also can maintain its stable operation when the submodule experiences an asymmetric fault. Simulation tests were completed in MATLAB/SIMULINK, and a five-level experimental platform was built to verify the simulation and test results.

In practical engineering, the issues of large equipment volume and high operating costs increasingly highlight the demand for lightweight modular multilevel converters (MMCs) in new power systems. The arm multiplexing MMC (AM-MMC) can improve the utilization of submodules and effectively achieve lightweight MMC. However, due to its complex topology and susceptibility to fluctuations in internal energy caused by the changes in operating conditions, the energy fluctuations are more severe and even prone to reuse errors, especially in the event of AC system failures, increasing the requirements for the AC fault ride-through capability of AM-MMC. For the three-phase short-circuit fault of the AC system under the most serious fault condition, the fault characteristics of the converter after the fault are analyzed, and an optimal configuration strategy for the fault capacitor based on a fast blocking mechanism is proposed. During the fault period, the multiplexing mode adjustment and the fast blocking are carried out simultaneously to realize the fault ride-through of the AM-MMC after the three-phase short-circuit of the fault ride-through AC system. Finally, to validate the proposed strategy, a singleended simulation model was constructed using MATLAB/Simulink. Simulation results demonstrate that re-locking the converter after switching the multiplexing mode can effectively resolve the three-phase short-circuit fault ridethrough issue in the AM-MMC.

The modular multilevel matrix converter (M3C) has a cascaded modular structure, which is suitable for scenarios with medium- and high-voltage, high power and wide frequency ranges such as driving high-voltage high-power motors and flexible low-frequency transmission systems. However, due to the large number of system components and the presence of floating capacitors, the difficulty in controlling the M3C is significant, especially in terms of voltage equalization and ripple suppression of numerous floating capacitors in the system. A strategy based on bridge arm energy control is proposed, which is intuitive and simplifies the number of control layers. The goal of directly controlling the energy of each bridge arm in the converter is mainly achieved by setting the energy reference value of the bridge arm and controlling the bridge arm energy at present in real-time in the form of negative feedback. Aimed at each cascaded submodule within the bridge arm, a voltage equalization controller is designed. Finally, the feasibility of the proposed control strategy was verified through simulations.

Since the inverter-interfaced distributed generations (IIDGs) in microgrids exhibit special output characteristics, traditional protection schemes are no longer suitable for microgrids. Simultaneously, the idea of “Signal-injected active protection” introduces a novel perspective on microgrid protection. However, the shunt problem of non-injection source results in a poor protection performance. Under this background, in light of the idea of control-protection coordination, an active protection method for microgrids based on harmonic component injection is proposed. First, the frequency response characteristics of IIDG are analyzed. Then, the controllers of injection source and non-injection source are designed, and the corresponding start-up criteria are established. When a fault occurs, the injection source injects the characteristic signal of harmonic component with constant amplitude into the microgrid, while the non-injection source limits the shunt of the characteristic signal, so that the characteristic frequency equivalent network of the microgrid exhibits a single-source structure. Simulation results show that the proposed strategy can reliably start the harmonic component injection and solve the shunt problem of the characteristic signal, providing a solution for signal-injected microgrid protection.

As the proportion of photovoltaic installed capacity continues to increase and the low inertia problem inpower systems becomes increasingly prominent, how to evaluate the inertia supporting capacity of a photovoltaic grid-connected system and how to set the frequency modulation parameters of the system to meet the requirements of grid-connected guidelines are the key issues to be solved urgently. The equivalent inertia of a typical grid-connected photovoltaic system with additional frequency modulation control is quantitatively evaluated by analogizing the rotor motion equation of synchronous generators. Then, considering the physical constraints of photovoltaic equipment and the requirement of new energy grid-connected guidelines, the frequency modulation control parameter setting method for a centralized photovoltaic grid-connected system is put forward. Furthermore, based on the physical constraints and grid-connected guideline requirements, a target of minimum network line loss is introduced additionally, and the frequency modulation control parameter setting method for a distributed photovoltaic grid-connected system is formulated. Finally, the simulation results based on MATLAB/Simulink verify the correctness of frequency modulation control parameter setting.

Due to its advantages of stable power and low cost, the capacitive energy extraction mode is suitable for local power supply scenarios of pole switches in medium-voltage distribution networks. The output characteristics of a capacitive energy extraction circuit are often in an approximate constant-current source range, and the primary voltage is required to normally supply power in the range of 5.7–42 kV under different operating conditions in the meantime, so voltage regulation measures must be taken to provide stable voltage for the load. First, based on the compensation characteristics in the capacitive energy extraction mode, a secondary reactive power compensation mode is proposed to make up for the difference between the energy extraction power and load consumption. Second, a voltage modulation dynamic reactive power compensation scheme based on PWM is put forward, so that the dynamic reactive compensation is realized through dynamic equivalent reactance. Third, a waveform modulation dynamic reactive power compensation scheme is constructed, and the transformer voltage waveform is actively controlled through the H-bridge of the third winding, thus realizing the reactive power dynamic voltage stabilization in the capacitive energy extraction mode. Under the proposed voltage stabilization and regulation scheme, the stable AC voltage output under low heat generation in the capacitive energy extraction mode is realized.

In the context of new energy era, aimed at the hot issue of the impact of electric vehicle charging loads on power grid, a short-term charging load forecasting method for electric vehicles based on a combined long short-term memory network and graph neural network (LSTM-GNN) model is proposed to address the shortcomings of existing models. First, three typical daily loads are correlated using the Copula theory, a feature input matrix containing the main factors and historical loads of charging stations is constructed, and a sliding window data set is established. Second, LSTM and GNN are used for feature extraction and graph data construction of the feature input matrix, and the GNN layer completes the feature aggregation and outputs the predicted load value and error evaluation indexes. The results of an example show that the proposed model can effectively extract the key features of charging load at charging stations in different regions, and it has better forecasting performance and generalization capability than traditional models.

The multi-pulse rectifier technology has advantages such as simple realization, low cost and high reliability, and it has been widely applied in high-power rectifier systems. Based on the topology of a traditional D-type 18-pulse autotransformer, a triangle is added to connect the primary side to form an isolated 18-pulse rectifier transformer, which is further used in a DC distribution system. Its circuit, working principle, output DC voltage and input current harmonics are analyzed theoretically in detail. In addition, the software Ansys-Maxwell is used to perform equivalent modeling of the transformer, and the reliability of simulation results is verified by a comparison with the theoretical analysis. Two methods of leakage flux method and short-circuit experiment are used to solve the primary-side leakage inductance of the transformer. The input current harmonics of the transformer are analyzed in depth, and the influence of transformer leakage inductance on the suppression of input current harmonics is also discussed.