Fault Diagnosis & Data

How fault patterns no human can see come about

How 200+ neural networks detect anomalies in 100,000+ signals that stay invisible to the eye

Dr. Lukas Kaupp
Dr. Lukas Kaupp University of Darmstadt

Most production failures begin long before they become visible. A bearing running two degrees too hot. A sensor whose drift is barely noticeable across the day. Insignificant on their own – a failure when combined.

Kauzen analyzes 100,000+ signals simultaneously. Hundreds of automatically trained neural networks learn the normal operation of your plant and detect deviations no human can oversee at this dimension. This creates a fault picture that warns maintenance teams before the machine stands still.

This article will be expanded by our team with further sections.

Use cases

Use-case demo
  • Feed
  • Print
  • Sensorics
  • Lighting systems
← Back to overview