Rapid screening of photovoltaic cells via an unsupervised neural network autoencoder

Photovoltaic (PV) energy harvesting systems need to quickly and accurately assess the performance of solar cells during production. This study introduces an automated method that uses machine learning to simplify and speed up this process. A neural network autoencoder, trained with simulated data, is used to analyze key performance characteristics of PV cells. The system can group and classify performance data through a clustering technique called K-means, helping manufacturers identify potential issues faster. By linking physical input parameters to performance data, this approach not only improves quality control but also provides valuable insights for optimizing solar cell designs.