Functional and structural resilience assessment in urban drainage networks: a physics-guided graph-based surrogate...

Area of Science:

• Environmental Engineering
• Hydraulic Engineering
• Network Science

Background:

• Effective management of urban drainage networks (UDNs) requires rapid resilience assessment under failure scenarios.
• Traditional hydrodynamic models like the Storm Water Management Model (SWMM) are computationally intensive, limiting real-time applications.
• Developing efficient surrogate models is crucial for UDN design, retrofitting, and management.

Purpose of the Study:

• To introduce a novel physics-guided, graph-based approach for rapid and accurate resilience assessment of UDNs.
• To offer a computationally efficient alternative to traditional hydrodynamic models for UDN resilience evaluation.
• To validate the proposed model's consistency with established methods using real-world case studies.

Main Methods:

• Developed a physics-guided surrogate model using graph theory for UDN resilience assessment.
• Utilized maximum subcatchment flow as input, employing a nonlinear method derived from SWMM outputs.
• Incorporated the weighted shortest path method and hydraulically informed graph metrics for flow routing and failure analysis.

Main Results:

• The graph-based model significantly accelerates resilience assessment compared to SWMM.
• The approach effectively evaluates UDN resilience under structural (e.g., pipe failure) and functional (e.g., intense rainfall) scenarios.
• Validation against two real-world case studies showed high consistency with SWMM-based resilience calculations.

Conclusions:

• The novel graph-based approach provides a computationally efficient and accurate method for UDN resilience assessment.
• This physics-guided model offers a viable alternative to traditional hydrodynamic models for real-time UDN management and design.
• The methodology demonstrates strong potential for enhancing the robustness and reliability of urban water infrastructure.