Category: Graduate Student Seminar Physics-Informed Machine Learning for Rapid Damage Assessment of Reinforced Concrete Bridge Columns

October 31, 2025 1:00 pm

In this quarterly IBT/ABC-UTC Research Seminar, Parmida Rahmani, Ph.D Graduate Student & Research Assistant, and Floriana Petrone, Ph.D., Associate Professor at the School of Civil Engineering and Environmental Engineering, University of Nevada - Reno, presents work related to Rapid Damage Assessment of Reinforced Concrete Bridge Columns following seismic events.

presentation begins:  3:31
Q & A begins:  49:26


Damage states' drift ratio and lateral force distribution

Description:
Following an earthquake, bridge owners will need to rapidly determine the status of hundreds of structures with a balance of speed and accuracy of nonlinear analysis. This presentation introduces a physics-informed machine learning approach that uses material and geometric features to predict the monotonic pushover response of RC bridge columns and the associated damage states (DS1–DS5). The model reconstructs the force–drift curve using an energy-constrained bi/trilinear fit with minimum area error, normalizing for typical stiffness bias. Yield and ultimate force/drift points are derived from a library of 480 representative OpenSees nonlinear models and validated against experimental data. Compact, uncertainty-constrained equations are delivered for application, accuracy calibrated for critical points as well as damage-state approximations. The result is a fast, interpretable tool—well-suited for network-level screening and prioritization—that bridges the gap between coarse fragility methods and computationally demanding high-fidelity simulations.

 Presenters:


Parmida Rahmani
Ph.D. Student & Research Assistant
Civil & Environmental Engineering
University of Nevada, Reno
Email:  prahmani@unr.edu

 

Ms. Rahmani’s research focuses on rapid post-earthquake assessment of bridge infrastructure through physics-informed machine learning, monotonic/cyclic pushover modeling, and OpenSees simulation. She develops interpretable, uncertainty-aware predictors of RC column response and damage thresholds to guide network-level decisions following earthquakes. By integrating domain mechanics with data-driven approaches, she creates actionable tools for practitioners and agencies.


Floriana Petrone, Ph.D.
Associate Professor
Civil & Environmental Engineering
University of Nevada, Reno
Email:  prahmani@unr.edu

Floriana is an Associate Professor in the Department of Civil and Environmental Engineering at the University of Nevada, Reno (UNR), with expertise in reinforced concrete modeling and novel simulation techniques for seismic hazard and risk assessment under large uncertainties. Dr. Petrone’s research aims to advance the scientific knowledge and computational tools needed to enhance the resilience of civil infrastructure and energy systems to natural hazards. Her group works on integrating high-performance computing, advanced numerical modeling, and machine learning to generate frameworks for innovation across infrastructure design, maintenance, operations, and rapid recovery.

Presentation Graphics:
graphics courtesy of UNR

Figure 1. Damage State Definition

Figure 2. Damage states’ drift ratio and lateral force distribution for all bridge column samples.

Figure 3. Real and predicted pushover response with energy-constrained piecewise fit; overlay of actual vs. predicted damage-state points (DS-1-DS5). 

For more information: Please contact us by sending an email to abc@fiu.edu.