21. Predicting temperature dependent solar-module performance


Klaus Jäger

Blog post – Klaus Jäger


  At the Department Optics for Solar Energy (SE-AOPT) of Helmholtz-Zentrum Berlin für Materialien und Energie we work on experimental and numerical optics in and for photovoltaic devices. Our current topics are “Light management for solar cells,” “Optical Optimization of bifacial solar plants” and “Photon up-conversion and optical sensing.” For the second topic on optical optimization we use real-life weather data and several data-science approaches.

What is the data science project’s research question?   We developed an optical model which can predict the illumination on solar modules in a large solar module field, when real-life weather data is given as an input. For predicting the energy yield of the module, however, also the temperature dependence of the solar-cell performance is important. In this project we aim to extend the model such that the energy yield of PV modules can be predicted accurately, when temperature and – if available – wind data are included.

What data will be worked on?   Publicly available data from PV installations including weather data, solar irradiance and corresponding module performance data.

What tasks will this project involve? 

– Study the existing model
– Develop a plan how to include temperature effects, e.g. by machine learning or with a physical model
– Implement the model
– Test the validity of the module and – if possible – benchmark existing physical models vs. machine learning approaches

What makes this project interesting to work on?   Photovoltaic solar energy is one of the most important technologies to supply humankind with environmental-friendly electricity. For further expanding photovoltaics, reliable energy-yield modelling is crucial. This project combines one of the most promising energy technologies with data science approaches, as large sets of weather and PV-module performance data are considered.

What is the expected outcome?   Contribution to research paper, Contribution to software development

What infrastructure, programs and tools will be used? Can they be used remotely?  Optical models previously developed by the group. Common machine learning algorithms. Office space available, remote work also possible.

What skills are necessary for this project?  Scientific computation, Computational models, Computer simulations, Data mining / Machine learning

Is the data open source?  Yes 

Interested candidates should be at Master level.  Klaus Jäger is looking for 1 visiting scientist, working on the project together with the team.