For asset health monitoring purposes, the iQunet sensors are developed to perform interval-based measurements (for example 1 measurement every hour or every day) instead of performing 1 measurement once a quarter or even once a year. In this way, a large amount of data is collected and saved on to your local iQunet Server. Based on this data, a consistent trend can be built up. Analyzing these trends will provide you with insights into the change of the behavior of the installation and allow for corrective measures to be taken if necessary.
iQunet offers a Machine Learning Service to assist you with this analysis. After you have subscribed to the service, all historical sensor data stored on your local iQunet Server for the specified training period will be automatically compressed and transferred once to the iQunet Machine Learning Servers (located in the iQunet premises) to calculate a machine learning data model. This model is then returned and saved on to your local iQunet Server for continuous local anomaly monitoring (subscription based). New measurements that differ too much from the calculated data model are detected as anomalies and can be followed up and flagged (difference based on the Mean Squared Error).
In the iQunet Sensor Dashboard you can visualize trend graphs of these anomalies and set alarm levels.
We provide you with a ready to useAnomaly Detection Service, integrated in the iQunet software. Start your AI journey today and request a demo.