PhD defence by Martin Bartholomäus

Abstract

Solar power has grown rapidly over the past decade and has become a cornerstone of the global energy transition. At the same time, solar technology itself is changing at an unprecedented pace. Traditional “monofacial” solar panels, which collect light only from the front, have largely been replaced by “bifacial” modules that also harvest reflected light from the rear side. In parallel, new solar cell technologies such as PERC and, more recently, TOPCon have pushed efficiencies ever higher.Because new solar technologies are deployed faster than long-term field testing can keep up, their reliability is mainly assessed using accelerated laboratory tests before products reach the market. This means that once panels are installed outdoors, monitoring their health and detecting faults becomes crucial. Today, operation and maintenance (O&M) of solar power plants still relies heavily on simple performance thresholds and labor-intensive on-site inspections.

This PhD thesis explores how modern solar modules degrade in the real world and how advanced electrical measurements can be used to detect problems remotely and more accurately. It shows that bifacial modules degrade in a broadly similar way to conventional modules, although some fault types change or disappear while new ones emerge. This underlines the need for continuous monitoring as solar technologies evolve. A key focus of the work is the use of current–voltage (IV) curves, which describe the electrical behavior of solar modules and strings. The thesis examines both how well bifacial modules can be modeled and how accurately IV curves can be measured using equipment already present in solar power plants, such as inverters. The results indicate that while modeling accuracy remains a challenge, measured IV curves contain rich information about the health of a PV system. By combining IV curve data with modern data-driven and machine learning methods, the thesis demonstrates the potential for highly accurate remote fault detection. This approach enables the identification of faults at an early stage, often before significant power losses occur.

Overall, this research demonstrates that reliable, remote fault detection in solar power plants is achievable using inverters. This opens the door to smarter, less labor-intensive O&M strategies, helping ensure that the rapidly expanding solar fleet delivers clean and reliable energy for decades to come.

Supervisor

• Sergiu Viorel Spataru

Evaluation Board

• Stela Canulescu - Technical University of Denmark
• David Moser - Becquerel Institute Italia, Italy
• George E. Georghiou - University of Cyprus

Master of Ceremony

• Gisele Alves dos Reis Benatto