Anticipating issues rather than correcting them, to save time and money - this could be the slogan for ASTRIIS, the start-up from GIPSA-lab currently in the SATT Linksium incubator. “Everything that rotates, moves or vibrates emits vibration signals that can be recorded and analysed in real time,” explains Nadine Martin, CNRS research director at the laboratory.
Drawing on 30 years of research in analysing signals with applications in preventive maintenance, she has combined her results in one piece of software, AStrion. It’s a program for automatic signal processing, capable of detecting all frequencies representing the condition of different parts for each kind of signal. “This kind of autonomous super scanner has been designed to assist with maintenance operations for machines,” specifies the researcher. “By tracking changes over time for indicators associated with an individual frequency or a physically relevant group of frequencies, the software can identify any abnormal signal changes thanks to learning algorithms based only on the data base being monitored.” The software can monitor each part of a system remotely and analyse large data bases. The intuitive interface means that the analysis can be understood by operators and facilitates the work of experts by allowing them to access all results in detail.
Demonstrated by proof of concept in the wind sector, AStrion is ready for deployment on industrial machinery to test its performance in real conditions.
*CNRS, Grenoble INP – UGA, UGA
