A non-isolated multiport DC-DC converter with ultra-high gain is proposed to meet the high gain DC-side conversion requirements of voltage sag control equipment, which integrates two unidirectional input sources, one bidirectional energy storage unit and output load. This converter has three different operating modes depending on the operation of the energy storage unit, and all of them can achieve ultra-high voltage gain. In particular, the converter can still operate reliably when the input of any one of the input sources is zero. In addition, the input current of multiport in the converter is continuous, and the input and output ports are common grounded, which helps to reduce electromagnetic interference in the system. A detailed analysis was performed on the operating principle, steady-state characteristics and performance comparison of the converter, moreover, a 250 W experimental prototype was built to verify the feasibility and effectiveness of the proposed converter.

Aimed at the problem that the three-level voltage space vector pulse width modulation method will cause the generation of common-mode voltage (CMV) in an inverter, a simplified modulation method that can reduce the CMC is proposed in this paper. First, the three-level voltage space vector diagram is simplified to a two-level space vector diagram by coordinate translation. Second, according to the principle of generating CMV by switching states, suitable vectors are selected to adjust the switching sequence of space vectors. Finally, the unified expression of modulated wave is obtained by derivation. A comparison with experimental results shows that the proposed method can effectively reduce the computational load and limit the CMV to 1/6 of the DC-side voltage.

With the development of renewable energy, DC microgrids are receiving increasing attention. Dual active bridge (DAB) converters have advantages such as zero voltage conduction and can achieve a high control efficiency, thus playing an important role in DC microgrids. To address the issue of single phase shift control which causes some switching tubes to lose their capability of conducting at zero voltage when the voltage transmission ratio is mismatched, a multi-mode switching method was proposed to achieve zero voltage turn-on across the entire load range. First, the expressions for the effective values of transmission power and inductance current in the internal and external modes of extended phase shift were analyzed and derived. Second, the optimization curves of phase shift ratio and transmission power were obtained through the Lagrange extremum method to ensure that the DAB converter can satisfy switching under full-load range while achieving soft switching in the full range. Finally, a hardware-in-the-loop platform was built to verify the feasibility and superiority of the optimized modulation strategies.

Aimed at the problem that the performance of model predictive control will decrease significantly when the predictive model is not accurate, a dynamic compensation strategy based on an extended state observer (ESO) is proposed, which is on the basis of voltage model predictive control. This strategy can not only improve the response speed and accuracy of the inverter system, but also suppress the influences of parameter perturbation and disturbance. First, the inverter state space equation of a LC filter is established, and the influences of parameter perturbation and current disturbance on the output voltage from the inverter system are analyzed. Next, a voltage model predictive controller is designed based on the nominal model of the inverter system. Then, an ESO is established through the local input and output information, so as to accurately estimate the internal and external disturbances of the system. In addition, a reverse compensation controller is designed, thus effectively suppressing the output voltage sag, swell and distortion of the inverter. Finally, this strategy was verified on an actual inverter platform PEK-530A. Experiments with various load disturbances and parameter perturbations under different working conditions were designed to validate the effectiveness and correctness of the proposed strategy.

As a mainstream high-efficiency DC-DC converter, the LLC resonant converter faces an issue of a narrow voltage regulation range, and a hybrid modulation technique combining pulse frequency modulation (PFM) and phase shift modulation (PSM) can effectively broaden the voltage regulation range. However, a single proportional integral (PI) controller cannot adequately balance the system stability and responsiveness in the two operating modes. Aimed at this problem, a hybrid modulation adaptive control strategy for the LLC converter is proposed. First, a small-signal model of the LLC converter in PFM and PSM modulation modes is established using an extended describing function method. Then, a model correction algorithm is developed with an objective of unanimous rate of change of the model gain, so as to obtain a unified model of the LLC resonant converter. Furthermore, based on this unified model, adaptive PI controllers are designed considering input voltage and load current, thus realizing the stable performance and dynamic response of the converter across the entire load and input ranges. Finally, the effectiveness of the proposed adaptive control strategy was validated by the construction of a simulation model and an experimental platform.

To solve the problem of reflux power and switching tube loss in a neutral point clamped three-level dual active bridge (DAB) DC-DC converter, an optimal control strategy for reducing the reflux power in the zero-voltage switching (ZVS) boundary and range is proposed. Under the premise of realizing ZVS for all the switching tubes, this strategy can reduce the reflux power. First, the working principle for the three-level DAB converter under dual phase-shift (DPS) control is analyzed, and the ZVS range is obtained. Then, the mathematical models of transmission power and reflux power are established, and the reflux power is optimized within the ZVS boundary and range, respectively. Finally, the effectiveness of the proposed control strategy was verified by simulations and an experimental platform.

For a neutral-point-clamped three-level converter with only one zero-level switching state, there exists a problem of unbalanced loss distribution in internal and external tubes, which seriously limits the converter capacity and power density. In comparison, an active neutral-point-clamped three-level converter with multiple redundant zero-level switching states has the capability of flexibly adjusting the loss distribution. All the zero-level switching states of the active neutral-point-clamped three-level converter that can ensure the normal operation of the converter are summarized, the commutation process in different zero-level switching states are analyzed in detail, and the loss distribution rule in different commutation modes is summed up. Then, the correctness of the loss distribution rule in each commutation mode was verified by simulation and experimental results.

Bidirectional DC-DC converters are widely applied in the fields such as distributed generation systems, converter development and testing, and electric vehicles. Owing to their advantages including a wide gain, a high efficiency, and input and output isolation, bidirectional LLC resonant converters are widely used in DC bidirectional DC-DC converters. Based on the topology of a three-phase bidirectional LLC resonant converter, a loss model is established, and the efficiency characteristics of the converter are analyzed in depth. Meanwhile, with the consideration of the efficiency and gain indexes of the converter, a resonant parameter design method for optimizing the efficiency characteristics is proposed. A prototype of a 100 kW three-phase bidirectional LLC resonant converter was built, and the effectiveness and feasibility of the proposed method were verified by experimental results.

In response to the demand for fully domestic and high-power-density modular power supplies in aviation/vehicle applications, a grid-side current harmonic suppression and digital control strategy optimization method for wide-frequency three-phase power factor correction (PFC) power modules is proposed, which aims to address the challenges of harmonic suppression under wide-frequency input conditions and improve the controller’s resource utilization efficiency. By establishing a d-axis equivalent model of a three-phase PFC system in a rotating coordinate system, a theoretical basis for indirect current control is provided. In addition, a proportional integral resonant (PIR) regulator is introduced into the DC current loop, so as to effectively enhance the harmonic suppression performance. At the embedded architecture level, an interrupt nesting mechanism and a parallel computing architecture using central processing unit/control law accelerator (CPU/CLA) are adopted, significantly increasing the execution frequency of CLA tasks. Experimental results validate the effectiveness of the proposed digital control method in harmonic suppression and resource utilization optimization.

Aimed at the multi-peak characteristics of P-U output characteristic curves when a photovoltaic array is partially shaded, as well as the problem that the traditional maximum power point tracking (MPPT) method is easy to fall into local optimum and difficult to strike a balance between its tracking accuracy and speed, an MPPT algorithm combining the improved Harris hawk optimization (IHHO) algorithm and variable-step incremental conductance (INC) method is proposed. This algorithm divides the peak region to determine the voltage range where the extreme point exists, so as to narrow the search scope of the algorithm, and then applies the IHHO algorithm for global search in this range. First, the IHHO algorithm optimizes the initial position near 0.8 times the open-circuit voltage in each interval sequentially. Then, the Levy flight is used to optimize the scaling factor at the exploration stage. After the algorithm converges, it is switched to the variable-step INC method for local development. Finally, four kinds of illumination conditions are set up for verification. Simulation results show that the proposed composite MPPT algorithm can effectively avoid falling into local traps under complex shading, low illumination and illumination mutation conditions. In addition, the search speed is improved by 50% on average.