PV cell manufacturers typically rely on data sent to Manufacturing Execution System (MES) to retrieve information on the machine status and performances, but this requires a complex IT infrastructure and technical skills to process data. Furthermore, most of the effort in a data analysis activity is spent on data cleaning and features engineering. The backbone of Applied's proposal is a software tool embedded in the metallization line to perform data collection, historicization, automatic stop reason labeling and visualization. The same software can integrate data from sources other than the metallization line: automated optical inspection (AOI) instrumentation, sensors, etc.
On top of the data lake thus built, multiple software features can be added with machine learning algorithms to optimize recipe tuning, implement closed-loop controls and predictive maintenance, thus increasing uptime and yield and reducing the need for highly skilled operators.
For more information contact firstname.lastname@example.org.