A method for monitoring and pressure inversion of water hammer in urban water supply network based on DAS and physical information neural network

CN122282004APending Publication Date: 2026-06-26NANJING UNIV OF POSTS & TELECOMM

Patent Information

Authority / Receiving Office: CN · China
Patent Type: Applications(China)
Current Assignee / Owner: NANJING UNIV OF POSTS & TELECOMM
Filing Date: 2026-04-01
Publication Date: 2026-06-26

AI Technical Summary

Technical Problem

Existing urban water supply network monitoring technologies suffer from blind spots and high false alarm rates in complex noise environments, making it impossible to achieve quantitative pressure monitoring and difficult to accurately assess the damage level and remaining lifespan of water hammer on pipelines.

Method used

A method based on DAS and physical information neural network is adopted. Data is collected by distributed fiber optic acoustic sensors, and noise is filtered out by combining variational mode decomposition and KL divergence algorithm. A deep learning model of graph convolutional network and temporal convolutional network is constructed to identify water hammer events. The physical information neural network is used for quantitative inversion, and Radon transform is combined to track the propagation of water hammer waves, so as to realize full-process monitoring and pressure inversion.

Benefits of technology

It achieves full-coverage, blind-spot-free water hammer monitoring in complex urban environments, accurately captures and quantitatively inverts the pressure field inside the pipe, provides precise data support, and provides a basis for pipeline leakage control, pipe burst early warning and water pump scheduling decisions, reducing false alarm rate and ensuring water safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

Figure CN122282004A_ABST

Patent Text Reader

Abstract

This invention discloses a method for full-process monitoring and pressure inversion of water hammer in urban water supply networks based on DAS and a physical information neural network, belonging to the field of urban lifeline safety monitoring technology. It utilizes distributed optical fiber sensing (DAS) to collect data; adaptively removes environmental noise using the VMD-KLD algorithm; decouples signal characteristics using a spatiotemporal dual-flow network (TCN+GCN) to accurately identify water hammer events and eliminate false alarms; and for the first time, quantitatively inverts the fluid pressure field (MPa) and velocity field inside the pipeline using fiber optic phase data by introducing a physical information neural network (PINN) constrained by fluid dynamics equations. Finally, it uses Radon transform to achieve wavefront tracing and source location. This invention solves the problems of blind spots, inability to quantitatively obtain pressure values, and high false alarm rates in traditional monitoring methods, providing a fully transparent and high-precision digital monitoring method for the safe operation of long-distance pipelines.

Need to check novelty before this filing date? Find Prior Art