Module

Road

Overview Analysis Results
Overview

Seismic performance assessment of road networks under ground deformation

The Road module evaluates how permanent ground deformation affects individual roads and the performance of the overall road network. Its practical goal is to estimate road failure risk and identify which links are most critical for connectivity and flow after hazard-induced damage.

Primary objective

Turn ground-deformation intensity into road-level damage probabilities and system-level network risk.

Core engine

Fragility functions, Monte Carlo sampling, and network importance metrics are combined in a single transportation-resilience workflow.

Main outputs

Road failure probabilities plus network-level CPIM and RF indicators that show which links matter most.

Method

How it works

The workflow propagates hazard intensity from PGD to road damage and then up to network-wide performance measures.

The Road algorithm starts from permanent ground deformation (PGD) and propagates that hazard into road-level damage and network-level consequences. In the current implementation, PGD is the main hazard intensity measure used to evaluate seismic damage to each road segment.

For each road, the module assigns a fragility class and evaluates the probability of exceeding each limit state using a lognormal fragility model:

$$P(LS \mid PGD)=\Phi\left(\frac{\ln(PGD)-\ln(\theta)}{\beta}\right)$$

Here, $\theta$ is the median threshold for a damage limit state, $\beta$ is the dispersion parameter, and $\Phi(\cdot)$ is the standard normal cumulative distribution function. Roads are mapped to HAZUS-style roadway fragility classes such as Hrd1 and Hrd2.

Those exceedance probabilities are converted into discrete damage-state probabilities, producing a probabilistic description for each road segment rather than a single deterministic outcome.

The module then performs Monte Carlo sampling to estimate failure behavior across the network. In each simulation sample, road damage states are drawn according to their probabilities and the resulting network condition is checked.

Two network-level importance measures are then computed. CPIM measures how important each road is to disconnection risk between the source and target nodes, while RF measures how much network flow capacity is reduced when a road is damaged or removed. As a result, the module not only estimates whether each road is likely to fail, but also which roads are most influential for overall system functionality.

Inputs

What you need

The run requires network geometry plus PGD values so road fragility and network disruption can be evaluated together.

  • The Road module requires three CSV files: road_info.csv, node_info.csv, and pgd_values.csv.
  • road_info.csv must include road identifiers, network connectivity, capacity, and road geometry. Key columns are id, capacity, from, to, and geometry.
  • node_info.csv must include node identifiers and point geometry, typically id and geometry.
  • pgd_values.csv must include a numeric PGD column in cm.
  • If a guid column is available in both road and PGD data, PGD can be mapped directly to each road segment.
  • If guid is not available, the module falls back to a mean-PGD-based spatial variation model derived from road geometry.
  • If a road type column is provided, it is used to map each road to a fragility class such as Hrd1 or Hrd2.
Implementation note: for meaningful CPIM and RF results in the current setup, the network should contain node IDs 1 and 13, since they are used as the fixed source and target nodes for network-level calculations.