Energy transition
Recycling flood
In the years ahead, many photovoltaic systems and wind turbines will go to the scrap pile due to ageing. Their raw materials, however, are much too good for that.
It is a sticky, hot summer’s day. Rain has been dripping onto the shiny blue-black panels of the solar system for days. This should actually not be a problem – unless the protective layer of plastic film on the back of a module is damaged: thin hairline cracks are sufficient for moisture to soak in. This can lead to electronic components no longer being sufficiently insulated – and the solar system automatically switching off for safety reasons. The failure of just a single module is enough for such a scenario.
Tiny defects such as a hairline crack in the film are almost impossible to detect, and the fact that it can paralyze an entire solar park at short notice is not the only problem: “Damage caused by defective backs contributes to limiting the service life of a solar park to around 20 years at present, while it could reliably supply electricity for 50 years,” says physicist Dr. Ian Marius Peters from the Helmholtz Institute Erlangen-Nürnberg for Renewable Energy (HI ERN), a Jülich branch office. He and his team are researching how to extend the service life of solar systems and minimize performance losses.
“Around 10 per cent of the modules don’t last as long as they could,” says Peters’ colleague Dr. Claudia Buerhop-Lutz from HI ERN. “Some of the materials in the modules show the first signs of age-related degradation after just five years.” This quickly ruins the efficiency gain of a new system with its initially higher efficiency levels. The researchers at HI ERN are therefore developing methods to recognize defects as early as possible, relying on three strategies: non-destructive measurements in the field, the use of high-resolution drone cameras, and the use of artificial intelligence (AI).
10
per cent
-
of the modules do not last as long as they could, says Dr. Claudia Buerhop-Lutz from the Helmholtz Institute Erlangen-Nürnberg for Renewable Energy (HI ERN).
Destruction is not necessary
Claudia Buerhop-Lutz is focussed on the back of solar modules, which is made up of several layers of plastic film and protects the solar cell from water. “Films are playing an increasingly important role in the reliability of solar modules,” says Buerhop-Lutz. “Corrosion, surface abrasion or peeling layers compromise the protection the films provide.” Such damage has increasingly occurred in PV modules built between 2010 and 2012. According to the expert, less suitable materials may have been used during this time, but she has also found damage in newer systems.
Buerhop-Lutz and her team have developed a new method to examine the condition of the films directly in the field without damaging them. To this end, the researchers first analyzed numerous variants of the module backs made from different combinations of plastic films in the laboratory – a total of more than 30,000 modules from 30 solar parks – and also any visible damage. This has yielded a “backside” library with around 250 variants, which now serves as a comparison for infrared measurements of the modules in the field and helps to detect possible defects before they lead to module failure.
Drones in pairs
Aerial monitoring of the solar installations using drones is focussed on the front of the modules. In the COSIMA project, Buerhop-Lutz and her team combine two different methods: thermographic drones that can be used to detect hotspots – solar modules with localized overheating – and drones that measure electroluminescence. For the latter, the researchers reverse the process in the solar cell: when current is applied, the solar cell emits light – at the undamaged points.
“This allows us to recognize defects that are hidden deeper in the material,” says the researcher. AI then sorts all the information. “The analysis not only provides the operators with the usual jumble of images, but also indicates where a loss of performance is imminent. This is true predictive maintenance,” says Ian Marius Peters.
AI finds errors in standard data
The HI ERN scientists can also filter out errors from the operators’ usual monitoring data and analyze them using machine learning in the dig4MorE project. AI recognizes both performance deficits and error patterns that indicate specific malfunctions. “Modules are sometimes wired incorrectly, malfunctions only occur in individual parts or in certain climatic conditions,” reports Buerhop-Lutz, who would like to get more data from the solar parks. “Operators often don’t even realize the value of the information contained in their data,” she observes. “We could unearth these treasures.” The work of HI ERN researchers is already helping to make green power generation with PV systems more reliable. This is important because renewable sources are expected to supply at least 80 per cent of gross electricity consumption by 2030. Uncontrolled outages are undesirable.
Text: Katja Engel | images: Adobe Stock #218474496 (top); HI ERN/Jessica Pölloth (2nd from the top)
Contacts
Further Information
Last Modified: 29.02.2024