کنفرانس بین المللی فناوری و مدیریت انرژی

صفحه اصلی / دهمین کنفرانس بین المللی فناوری و مدیریت انرژی

Optimizing Solar Panel Performance Through Advanced CNN Architectures for Fault Classification

نویسندگان :

Mahmood Seyyedzadeh¹ Alireza Tajdid² Amir A. ghavifekr³ Mohammad Hassanzadeh⁴ Mohammad Mehdi Paikane⁵

1- دانشگاه تبریز 2- دانشگاه تبریز 3- دانشگاه تبریز 4- دانشگاه شهیدمدنی آذربایجان 5- دانشگاه شهیدمدنی آذربایجان

کلمات کلیدی :

solar panel،fault detection،CNN architecture،Optimization

چکیده :

Optimizing the performance of solar panels is crucial for enhancing energy efficiency and sustainability. This study explores advanced CNN architectures—VGG19, InceptionV3, and ResNet-50—for fault classification in solar panels. Using a dataset of six fault categories, images were preprocessed, resized, and balanced with class weights. Models leveraged pre-trained ImageNet weights, fine-tuning, and the Adam optimizer for efficient training. Performance was evaluated using standard metrics, and experiments were conducted on different hardware setups to assess classification accuracy and computational efficiency. Results provide insights into optimizing solar panel diagnostics for improved maintenance and energy output.