Artificial Intelligence Applied to Asset Management and Vibration Analysis

Peter Caldwell, Senior Engineer, ALS Industrial

Description

This presentation will have two focus areas which are:

  • Anomaly Detection Systems for the monitoring of asset performance and health.
  • Artificial Neural Networks for automating diagnosis within vibration analysis.

Anomaly Detection Systems can be used to identify occurrences when an asset is operating outside of its historical limits that may be due to operational changes or asset health deterioration. This presentation will overview the theory behind Anomaly Detection Systems before looking at current industrial applications including their integration into asset management workflows. Concluding this section, the strengths and challenges when implementing an Anomaly Detection System will be reviewed.

Artificial Neural Networks allow computers to perform categorisation tasks that have typically been performed by experienced people. Categorisation forms an integral part of the diagnostic phase of vibration analysis and hence provides an opportunity to leverage an Artificial Neural Network. Through some light-hearted demonstrations using Artificial Neural Networks, this presentation will build an understanding of the strengths and limitations of the concept. A specific look at applying Artificial Neural Networks to vibration analysis will explore the options that may be used by developers in this field.

The presentation will be concluded through exploring the likely day to day implications of Artificial Intelligence on personnel and assets.

This presentation will have two focus areas which are:

  • Anomaly Detection Systems for the monitoring of asset performance and health.
  • Artificial Neural Networks for automating diagnosis within vibration analysis.

Bio

I currently work as a online condition monitoring specialist for ALS Inustrial charged with managing our online systems and assessing new and emerging systems. I graduated from Curtin University as a Mechanical Engineer I spent many years working as a reliability consultant specialising in test and measurement where vibration analysis was a key skill. For the last 10 years I have been working almost exclusively with online vibration condition monitoring systems and received my VCAT4 qualification in 2019. I have studied Data Science in an effort to better understand the value of the data being produced by online systems. These studies have been invaluable when discussing and assessing commercially available systems which are looking to leverage existing and new data sources to optimise the monitoring and management of assets.