Search

As one of popular high-power power interface topologies between the energy storage devices and the dc link, the three-phase dual-active-bridge (3$\Phi$-DAB) dc–dc converter has received increasing attention. However, due to high number of semiconductor devices and multiple possible operation modes in the 3$\Phi$-DAB converters, the reliability and healthy operation become increasingly challenging compared with traditional single-phase DAB converters. This study presents detailed waveform analyses for 3$\Phi$-DAB converters under the normal and open-circuit faulty mode operation. Based on the analysis, a fast, cost-effective, reliable and sensor-less fault detection (FD) and fault isolation (FI) mechanism based on finite state machines for 3$\Phi$-DAB converters is presented in this article. Specifically, switch node voltages between the primary and secondary bridges of 3$\Phi$-DAB converters are adopted as the open-circuit fault (OCF) diagnostic signature and a low-cost OCF diagnostic circuit is utilized to perform the required signal processing. Using this technique, the open-circuit FD was achieved within 1/25th of the switching frequency and the open-circuit FI was performed within 1/6th of the switching frequency, which is the fastest FD and FI time for 3$\Phi$-DAB converters reported so far. Experimental results were acquired from a built prototype under various OCF scenarios to validate the effectiveness of the proposed technique.

The prediction of bifacial photovoltaic (bPV) system performance under variable conditions has persistently challenged researchers and practitioners alike, largely due to the unstable and imprecise irradiance measurements and the extensive training processes required for machine learning-based methods. Addressing these issues, this study introduces an innovative digital twin system that integrates a novel circuit-long short-term memory (LSTM) model with the newly proposed triangle-shading pattern estimation method, eliminating dependencies on direct irradiance measurements and historical data. Our approach uniquely combines the adaptability of LSTM networks with circuit models, facilitating real-time power prediction with unprecedented accuracy and efficiency. Comprehensive evaluations across various shading scenarios demonstrate the proposed model's superior performance, consistently reducing mean absolute error, mean squared error, and root mean squared error by over 50% compared with existing methods. This breakthrough offers a scalable, cost-effective solution for optimizing the deployment and management of bPV systems, marking a significant advancement in the field of photovoltaic research.

The control coupling and stability of three-phase photovoltaic systems with constant power loads (CPLs) are two key technical issues that urgently need to be addressed. However, current methods are difficult to solve both technical challenges simultaneously. To address these issues, this article proposes a novel decoupled continuous control set model predictive control (CCS-MPC) strategy for LC-filter T-type three-phase four-leg three-level voltage source inverters (3P4L-3L-VSIs) driving CPLs. The control strategy is developed in αβγ frame. The model of 3P4L-3L-VSIs is first transformed from abc frame to the αβγ frame to remove the coupling effects in the abc frame. Then based on the decoupled model in the αβγ frame, a novel CCS-MPC strategy is proposed to provide balanced sinusoidal voltages. The α and β components are regulated by the A, B, and C legs, while the zero-sequence component is regulated by the N leg. To confirm the effectiveness of the proposed strategy, both simulation and experimental results are presented to verify the proposed decoupled CCS-MPC strategy. The results show that the proposed control achieves low computational burden and total harmonic distortion and can increase the stability margin especially for CPLs compared with the commonly used PI-based control, PR-based control, and deadbeat control.

Voltage equalizers (VEs) are essential for partially shaded photovoltaic (PV) modules by equalizing the voltage of PV modules and preventing anti-paralleled diodes from bypassing the shaded PV modules, resulting in improved power yield under partial shading conditions (PSCs). Recently, different topologies for VEs have been discussed based on the distributed maximum power point tracking (DMPPT) principle at the module level. Considering that the power flow distribution, operation modes, and actual performance of these VEs show distinct differences, it becomes increasingly important for the performance quantification and comparative assessment of various topologies both theoretically and experimentally. Here, three typical differential-power-processing-based VEs are selected, including series-resonant-voltage-multiplier (SRVM), flyback-based PV-to-IP (Flyback-PV-IP), and flyback-based PV-to-Bus (Flyback-PV-Bus). Key performance indexes for VEs have been defined, including the processed power, power losses, and overall system efficiency. To quantify the performance of different topologies of VEs, an algorithm is developed in MATLAB with daily irradiation and temperature under various PSCs. Moreover, three experimental prototypes for the selected topologies have been built and main tests under different mismatching conditions have been conducted. With a systematic performance quantification and a fair comparison of typical VEs, this article will propose a systematic evaluation method for VE schemes. Meanwhile, the optimal VE topology with its control for typical PSC cases will be identified.

Recent power curtailment-based photovoltaic (PV) flexible power point tracking (FPPT) algorithms mainly adopted intricate curve fitting or sophisticated curve-scanning mechanisms to ensure the grid supportive functionalities under partial shading conditions (PSCs), showing the limitations of mathematical solidity or system dynamics improvement. Accordingly, a novel fast-speed partial-shading-tolerant FPPT (PST-FPPT) algorithm is proposed in this article. Regarding the proposed scheme, a modified explicit PV model is developed to express the key operation points with the assistance of several representative current-voltage samples from the initialization process, which is beneficial to computational burden reduction and irradiance sensors removal. Additionally, to guarantee the tracking speed to system dynamics, a set point estimation-based direct voltage regulation strategy is proposed in this article, eliminating the redundant searching in approaching the predefined power command. Simulation and experimental evaluations under various PSCs and operational circumstances validated the effectiveness of the proposed control.