Verfahren und System zur Verwaltung von Ereignissen in einem Netzwerk aus Speicherprogrammierbaren Steuerungen unter Verwendung von Maschinellem Lernen

Details

Veröffentlichungs-Nr.

EP4715500

Anmeldetag

19. September 2024

Veröffentlichung

25. März 2026

Rechtsraum

EP

Abstract

A method and system for managing events in a network of programmable logic controllers (PLCs) 102A-N using machine learning models 104A-N are disclosed. The method involves a first PLC 102A periodically receiving a data package from a second PLC 102B, where the data package includes information used by a first machine learning model 104A to identify occurrences of events in the second PLC 102B. The first PLC 102A retrains a second machine learning model 104B by applying a learning algorithm to the received data package, ensuring the model's accuracy and effectiveness. The retrained second machine learning model 104B is then deployed into at least a third PLC 102C. The third PLC 102C uses the retrained model to predict future occurrences of events and execute corrective actions based on these predictions, thereby managing and mitigating potential risks.

Anmelder

Firma

Siemens Aktiengesellschaft

Land

🇩🇪 Deutschland