DIVIDER: Dividing and Allocating Current Flow for Predicitve Maintenance
Duration:
10/2022
—
12/2023
Total Time:
1 Year
Predictive maintenance for automated systems and processes, Photo: JOANNEUM RESEARCH
The Project
How can maintenance be identified at an early stage?
Maintenance is often carried out according to an interval schedule or predicted proactively through monitoring of various parameters. This monitoring requires the deployment of numerous sensors and measuring devices to detect wear and sudden damages at an early stage. DIVIDER aims to determine whether it is technically possible to identify those anomalies solely through measurements and characterization of the main power supply.
Our activities within the project
We develop mathematical models of systems with the aim of modelling the entirety of the individual system components. At the ROBOTICS Solution Centre, we commission these systems and apply forces to them to simulate wear or anomalies while measuring the main power supply in high resolution. Subsequently, we highlight these measurement data and make them accessible to machine learning. Afterwards, we validate these models ROBOTICS Solution Center.
Research group
Downloads
Sponsors
FFG
Bundesministerium für Klimaschutz, Umwelt, Energie, Mobilität, Innovation und Technologie (BMK)
Project Partners
Fraunhofer Austria
Messfeld GmbH
Project Details
The project DIVIDER is about the continuous monitoring of machines and systems to detect potential maintenance needs at an early stage. Normally, this monitoring requires the use of many sensors and measuring devices to identify wear or sudden anomalies. For electric machines, both electrical and mechanical measurement techniques are employed to monitor their condition.
The exploratory project DIVIDER aims to determine whether it is fundamentally possible to identify wear or spontaneously occurring anomalies based solely on measurements of the main power supply. We are using a high-precision sensor system to record the smallest fluctuations in the main power supply and store them as data points. The large datasets are then analysed, characterised and assigned using AI methods.
