A dual-active-bridge DC-DC converter is taken as the research object, and a triple-phase-shift (TPS) control strategy with an optimization objective of reducing the backflow power and inductance peak current is proposed. Through the analysis of characteristics in six operating modes with TPS, the input-side zero-backflow power mode and bilateral zero-backflow power mode with smaller inductance peak current in Boost and Buck modes are obtained, and the constraint of phase-shift ratio and power range in different operating modes are given. The inductance peak current in bilateral zero-backflow power mode is smaller, and the zero switching of the power device can be realized. On this basis, the working point with the smallest inductance peak current at different values of transmission power is further obtained through optimizations using the KKT method. Finally, the proposed TPS optimization method and the inductance peak current optimization method under extended-phase-shift were verified by experiments. Experimental results show that the proposed optimization strategy can eliminate the backflow power, reduce the current stress and improve the conversion efficiency, thus verifying the theoretical analysis.

Aimed at the problems of insufficient dynamic performance and poor robustness in the traditional linear control, a dynamic response optimization method based on the model predictive control of a multi-phase interleaved Buck converter is proposed in this paper. First, the state space model of the interleaved Buck converter is optimized, and a virtual impedance method is put forward to realize the decoupling control of each phase converter. Second, a load disturbance observer and a continuous set model predictive controller based on one-step prediction are designed. The reference current is calculated based on the load current observation and the output voltage deviation, and the optimal duty cycle in each phase is obtained by substituting it into the predictive model with an objective of minimizing the current deviation. Finally, a simulation platform was built on Matlab/Simulink for simulation verification, and experiments were conducted on a three-phase Buck converter. Simulation and experimental results show that the proposed algorithm can effectively suppress the output voltage variations caused by load disturbances, improve the dynamic performance of the system, and ensure its robustness at the same time.

Aimed at the problems in traditional on-board charging DC-DC converters such as a narrow soft switching range, poor dynamic response and rectifier diode voltage spike, a novel wide-range soft switching full-bridge converter is designed, and a double-closed-loop control strategy combining variable universe fuzzy PI with model prediction is adopted. First, the working mode of the novel full-bridge converter is analyzed, and its simplified state space average model is established. Second, model predictive control is used in the current loop, and compensation improvements are made to obtain the predictive expression of inductance current at the time point of k+2, thereby improving the large time delay of an actual system. Third, the control of voltage loop is changed to the transform domain fuzzy PI control with fuzzy universe self-adaptation, so as to optimize the system adaptability and improve the control precision. Simulation results show that the soft switching range of the novel converter is extended to 30% of load, and the rectifier diode voltage spike is also eliminated. Under sudden changes in load, the proposed control strategy reduces the output voltage regulation time by 17.8% and 25.6%, as well as 12.9% and 9.1%, respectively. Finally, the effectiveness of the proposed topology and control strategy was verified by experimental results.

To improve the dynamic performance of passivity-based control (PBC) of a Boost converter with constant power load under external disturbances, a passivity-based adaptive control strategy based on neural network was proposed. Under this strategy, a passivity-based adaptive control structure with proportional and integral terms was adopted, and a single-neural network was used to optimize the injection damping, proportional coefficient and integral coefficient in real time. In addition, a power observer was used to observe the power of unknown constant power load (CPL) in real time, thus improving the dynamic and steady-state performances of the Boost converter when the CPL changes. Compared with PI+PBC and PI double-closed-loop control strategies, the passivity-based adaptive control based on neural network can make the Boost converter have a better performance. The results of computer simulations show that the proposed control strategy is feasible.

The input-series output-parallel combination technology for DC-DC converters is widely applied in high-power and high-voltage fields. To realize an average operation after the expansion and connection of modules, the characteristics of input voltage-sharing and output current-sharing after the series-parallel combination of module power supplies are studied, and a control strategy for module current-sharing on the output side is proposed, which consists of an output voltage loop and an output current-sharing loop. The common output voltage loop of all the modules samples the output voltage from the combined system for feedback regulation, while the output current-sharing loop of each module samples the output current from the module for average control. All the sampling and control chips under the proposed control strategy are on the low-voltage output side, which can eliminate an isolation unit and improve the system reliability. A half-bridge LLC resonant topology was selected as a sub-module to build a prototype, and results verified the effectiveness of the proposed output current-sharing control method.

In view of the poor tiansient response of the output-capacitorless, an output-capacitorless low dropout (LDO) regulator with fast transient response is proposed, and the voltage overshoot and undershoot suppression circuits are used to suppress voltage jumps, thus improving the transient response characteristics of the LDO regulator. An improved error amplifier is designed, whose gain is larger than that of a general error amplifier in static working states. In addition, the transient output current from the designed error amplifier can be increased when the output voltage jumps, and the charging and discharging of the power tube can be accelerated. By using Current buffer Miller compensation, the phase margin of the circuit is compensated with a smaller capacitive area. Simulation results show that when the voltage range is 2.8~3.3 V, the output voltage from the circuit is 1.8 V and the maximum load current is 50 mA. When the load current jumps between 1 and 50 mA, the undershoot amplitude of output voltage is reduced by 46 mV (51%) and the response time is 1.16 µs under the effect of the transient enhancement circuit. Meanwhile, the overshoot amplitude is reduced by 29 mV (35%), and the response time is 1.18 µs.

Under the non-ideal conditions of grid such as frequency fluctuation due to the access of various new energy to power system, the quality of grid-connected current of a grid-connected converter will be seriously degraded under the traditional repetitive controller with fixed parameters. Therefore, it is important to improve the frequency self-adaptability of the grid-connected converter under non-ideal grid conditions. To solve this problem, a frequency self-adaptive and selective-order harmonic repetitive controller (FSA-SHRC) is proposed, which can accurately approximate and compensate the part of parameter transformation due to fundamental frequency fluctuation by combining the internal mode with an infinite impulse response (IIR) filter. Compared with the conventional repetitive controller, the proposed FSA-SHRC has advantages of fewer memory cells required by the internal mode, a better error convergence rate and an accurate performance, and it has certain frequency adaptive performance by tuning the coefficients of IIR online when the fundamental frequency fluctuates. FSA-SHRC was applied to a three-phase grid-connected inverter control system, and the superiority of the theory and the corresponding adaptability under non-ideal grid conditions were verified.

To solve the problems of slow dynamic response speed and poor robustness in the conventional voltage and current dual-closed-loop control for grid-connected inverters based on virtual synchronous generator (VSG) control, a novel dual-closed-loop control strategy for VSG based on the differential flatness theory was proposed. First, a mathematical model under the dq coordinate axis was established according to the topology of a three-phase grid-connected inverter, and the flatness of the system was proved based on the differential flatness theory. Second, a voltage and current differential flatness controller including feedforward reference locus and error feedback compensation was designed. Finally, a simulation model of the system was constructed in MATLAB/Simulink. Simulation results show that the proposed control strategy improves the dynamic response speed and robustness of the VSG-based grid-connected inverter under the operating conditions of power disturbance and three-phase short-circuit fault of the grid, thereby verifying the effectiveness and feasibility of the differential flatness control strategy when it is applied to VSG.

With the higher requirements for the safety and stability of traditional power grids, the energy crisis and environmental pollution problems are increasing. As a result, “Green” and “Low carbon” have become the urgent needs of the development of power systems. Based on the current situation, a multi-functional photovoltaic grid-connected inverter with power quality control is proposed. First, according to an improved adaptive harmonic detection method, the harmonic component in the grid-connected current is detected. Second, the command current is tracked and controlled by quasi-proportional resonance. Through the subjective and objective criteria importance through intercriteria correlation (CRITIC) game theory combination weighting method, a comprehensive evaluation model of power quality is established to evaluate the harmonic and reactive components. Third, considering the limited capacity of the inverter in practice, a capacity matching coefficient is formulated to compensate the dominating problems. Finally, simulation results show that the harmonic and reactive components are optimized and compensated in the grid-connection process of the inverter, which improves the power quality at the grid-connected point and greatly increases the utilization rate of the inverter. Therefore, the effectiveness and feasibility of the proposed strategy are verified.

To solve the problems of computational complexity and exponential amplification of measurement noise under the traditional backstepping control strategy applied to an LCL grid-connected inverter, an improved backstepping control strategy based on weighted average current control (WACC) is proposed in this paper. First, the order of the LCL filtering system is reduced from third-order to first-order by means of the WACC strategy, which effectively alleviates the overall computation load and decreases the number of derivations required for virtual control quantities. Then, a sliding-mode observer is utilized to observe the weighted average current, so as to lessen the influence of LCL filter parameter perturbation on the accuracy of weighted average current. Finally, a backstepping controller is designed to regulate the grid-connected current based on the established equivalent system model. Simulation and experimental results prove that the proposed control strategy has a good steady-state performance and high robustness.