Scholarlink

<p>The efficient and accurate detection of faulty solar panels is crucial for optimizing a photovoltaic system performance, and ensuring its long-term sustainability. This study proposes a hybrid deep learning and metaheuristic optimization-based approach for detecting faulty solar panels using VGG-16-ConvLSTM for feature extraction and a hybrid Gray Wolf Optimizer Harris Hawks Optimization (GWO-HHO) algorithm for feature selection. Initially, PV fault images undergo preprocessing, including image resizing, normalization and augmentation to increase dataset variability. Extracted deep features from VGG-16-ConvLSTM, which integrates spatial and temporal dependencies, are then optimized using GWO-HHO to select the most informative features while reducing dimensionality. The optimized feature subset is subsequently classified using machine learning models such as Support Vector Machine, Decision Tree, Naïve Bayes, Random Forest, and Artificial Neural Network. Performance evaluation is conducted using key metrics including sensitivity, specificity, accuracy, precision, and F1-score. Experimental results demonstrate that the proposed VGG-16-ConvLSTM + GWO-HHO framework achieves superior classification accuracy, effectively identifying faulty solar panels with improved computational efficiency and reduced feature redundancy. The findings highlight the potential of integrating deep learning and metaheuristic optimization for automated photovoltaic fault detection, thus rendering it a promising solution for real-world solar energy monitoring systems.</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

Detection of Faulty Solar Panels with GWO-HHO Algorithm Based on VGG-16-ConvLSTM

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月1日

查看全文
  • Esmael Ommar Mahdi Ommar,
  • Mehmet Şimşir,
  • Javad Rahebi

摘要

The efficient and accurate detection of faulty solar panels is crucial for optimizing a photovoltaic system performance, and ensuring its long-term sustainability. This study proposes a hybrid deep learning and metaheuristic optimization-based approach for detecting faulty solar panels using VGG-16-ConvLSTM for feature extraction and a hybrid Gray Wolf Optimizer Harris Hawks Optimization (GWO-HHO) algorithm for feature selection. Initially, PV fault images undergo preprocessing, including image resizing, normalization and augmentation to increase dataset variability. Extracted deep features from VGG-16-ConvLSTM, which integrates spatial and temporal dependencies, are then optimized using GWO-HHO to select the most informative features while reducing dimensionality. The optimized feature subset is subsequently classified using machine learning models such as Support Vector Machine, Decision Tree, Naïve Bayes, Random Forest, and Artificial Neural Network. Performance evaluation is conducted using key metrics including sensitivity, specificity, accuracy, precision, and F1-score. Experimental results demonstrate that the proposed VGG-16-ConvLSTM + GWO-HHO framework achieves superior classification accuracy, effectively identifying faulty solar panels with improved computational efficiency and reduced feature redundancy. The findings highlight the potential of integrating deep learning and metaheuristic optimization for automated photovoltaic fault detection, thus rendering it a promising solution for real-world solar energy monitoring systems.