Neuron soundware predictive maintenance protects gearbox production
Our goal in working with this recognised European automotive manufacturer was to keep gearbox production on target and avoid unplanned failures in the helium compressor, a critical production line component.
A helium compressor used in manufacturing gearboxes has multiple operating modes. Our acoustic anomaly detection solution will learn all the standard acoustic modes for these machines so they can recognise a non-standard sound from the compressor and send a real-time warning of a possible upcoming incident. Each anomaly detected triggers an automatic warning for a maintenance specialist, who can inspect the equipment and head off an impending fault.
More than 50% of production managers are looking for solutions to help prevent these costly production downtimes.
Suspending production often causes significant losses. That’s why more than 50% of production managers are looking for solutions to help prevent these costly production downtimes. Combining IoT tech with AI gives us a revolutionary solution that digitizes machine maintenance and moves it into the future, to Industry 4.0.
“Our hardware is at a level that can handle advanced machine learning methods and the volume of data processed by neural networks for monitoring dozens of machines at a time 24 hours a day, seven days a week. This type of maintenance represents the future of manufacturing processes and we are happy to be able to support equipment all around the globe,” comments Pavel Konečný, CEO of Neuron Soundware.
“Our aim in this project is to catch any deviations from the standard sound of the compressors. We installed six highly sensitive and durable sensors on different parts of the compressors and connected them to our industrial IoT device, the nBox. This ‘box’ processes data from the sensors using edge computing. After processing the data, we send and save the most important acoustic output to the cloud for further analysis by AI, which also learns from the data and improves outcomes so that the customer learns of any imminent problems in a particular machine before they happen,” adds Pavel Konečný.