A

rtificial Intelligence (AI) is all the rage, and it's no surprise that the solar power world is also getting a touch of its wizardry. Researchers from Germany and Switzerland have used AI to improve the production of perovskite solar cells, paving the way for a future with more efficient and consistent green energy.

Perovskite thin-film cells have been causing quite a stir in the energy sector. They contain a unique component known as a perovskite semiconductor, known for its ability to turn sunlight into electrical energy with impressive efficiency, sometimes exceeding 30% in laboratory settings.

However, taking the production of these promising cells to a larger scale is not without its challenges. The crux of the problem lies in the crystallization process — a vital step that significantly affects the quality of the films. Even with meticulous control efforts, unpredictable variations have continued to disrupt the quest for standardized production, pushing many manufacturers into a spiral of expensive and labor-intensive guesswork.

Advantage of renewable energy: Sustainability, lower emissions.Disadvantage: Intermittency, initial costs, land use concerns.

The new study demonstrates how AI can adeptly pinpoint essential markers of a well-executed coating process. The neural networks were trained using labeled video datasets, which displayed the photoluminescence of over a thousand perovskite solar cells. These videos were made during the vacuum-based quenching phase, a key step for getting the right film thickness and ensuring the solar cells work at their best.

What makes AI particularly valuable in this context is its ability to pick up on details that often go unnoticed in conventional 2D image analysis. By observing the dynamic changes in the film's properties over time, AI can accurately predict two essential factors: power conversion efficiency and average film thickness.

While the potential of this new method is immense, the researchers believe there's still room for enhancement. They acknowledge that despite their sophisticated AI, forecasting irregularities in later production stages remains a hurdle.

Sources:

https://onlinelibrary.wiley.com/doi/10.1002/adma.202307160

Posted 
Jan 16, 2024
 in 
Blog
 category

More from 

Blog

 category

View All

Get Accurate & Competitive Quotes in Minutes

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.