Audio Analytics on Edge

We present an extensive use case of "XAL Edge Audio Analytics" applied to non-invasive monitoring of heavy hydropower assets. The key question is to what degree audio analysis on industrial equipment can provide an effective and inexpensive alternative to traditional monitoring and maintenance. Our edge computing system processeses real-time audio data feeding into analysis models for anomaly classification and provides access to analysis results.

The first part of this talk covers the overall architecture and points out specific design considerations. The second part presents some of the analytical models running currently on the edges and their results. We finish with lessons learned.


When done correctly, edge computing offers a remakably cost efficient and performant solution for doing predictive maintence and health monitoring for high frequency data. Close collaboration and interaction with the asset owners and their subject matter experts are keys to success.



Vinzenz Eck
Vinzenz Eck is founder and managing director of Expert Analytics GmbH in Munich. He is an expert in data science and computational engineering with more than 6 years industrial experience. His main areas of expertise are information modelling, numerical methods, data science, software architecture, sensitivity analysis and uncertainty quantification.

Ola Skavhaug
Ola Skavhaug is a founder of Expert Analytics AS, a private research and development company in Norway, and Expert Analytics GmbH in München. His research interests include numerical methods, biomedical computing, hydrology and energy market modelling. He is the main author of several open source software projects, most recently Rasputin and Shyft.


OPC Foundation






Sie möchten über die building IoT
auf dem Laufenden gehalten werden?