My research mission centers on enhancing the reliability, durability, and sustainability of solar photovoltaic (PV) systems through the strategic integration of Artificial Intelligence (AI) and smart renewable energy technologies. The goal is to leverage AI's potential to significantly improve the operational efficiency, lifespan, and eco-friendliness of solar energy installations. By applying advanced machine learning algorithms and data analytics, I aim to develop innovative solutions for predictive maintenance, accurate energy yield forecasting, and optimised performance of solar PV systems. This endeavor seeks not only to bolster the resilience and longevity of these systems but also to ensure their sustainable integration into the energy grid. Emphasising the critical importance of reliability and durability, my research strives to pave the way for a future where solar energy systems can deliver on their promise of providing a clean, abundant, and sustainable energy source. This work represents a pivotal step towards realising a more sustainable and smartly-powered world, with solar PV technology at the forefront of this transformative journey.
Featured Publications
Comprehensive study on the efficiency of vertical bifacial photovoltaic systems: a UK case study
​
Nature Scientific Reports, 2024
​
This paper presents the first comprehensive study of a groundbreaking Vertically Mounted Bifacial Photovoltaic (VBPV) system, marking a significant innovation in solar energy technology. The study reveals that the VBPV system significantly outperforms both a vertically mounted monofacial PV (VMPV) system and a conventional tilted monofacial PV (TMPV) system in energy output. Key findings include a daily power output increase of 7.12% and 10.12% over the VMPV system and an impressive 26.91% and 22.88% enhancement over the TMPV system during early morning and late afternoon hours, respectively. Seasonal analysis shows average power gains of 11.42% in spring, 8.13% in summer, 10.94% in autumn, and 12.45% in winter compared to the VMPV system. Against the TMPV system, these gains are even more substantial, peaking at 24.52% in winter.
Dual spin max pooling convolutional neural network for solar cell crack detection
​
Nature Scientific Reports, 2023
​
This paper introduces a highly accurate solar cell crack detection system using four CNN architectures. It assesses cell condition through electroluminescence images, achieving up to 99.5% acceptance rate. Thermal tests validate its effectiveness, highlighting potential PV industry benefits.
Nature npj Materials Degradation, 2023
​
In this study, we analysed thermal defects in 3.3 million PV modules located in the UK. Our findings show that 36.5% of all PV modules had thermal defects, with 900,000 displaying single or multiple hotspots and ~250,000 exhibiting heated substrings. We also observed an average temperature increase of 21.7 °C in defective PV modules. Additionally, two PV assets with 19.25 and 8.59% thermal defects were examined for PV degradation, and results revealed a higher degradation rate when more defects are present. These results demonstrate the importance of implementing cost-effective inspection procedures and data analytics platforms to extend the lifetime and improve the performance of PV systems.
Inequalities in photovoltaics modules reliability: From packaging to PV installation site
Renewable Energy, 2023
​
This work assesses the reliability of 186 PV modules, tracking them from packaging to installation. Pre-packaging, no significant issues were found. After installation, 2.2% developed cracks, leading to output power losses. Thermal inspections revealed hotspots in cracked modules, indicating potential degradation. Additionally, PID tests showed greater impact on modules with cracks.
