Towards a physics-informed network paradigm with data generation and background noise removal for different distributed acoustic sensing applications
Peer-Reviewed Publication
Updates every hour. Last Updated: 25-Oct-2026 23:16 ET (26-Oct-2026 03:16 GMT/UTC)
A research team from Shanghai Jiao Tong University has developed a physics-informed neural network for distributed acoustic sensing. The approach generates synthetic event data and removes background noise without relying on real-world event data. Demonstrated in public datasets and belt conveyor fault monitoring, the method achieves high accuracy and generalizes across different sites, offering a scalable solution for industrial monitoring and optical sensing applications.