Belhadj Souheyb Mohammed, Meliani Bouziane, Benbouhenni Habib, Colak Ilhami, Elbarbary Z M S, Al-Gahtani Saad F
Laboratory GIDD, University of Relizane, Relizane, Algeria.
Laboratoire LAAS, Department of Electrical Engineering, Ecole Nationale Polytechnique d'Oran, EL M'naouer, Bp 1523, Oran, Algeria.
Sci Rep. 2025 Apr 26;15(1):14631. doi: 10.1038/s41598-025-99551-2.
The conventional DC-DC architecture is often used in photovoltaic systems, with control based on incremental conductance-based maximum power point tracking (MPPT-IC) algorithms. Despite its simplicity and ability to ensure voltage balance across the output capacitors, this architecture suffers several drawbacks. These drawbacks cause undesirable problems such as high power ripples, overshoot, and limited dynamic response. Therefore, this paper proposes a three-level quadratic DC-DC boost converter as a suitable solution to replace conventional inverters in photovoltaic systems, while combined with an advanced MPPT method. The new approach is MPPT based on NARX neural network (NARAX-NN) algorithms. This proposed strategy is designed for high accuracy, robustness, and fast dynamic response compared to the MPPT-IC strategy. In this work, the MPPT-NARX-NN strategy of a three-level quadratic DC-DC boost converter is compared with several different strategies (MPPT-IC, MPPT based on type 1 fuzzy logic (MPPT-T1FL), and MPPT based on type 2 fuzzy logic (MPPT-T2FL)). This comparison is performed using MATLAB under different operating conditions. Simulation results indicate that the MPPT-NARX-NN approach significantly improves the operational performance of a three-level quadratic DC-DC boost converter compared to other strategies, increasing the reliability and deployment of photovoltaic systems. The numerical results show that the MPPT-NARX-NN strategy improves the rise time by 96.43, 97.34, and 94.50% compared to MPPT-IC, MPPT-T1FL, and MPPT-T2FL, respectively. Also, the settling time is improved by 50, 66.66, and 6.66% compared to MPPT-IC, MPPT-T1FL, and MPPT-T2FL, respectively. Furthermore, the strategy increases the average tracking efficiency (%) by 3.86 and 1.12% for MPPT-IC and MPPT-T1FL, respectively. These results highlight the effectiveness of the three-level quadratic DC-DC boost converter based on the MPPT-NARX-NN strategy in extracting energy, increasing performance and flexibility, and improving system reliability, making solar PV systems more efficient and a promising and indispensable solution.
传统的DC-DC架构常用于光伏系统，其控制基于基于增量电导的最大功率点跟踪（MPPT-IC）算法。尽管这种架构简单且能够确保输出电容器两端的电压平衡，但它存在几个缺点。这些缺点会导致诸如高功率纹波、过冲和动态响应受限等不良问题。因此，本文提出一种三电平二次DC-DC升压转换器，作为在光伏系统中替代传统逆变器的合适解决方案，并结合一种先进的MPPT方法。新方法是基于NARX神经网络（NARAX-NN）算法的MPPT。与MPPT-IC策略相比，该提出的策略设计用于实现高精度、鲁棒性和快速动态响应。在这项工作中，将三电平二次DC-DC升压转换器的MPPT-NARX-NN策略与几种不同策略（MPPT-IC、基于1型模糊逻辑的MPPT（MPPT-T1FL）和基于2型模糊逻辑的MPPT（MPPT-T2FL））进行比较。这种比较是在不同运行条件下使用MATLAB进行的。仿真结果表明，与其他策略相比，MPPT-NARX-NN方法显著提高了三电平二次DC-DC升压转换器的运行性能，提高了光伏系统的可靠性和可部署性。数值结果表明，与MPPT-IC、MPPT-T1FL和MPPT-T2FL相比，MPPT-NARX-NN策略分别将上升时间提高了96.43%、97.34%和94.50%。此外，与MPPT-IC、MPPT-T1FL相比，稳定时间分别提高了50%、66.66%和6.66%。此外，该策略分别将MPPT-IC和MPPT-T1FL的平均跟踪效率（%）提高了3.86%和1.12%。这些结果突出了基于MPPT-NARX-NN策略的三电平二次DC-DC升压转换器在提取能量、提高性能和灵活性以及提高系统可靠性方面的有效性，使太阳能光伏系统更高效，成为一种有前途且不可或缺的解决方案。
