.png){kind=icon}
.png){kind=icon}
Return
Unsupervised Graph-Based Deep Clustering for Optimal Monitoring Point Placement in Urban Drainage Networks
.png){kind=icon}
OA
AI
Q
Q
G
L
B
DOI：10.1002/eng2.70544.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The optimal placement of monitoring points in urban drainage networks is critical for operational efficiency, yet traditional methods suffer from high costs and redundancy. To address this, we propose a novel unsupervised framework that integrates a graph convolutional network for topological feature extraction with a structured deep clustering network to identify critical monitoring zones without requiring labeled data. On both a self-collected dataset and the public C-Town benchmark, the model demonstrated high performance, achieving over 94% clustering accuracy against expert-defined labels and identifying functional network zones with high precision. This high-precision clustering yields significant operational improvements: the number of required monitoring points was reduced by over 40% while critical node coverage increased to over 90%. This optimized layout also enhanced monitoring reliability by over 8% and shortened pollution event response times by nearly 36%, offering a scalable, data-driven solution for intelligent and efficient urban water management.
Keywords:
drainage networks
graph convolutional networks
monitoring point optimization
structured deep clustering
unsupervised learning
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
2
Papers:
362
Citations:
1.7K
Organization
Cited Papers