Neuron Soundware (NSW) Neuron Soundware: detecting machine failures early combining sound, AI and IoT technologies

Summary

NSW is an AI-based acoustic & vibration diagnostic technology for early detection of broken machines and process monitoring, providing unparalleled insights into how mechanical systems operate. Combined with its large-scale applicability and ease of use, it disrupts the way machines are being diagnosed, facilitating industrial sustainability, catering to the needs of Industry 4.0. Our technology functions as an auditory cortex, understanding audio signals and detecting prominent issues, providing highly reliable results reaching 99.6% accuracy. NSW integrated hardware with an embedded edge software and the cloud-based software platform provides automatic detection of machine failures before any other methods, preventing costly downtimes and unnecessary maintenance works. Our detection process utilizes the audio datasets holding terabytes of data, further extended with each new anomaly, continuously improving algorithms' accuracy, addressable failures and types of machines.

Resources

Show all and search (1)

Unfold all

Fold all

More information & hyperlinks

Web resources:	https://cordis.europa.eu/project/id/190167814
Start date:	01-02-2023
End date:	31-12-2025
Total budget - Public funding:	5 116 185,25 Euro - 2 500 000,00 Euro

Cordis data

Original description

NSW is an AI-based acoustic and vibration diagnostic technology for early detection of broken machines and process monitoring, providing unparalleled insight into the operation of mechanical systems. Combined with its large-scale applicability and ease of use, it disrupts the way machines are diagnosed, facilitating industrial sustainability and catering to the needs of Industry 4.0. Our technology acts as an auditory cortex, understanding audio signals and detecting salient problems, delivering highly reliable results with 99.6% accuracy. NSW's integrated hardware with embedded edge software and cloud-based software platform automatically detects machine faults before other methods, preventing costly downtime and unnecessary maintenance. Our detection process utilizes our proprietary audio dataset, which contains terabytes of data and is continuously expanded with each new anomaly, improving the accuracy of the algorithms, the faults addressed and the types of machines.