Bridge Digital Twinning Using an Output-Only Bayesian Model Updating Method and Recorded Seismic Measurements

Area of Science:

• Civil Engineering
• Computational Mechanics
• Seismology

Background:

• Manual post-earthquake bridge inspections are time-consuming and hazardous.
• Existing Finite Element (FE) models for seismic response simulation have uncertainties in material properties and input motions.
• Digital technologies offer potential for faster and more accurate damage diagnosis.

Purpose of the Study:

• To develop and validate a sequential Bayesian model updating technique for joint identification of bridge FE models and foundation input motions.
• To reduce uncertainties in seismic response analysis of bridges by integrating sensor data with prior models.
• To demonstrate the application of digital twinning for real-time structural health monitoring and damage diagnosis.

Main Methods:

• Sequential Bayesian model updating to jointly identify linear/nonlinear FE models and foundation input motions.
• Incorporation of soil-structure interaction effects within the FE models.
• Numerical verification studies and a real-world case study using seismic data from the San Rogue Canyon Bridge.

Main Results:

• The model updating technique effectively identified bridge responses with reasonable agreement to recorded data.
• Valuable insights into soil-structure interaction effects were obtained by comparing free-field and foundation input motions.
• The updated FE model, acting as a digital twin, showed potential for detailed damage diagnosis at various structural levels.

Conclusions:

• The presented Bayesian model updating technique is effective for seismic response analysis and damage diagnosis of bridges.
• Digital twinning, through updated FE models, offers a powerful tool for post-earthquake assessment and structural health monitoring.
• The methodology has significant potential for real-world applications in bridge safety and resilience.