Scholarlink

<p>Partial shadowing (PS) and changing environmental factors like temperature and irradiation cause performance deterioration in photovoltaic (PV) systems. These dynamic elements can produce both local and global maximum power points (GMPP and LMPP), hence producing several peaks in the power-voltage characteristics of PV modules. This study proposes a revolutionary approach to dynamically monitor the MPP across varying irradiation and temperature conditions, thereby enhancing the performance of PV systems. The proposed method aims to improve energy generation by continually adapting to shifts in environmental conditions, ensuring efficient power extraction from the system. The integration of a deep convolution neural network (CNN) with the modified coati optimization algorithm (MCOA) for maximum power point tracking (MPPT) is a main contribution of this work. The deep CNN-based model, developed through a multi-criteria optimization technique, is designed to optimally adjust its weights. This technique adheres to the GMPP while addressing time-varying PS patterns, facilitating rapid responses to environmental fluctuations. The proposed method demonstrates superior performance, exhibiting improved tracking speed, increased efficiency, and reduced steady-state oscillations compared to conventional MPPT techniques, as evidenced by simulation results. Furthermore, the study illustrates a 2 kW PV system designed for integration with a 3-phase conventional grid. The system includes a 3-phase, 2-level pulse width modulation voltage-source inverter (PWM-VSI) equipped with an LC filter, a PV array, and a DC/DC converter, along with a constant current hysteresis controller (CCHC) integrated with a phase-locked loop (PLL). Simultaneously, the CCHC ensures that the inverter output remains aligned with the grid voltage. Therefore, reducing total harmonic distortion (THD) in current and voltage, MATLAB/Simulink simulations corroborate the suggested strategy, demonstrating precise maximum power point tracking amid variations in irradiation and temperature. Evaluating the proposed MPPT methodology against existing methods, including the MCOA and the coati optimization algorithm (COA), reveals substantial enhancements in tracking precision and efficiency. The simulation results indicate that employing a hybrid deep CNN with the MCOA methodology can improve PV system performance, especially in challenging climatic circumstances.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Deep CNN-optimized MCOA for maximum power point tracking of PV systems connected to a 3-phase grid using constant current with hysteresis controller

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月5日

查看全文
  • Radhika Guntupalli,
  • S. K. B. Pradeep Kumar Ch,
  • Prabhakara Sharma P

摘要

Partial shadowing (PS) and changing environmental factors like temperature and irradiation cause performance deterioration in photovoltaic (PV) systems. These dynamic elements can produce both local and global maximum power points (GMPP and LMPP), hence producing several peaks in the power-voltage characteristics of PV modules. This study proposes a revolutionary approach to dynamically monitor the MPP across varying irradiation and temperature conditions, thereby enhancing the performance of PV systems. The proposed method aims to improve energy generation by continually adapting to shifts in environmental conditions, ensuring efficient power extraction from the system. The integration of a deep convolution neural network (CNN) with the modified coati optimization algorithm (MCOA) for maximum power point tracking (MPPT) is a main contribution of this work. The deep CNN-based model, developed through a multi-criteria optimization technique, is designed to optimally adjust its weights. This technique adheres to the GMPP while addressing time-varying PS patterns, facilitating rapid responses to environmental fluctuations. The proposed method demonstrates superior performance, exhibiting improved tracking speed, increased efficiency, and reduced steady-state oscillations compared to conventional MPPT techniques, as evidenced by simulation results. Furthermore, the study illustrates a 2 kW PV system designed for integration with a 3-phase conventional grid. The system includes a 3-phase, 2-level pulse width modulation voltage-source inverter (PWM-VSI) equipped with an LC filter, a PV array, and a DC/DC converter, along with a constant current hysteresis controller (CCHC) integrated with a phase-locked loop (PLL). Simultaneously, the CCHC ensures that the inverter output remains aligned with the grid voltage. Therefore, reducing total harmonic distortion (THD) in current and voltage, MATLAB/Simulink simulations corroborate the suggested strategy, demonstrating precise maximum power point tracking amid variations in irradiation and temperature. Evaluating the proposed MPPT methodology against existing methods, including the MCOA and the coati optimization algorithm (COA), reveals substantial enhancements in tracking precision and efficiency. The simulation results indicate that employing a hybrid deep CNN with the MCOA methodology can improve PV system performance, especially in challenging climatic circumstances.