The AQUAM project set out to create a software model that can predict the performance of installed photovoltaic (PV) modules in the field. Central to this goal was the design, construction and testing of an expert system that would carry out infrared (IR) and electroluminescence (EL) measurements automatically, reliably and efficiently. The focus was on understanding and modelling the ageing behaviour of defective modules under real operating conditions, taking into account thermal and mechanical stresses such as day‑night cycles, wind, snow loads and hail impacts. These stresses were considered crucial for determining degradation pathways, material fatigue and performance forecasts.

IRCAM GmbH was responsible for developing a flight‑capable EL camera system. The system was integrated into a drone platform, enabling rapid, non‑contact inspection of large PV arrays. The EL camera required an external electrical stimulus to excite the modules and a controlled lighting environment to capture the weak EL signal. The drone‑based IR camera, developed in parallel, could be operated in daylight without contacting the modules, providing quick visualisation of performance‑relevant faults. Together, the two imaging modalities supplied complementary data: IR thermography highlighted hot spots and temperature gradients, while EL imaging revealed electrically active and inactive regions, including micro‑cracks and shunts.

ZAE Bayern’s role was to develop the models and software algorithms that would translate the field‑generated image data (IR + EL) into quantitative performance predictions for faulty modules. The team built an automated expert system that processed the imaging data, identified defect signatures, and fed them into a performance model. This model incorporated degradation mechanisms identified in the field, such as potential induced degradation (PID) and cell cracking, and used them to forecast module output over time. The ageing model was validated against long‑term field data collected during a dedicated study (AP6), which monitored selected PV plants over several years. The study confirmed that the combined IR/EL approach could detect early signs of degradation and that the predictive model accurately tracked the decline in power output, improving maintenance planning and system optimisation.

Additional technical achievements included the development of a simulation framework (AP5) that allowed rapid evaluation of design scenarios, and the creation of new design strategies for PV plants (AP8) that leveraged the predictive insights to optimise module placement and system configuration. The project also produced a comprehensive set of performance‑prediction tools that can be applied to existing installations, enabling operators to identify underperforming modules and schedule targeted interventions.

Collaboration among the partners was tightly integrated. Rauschert GmbH, a PV system provider and plant operator, selected the PV installations for study, identified critical faults, and integrated the expert system into their operations. IRCAM GmbH supplied the hardware expertise for the drone‑based EL camera, while ZAE Bayern provided the analytical and modelling expertise. The project ran from early 2015 through to 2018, with key milestones such as the first field deployment of the drone system in 2016 and the completion of the long‑term study in 2018. Funding was provided by German research agencies, supporting the development of the hardware, software, and field trials. The project’s outcomes were disseminated through numerous peer‑reviewed journal articles, conference presentations, and industry workshops, underscoring its impact on the PV inspection and optimisation community.