IoT Platform Combined with Machine Learning Techniques for Fault Detection and Diagnosis of Large Photovoltaic Plants

The maintenance of solar photovoltaic plants is critical to ensure reliable energy production and to reach the competitiveness in the energy market. Maintenance in photovoltaic industry is evolving into new condition monitoring systems based on thermal cameras and unmanned aerial vehicles due to high acquisition rate. Novel fault detection techniques and data analysis methods are required to analyze large volumes of images and, therefore, data. The automatization of fault detection is a key factor to enhance maintenance activities. This work introduces a novel approach to develop an Internet of Things platform built with CakePHP, HTML, CSS and JavaScript for hot spot detection of aerial thermographic images. This platform applies two consecutive artificial neural networks, and the methodology is tested with a real case study with thermal images from different photovoltaic solar plants. The analysis is performed in the platform, and the results obtained showed a 100% of accuracy for panel detection, and around 90% of accuracy for hot spot detection.