PSI - Issue 78

Simone Reale et al. / Procedia Structural Integrity 78 (2026) 1657–1664

1661

Table 2. RVs used in the simulation, based on (Ferreira et al., 2015).

RV

Distribution

Mean ( μ )

Std ( σ )

Bounds

D rcm,0 [10

-12 m 2 /s]

Normal Normal Normal Normal

4.5 0.4

0.2μ 0.16

≥ 0

α [ -]

0≤ α ≤1

T real [K]

292

5

- -

4800

700

b

C s,Δx [wt-%/c]

Lognormal

4

1.8

≥ 0

Δx [mm]

Beta Beta

10

5

0≤ Δx ≤ c-1 0.2≤ C crit ≤2

C crit [wt-%-c]

0.6

0.15

c [mm] f c [MPa] f y [MPa]

Normal

6 5

≥ φ t +1

c nom

Lognormal Lognormal

43

≥ 0

495 ≥ 0 The corrosion propagation phase can be described following Faraday’s Law (Liu and Weyers, 1998): ( ) = 0.0116 ∙ ∙ ∙ ( ) In Equation (17), R is the pitting factor, t p is the propagation period and i corr (t p ) is the corrosion current density [μA/cm 2 ]. R is sampled from a Gumbel distribution (Stewart and A. Al-Harthy, 2008), while the time-dependent value of i corr can be estimated as per Equation (18) (Vu and Stewart, 2000): ( )= 37.8∙(1− ) −1.64 0.85 ∙ − 0.29 [ / 2 ] (18) 4.2. Ground Motion Selection The distribution of the Engineering Demand Parameters (EDPs) conditioned on the Intensity Measure (IM) level is estimated through Multiple Stripe Analysis (MSA) (Jalayer and Cornell, 2009) considering 10 IM levels with increasing return period T R of 10, 50, 100, 250, 500, 1000, 2500, 5000, 10000, 100000 years. The spectral acceleration corresponding to the bridge fundamental period S a (T 1 ) has been selected as IM. The site hazard curve has been computed using REASSESS (Chioccarelli et al., 2019) assuming a V S,30 of 300 m/s. Seismic hazard deaggregation has been performed at each IM level, and hazard consistent ground motions have been selected through the Conditional Spectrum method (Jayaram et al., 2011; Baker and Lee, 2017). For each IM level, 30 pairs of ground motions have been retrieved from the PEER database (PEER, 2011) matching mean and variance of the target S a distribution. 4.3. Numerical Modeling The numerical model of the case study bridge has been developed using OpenSees (McKenna, 2011). Elastic beam elements have been used to model the deck, according to the common assumption of linear elastic response of the superstructure in the presence of seismic events. Bearing devices have been modeled through ElastomericBearingPlasticity elements. The bent cap has been described through elastic beam elements. Bent columns have been modeled using inelastic frames with force-based formulation and Gauss-Radau integration scheme. Fiber based sections are used. Cover concrete is modeled through Concrete01 uniaxial material, while Concrete04 is used for the core concrete. Steel02 is used to describe the reinforcement fibers, combined with a MinMax criterion to simulate bar fractures and the Fatigue wrapper to include LCF effects. Zero length elements combined with the ElasticPPGap material have been used to model the interaction between deck and abutment backwall, following the indications reported in (Caltrans, 2019). Geometrical nonlinearities have been included in the model. 30 (17)