Work

Optimal reduced-order model for Vehicle-Bridge Interaction and bridge damage detection using model based SHM-ML techniques

Year: 2026

Type: article

Abstract: • Optimal reduced-order vehicle-bridge interaction model. • Reliable Structural Health Monitoring (SHM) and damage detection. • VBI computational framework with significantly reduced computational cost. • Enriched model with realistic rail irregularities. • Experimental validation over multiple operating speeds, loads, and damage scenarios. Bridge structures subjected to moving vehicle loads exhibit complex dynamic behavior influenced by structural flexibility and rail surface irregularities. The correct approach and optimal modeling of vehicle-bridge interaction effects are crucial for the accurate computational determination of dynamic bridge vibrations, for reliable Structural Health Monitoring (SHM) and damage detection. The purpose of this work is twofold. First, a novel computational framework is introduced to generate an optimal reduced-order Vehicle-Bridge Interaction (VBI) model that captures the essential dynamics of the system with significantly reduced computational cost. This process begins with the development of a high-fidelity finite element (FE) model of the bridge which is updated using experimental modal data. This model is then reduced using Component Mode Synthesis (CMS) to enable efficient simulation while preserving the essential dynamic behavior and is combined with a simplified vehicle model enriched with realistic rail irregularities, producing an accurate and computationally efficient VBI system. Second, the study introduces a classification framework based on a lightweight Convolutional Neural Network (CNN) trained on numerically generated data from the validated VBI model. The CNN operates on frequency-domain features derived from bridge acceleration responses, which reflect damage-induced changes in stiffness and energy distribution of the structure. Using these features and an ensemble of independently trained CNNs, the framework classifies the structural health condition of the bridge.

Source: Mechanical Systems and Signal Processing

DOI: 10.1016/j.ymssp.2026.114188

Authors: George Karyofyllas, Dimitrios Giagopoulos

Institution: Aristotle University of Thessaloniki

Language: English

Topic: Structural Health Monitoring Techniques

Subfield: Civil and Structural Engineering

Field: Engineering

Domain: Physical Sciences

SDGs: Industry, innovation and infrastructure, Sustainable cities and communities

Open Access status: hybrid

Estimated APC paid: $4,730