Failures pattern detection in electronic plate manufacturing chain and establishing correlations between multiple sensor measurements lead to improvement in operations performance.
La compañía
Electrónica Cerler SA is a SME (+300 staff) that offers services of industrialization, purchase, and production of electronic products, focusing on innovation and sustainability, within a context of lean manufacturing culture . Outsources electronics manufacturing for sectors oh high added value. Some company milestones are: introduction of RFID technology for intelligent management of production lines, automatic system of production and data collection, etc.
El reto
Preventing failures in the electronic plate manufacturing chain drives to higher operations performance. Pattern detection and correlations between the multiple sensor measurements lead to failures prevention in electrical and functional tests.
La solución
esHPC (BIFI) performed data analysis and processing to develop a software (R, Python, elastic search) to simulate all the potential mistakes in the production chain processing thousands of manufacturing variables such as solder paste volume, welding temperature, optical measurements, electrical measurements (resistance, inductance, capacity coefficient). Design of fault prediction models that estimate, before the end of the assembly process, the malfunction probability of an electronic board.
The project was carried out using the Caesaraugusta supercomputer resources and funded by the company (approx. 50.000 €)
ParticipantesCentro de supercomputación
|
Ficha técnicaTipo de problema
|
