Scholarlink

This research paper comprehensively analyzes a solar photovoltaic (PV) system integrated with battery storage, focusing on the design and simulation aspects. The research implements artificial neural network (ANN)-based maximum power point tracking (MPPT) using two algorithms: Bayesian Regularization (BR) and Levenberg–Marquardt (LM). A comparative analysis of these algorithms is conducted to optimize energy harvesting from the PV system. The ANN-based MPPT system is designed in a MATLAB/Simulink environment and trained using a dataset of 1000 samples, incorporating solar irradiance, temperature, current, and voltage data. The trained dataset achieves nearly zero mean squared error after 1000 epochs, with regression values of 0.95 and 0.99 for BR and LM algorithms, respectively. A bidirectional DC-DC converter is integrated into the PV battery system to ensure efficient power transfer. The system is designed to manage fluctuating solar irradiance conditions and battery charging/discharging cycles. Simulation results confirm the system's effectiveness in maintaining performance under variable environmental conditions. This research advances the optimization of PV systems integrated with energy storage, paving the way for more reliable and efficient renewable energy solutions while providing critical insights for academics and professionals in solar energy and power systems engineering.

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

Intelligent Energy Exchange Prediction Between Solar PV Systems and Energy Storage Using AI for Enhanced Performance

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • A. Aswini,
  • P. Sivakumar,
  • P. Arokiya Prasad,
  • M. Kaleeswari

摘要

This research paper comprehensively analyzes a solar photovoltaic (PV) system integrated with battery storage, focusing on the design and simulation aspects. The research implements artificial neural network (ANN)-based maximum power point tracking (MPPT) using two algorithms: Bayesian Regularization (BR) and Levenberg–Marquardt (LM). A comparative analysis of these algorithms is conducted to optimize energy harvesting from the PV system. The ANN-based MPPT system is designed in a MATLAB/Simulink environment and trained using a dataset of 1000 samples, incorporating solar irradiance, temperature, current, and voltage data. The trained dataset achieves nearly zero mean squared error after 1000 epochs, with regression values of 0.95 and 0.99 for BR and LM algorithms, respectively. A bidirectional DC-DC converter is integrated into the PV battery system to ensure efficient power transfer. The system is designed to manage fluctuating solar irradiance conditions and battery charging/discharging cycles. Simulation results confirm the system's effectiveness in maintaining performance under variable environmental conditions. This research advances the optimization of PV systems integrated with energy storage, paving the way for more reliable and efficient renewable energy solutions while providing critical insights for academics and professionals in solar energy and power systems engineering.