🎓 Lesson 14
D5
Critical Node Identification and Vulnerability Mapping
Critical node identification finds the most important pipes, valves, or reservoirs in a water network whose failure would cause the biggest disruption to service.
🎯 Learning Objectives
- ✓ Analyze a water distribution network using graph-theoretic metrics (degree, betweenness, closeness centrality) to rank node criticality
- ✓ Calculate nodal vulnerability indices by combining centrality scores with exposure and consequence factors
- ✓ Design targeted mitigation strategies (e.g., redundancy placement, valve segmentation) for top-3 critical nodes identified in a given network
- ✓ Explain how regulatory requirements (e.g., AWWA standards) influence criticality thresholds and reporting criteria
📖 Why This Matters
In 2021, a single failed gate valve in Louisville’s water system left 50,000+ residents without pressure for 36 hours — not because it was old, but because its network position made it a hidden bottleneck. Critical node identification turns invisible interdependencies into actionable intelligence: it answers *which* component failures hurt most, *why*, and *where* to invest limited capital for maximum resilience gain — especially vital as climate extremes and aging infrastructure increase failure risks.
📘 Core Principles
Criticality arises from three intersecting dimensions: (1) Topological centrality — how connected and bridging a node is within the network graph; (2) Hydraulic sensitivity — how much pressure, flow, or service area changes when the node fails (assessed via EPANET-based scenario modeling); and (3) Consequence weighting — incorporating population served, critical facility proximity (hospitals, fire stations), and economic value. Vulnerability adds exposure: e.g., a high-centrality node in a flood-prone zone or with known pipe material degradation receives elevated vulnerability score. Modern practice uses multi-criteria decision analysis (MCDA) to fuse these layers objectively — avoiding over-reliance on any single metric.
📐 Composite Vulnerability Index (CVI)
The Composite Vulnerability Index quantifies relative risk by normalizing and weighting centrality, exposure, and consequence metrics. It enables direct comparison across diverse node types (reservoirs, pumps, valves, junctions) and supports ranking for intervention planning.
Composite Vulnerability Index (CVI)
CVI_i = w_c × C_i + w_e × E_i + w_k × K_iWeighted sum of normalized centrality (C), exposure (E), and consequence (K) scores for node i.
Variables:
| Symbol | Name | Unit | Description |
|---|---|---|---|
| CVI_i | Composite Vulnerability Index for node i | dimensionless (0–1) | Overall vulnerability score; higher values indicate greater priority for mitigation |
| w_c | Weight for centrality metric | dimensionless | Assigned based on system priorities (e.g., 0.4 if topology dominates reliability concerns) |
| C_i | Normalized centrality score for node i | dimensionless (0–1) | e.g., betweenness centrality scaled to [0,1] using min-max normalization across all nodes |
| w_e | Weight for exposure metric | dimensionless | Reflects regional hazard profile (e.g., higher in seismically active zones) |
| E_i | Normalized exposure score for node i | dimensionless (0–1) | Based on GIS hazard layers (flood depth, soil corrosivity, crime density, etc.) |
| w_k | Weight for consequence metric | dimensionless | Emphasizes societal impact (e.g., weighted toward hospitals in healthcare-critical networks) |
| K_i | Normalized consequence score for node i | dimensionless (0–1) | e.g., population served / max_population, or fire flow deficit / design requirement |
Typical Ranges:
Low-risk suburban network: 0.15 – 0.45
High-consequence urban core: 0.55 – 0.85
💡 Worked Example
Problem: Given: Junction J-42 has betweenness centrality = 0.38 (normalized 0–1), flood exposure score = 0.72 (0–1), and serves 12,500 people (population weight = 0.89, normalized against max served = 14,200). Weights: centrality = 0.4, exposure = 0.3, consequence = 0.3.
1.
Step 1: Normalize population served → 12,500 / 14,200 = 0.880
2.
Step 2: Apply weights → CVI = (0.38 × 0.4) + (0.72 × 0.3) + (0.880 × 0.3)
3.
Step 3: Compute → 0.152 + 0.216 + 0.264 = 0.632
Answer:
The CVI for J-42 is 0.632 (range 0–1), placing it in the 'High Vulnerability' tier (≥0.6), warranting priority inspection and valve isolation upgrade.
🏗️ Real-World Application
In the City of Austin’s 2022 Resilience Master Plan, engineers applied CVI analysis to its 2,100-km network. Using EPANET simulations and GIS-based hazard layers (flood, subsidence, corrosion risk), they identified 17 ‘Tier-1 Critical Nodes’ — including Pump Station PS-8 (high betweenness + located in 100-year floodplain) and Valve V-114 (closes off 42% of west-side pressure zone). Post-identification, targeted smart valve installation at V-114 reduced potential outage scope by 78%, validated in a simulated main break drill.
✏️ Student Exercise
Using the provided small network (5 junctions, 2 reservoirs, 1 pump, 1 tank), calculate betweenness centrality for all nodes (manual graph method), assign exposure scores (0.2–0.9) based on proximity to fault lines, and compute CVI for Junction J3 using equal weights (0.33 each). Then, justify which node should receive first mitigation funding based on your results and AWWA M36 guidance on service continuity.
🔧 Interactive Calculator
🔧 Open Water Distribution Network Analysis Calculator📋 Case Connection
📋 Calibration of Lagos Metropolitan Water Network
Persistent model–field mismatch (>25% pressure error) due to undocumented pipe replacements and unaccounted demand growt...
📋 Leak Localization in Tokyo’s Historic Cast-Iron Network Using ITA
Acoustic methods ineffective due to soil attenuation and ambient noise; conventional pressure zoning lacked resolution
📋 Climate-Adaptive Reinforcement of Cape Town’s Drought-Resilient Network
System unable to maintain minimum pressure during prolonged low-storage operation and projected 20% rainfall decline
📋 Water Quality Model Validation for Singapore’s Deep Tunnel Sewerage System (DTSS) Supply Branch
Disinfectant residual dropping below 0.2 mg/L at farthest nodes despite design dosing; suspected wall reaction dominance