STMicroelectronics and Hurence conference: rule engine and machine learning for the detection of execution faults in a complex industrial production line

[news]
Mar 21, 2019
image

Conference Big Data Paris 2019 Tuesday, March 12, 2019 from 17:00 to 17:20 in the Grand Amphitheater of Convention Center paris

Guillaume Lepelletier, Senior Staff Engineer STMicroelectronics and Thomas Bailet, Big Data CTO Hurence SAS will present “rules engine and machine learning for the detection of execution faults in a complex industrial production line”

The semiconductor manufacturing industry faces greater challenges than any other industry due to the nature and complexity of products, manufacturing processes and equipment. The silicon wafer production facility (fab) of STMicroelectronics located in Crolles (Isère) is an extremely automated production system with title already rocked into the era of Industry 4.0. At the heart of the unit is a 10,000m² clean room where several hundred different products are manufactured simultaneously on a wide range of technologies and several hundred different generation machines and features. Due to the versatility of the markets addressed, operational agility is therefore essential to industrial success.

This agility has a cost in terms of complexity and heterogeneity of information and control systems and, parallel to the deployment of automation, there is an evolution of the manufacturing trades. At STMicroelectronics in Crolles, the human being finds more and more its place as a supervisor: it guarantees the good functioning of the system in its globality by being able to unblock as quickly as possible any abnormal situation.

STMicroelectronics collaborated with Hurence (Lumbin , Isère), a recognized specialist in Big Data systems, to set up a system allowing:

  • to detect abnormal situations in the production line (whether of computer, mechanical or human origin);
  • provide real-time diagnostic assistance on the potential origin of the anomalies and the corrective actions to be taken.

This system is based on a Big Data architecture structured around components to develop a system capable of:

  • Real-time processing of significant volumes of industrial data driving the STMicroelectronics plant.
  • to exploit the experience accumulated by the operators thanks to the use of a rules engine
  • to embark on supervised learning in order to automatically improve the detection of abnormal situations as well as the diagnosis
Cookies policy

This site uses cookies to optimize your browsing experience and for audience measurement purposes. By continuing your navigation on this site without changing the settings of your cookies you accept its use.

I agree learn more