Self-learning condition monitoring system

Production processes are affected by a wide variety of anomalies and errors. Such errors can affect the quality of the products or lead to unplanned machine and production downtime.

The remedy here is reliable process monitoring, which makes it possible to detect errors and problems in the process flow at an early stage and in an automated manner. 

The self-learning condition monitoring system fits seamlessly into existing automation landscapes without any configuration effort.

A simple and effective protection system that contributes to early error detection in real time and thus significantly to the stability of the process.

The System

Simple and powerful data acquisition

  • Connectors for easy connection to the production environment
  • Efficient data management in an Influx DB

Web application

  • Automatic model creation: learning process models from historical process data
  • Application of the learned models for process monitoring in real time

 

The expansion modules

Connectors

  • OPC UA
  • MQTT
  • Profinet connector with patented method for self-configuration

Process models and condition monitoring methods

  • Discrete process models (parallel automata and Petri nets)
  • Hidden Markov models
  • LSTMs
  • kNN classifiers

SmartBox: Robust and space-saving hardware design of the CMS

 

Highlights

  • Reliable error detection in real time
  • Simple integration without configuration effort
  • Execution as web application or as SmartBox
  • Connectors for easy connection to your production line
  • Self-learning condition monitoring methods for a wide range of applications

 

Added value

  • Minimization of errors and system downtimes
  • Improvement of product quality