Scholarlink

Fault detection in photovoltaic (PV) arrays is a critical challenge for maintaining the efficiency and reliability of PV power systems. To address this issue, this study presents a deep learning-based approach by introducing a hybrid CNN-GRU-Attention model. The proposed model combines a Convolutional Neural Network (CNN) for spatial feature extraction from PV array output characteristics, a Gated Recurrent Unit (GRU) to capture temporal dependencies in fault signals, and an attention mechanism to dynamically prioritize significant features. Experimental results demonstrate that the proposed model significantly outperforms traditional methods in terms of both detection accuracy and generalization ability for common PV array faults, such as short-circuits, open-circuits, aging, and partial shading. This approach offers a robust technical foundation for enhancing intelligent maintenance in PV plants.

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

CNN-GRU-Attention Hybrid Network for Photovoltaic Array Fault Diagnosis

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Yi Lu,
  • Tianyou Li,
  • Jun Su

摘要

Fault detection in photovoltaic (PV) arrays is a critical challenge for maintaining the efficiency and reliability of PV power systems. To address this issue, this study presents a deep learning-based approach by introducing a hybrid CNN-GRU-Attention model. The proposed model combines a Convolutional Neural Network (CNN) for spatial feature extraction from PV array output characteristics, a Gated Recurrent Unit (GRU) to capture temporal dependencies in fault signals, and an attention mechanism to dynamically prioritize significant features. Experimental results demonstrate that the proposed model significantly outperforms traditional methods in terms of both detection accuracy and generalization ability for common PV array faults, such as short-circuits, open-circuits, aging, and partial shading. This approach offers a robust technical foundation for enhancing intelligent maintenance in PV plants.