Method, device and system for identifying rainwater and sewage mixed connection of sewer network

By constructing a GNN model based on SWE-GNN and combining pipeline topology and dynamic features, the problems of low discrimination accuracy and insufficient computational efficiency in existing technologies are solved, and efficient and accurate identification of mixed rainwater and sewage connections is achieved.

CN122286104APending Publication Date: 2026-06-26POWERCHINA HUADONG ENG CORP LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
POWERCHINA HUADONG ENG CORP LTD
Filing Date
2026-05-28
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Existing methods for identifying mixed rainwater and sewage connections in drainage pipe networks have low accuracy and weak generalization ability in complex topological pipe networks, and are computationally expensive, making it difficult to meet the requirements for real-time identification.

Method used

The encoder-processor-decoder architecture of SWE-GNN is adopted, and the flux exchange principle of shallow water equation is incorporated to construct a GNN model suitable for identifying mixed stormwater and sewage connections. Combining the pipe network topology and dynamic characteristics, the accuracy and efficiency of the identification are improved through autoregressive prediction and multi-step lead loss function.

Benefits of technology

It improves the discrimination accuracy by 15%-20%, the dynamic scene discrimination accuracy is ≥88%, the cross-pipeline generalization accuracy is ≥85%, and the calculation speed is two orders of magnitude faster than the traditional numerical hydraulic model, meeting the needs of emergency response.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122286104A_ABST
    Figure CN122286104A_ABST
Patent Text Reader

Abstract

This application provides a method, apparatus, and electronic device for identifying mixed stormwater and sewage connections in drainage pipe networks. The method includes: acquiring real-time GIS data of the drainage pipe network, constructing an undirected graph model, and determining the graph node features and edge features in the model; in this model, manholes and pipe network endpoints are used as graph nodes, and pipe segment connection relationships are used as edges; preprocessing the graph node features; and inputting the edge features and the preprocessed graph node features into the mixed stormwater and sewage connection identification model so that the model outputs the mixed stormwater and sewage connection point identification results. The mixed stormwater and sewage connection identification model includes a GNN model suitable for mixed stormwater and sewage connection identification, constructed using an encoder-processor-decoder architecture of SWE-GNN and incorporating the flux exchange principle of shallow water equations. This application can integrate pipe network topological relationships, capture the dynamic evolution of mixed stormwater and sewage, and has strong generalization ability of the mixed stormwater and sewage connection identification model, thereby improving the accuracy and efficiency of mixed stormwater and sewage connection identification in drainage pipe networks.
Need to check novelty before this filing date? Find Prior Art