Reduced-Order Physics-Informed Neural Network with Adaptive Basis Refinement for Structural Identification

Summary

Physics-informed neural networks (PINNs) offer a flexible framework for forward and inverse problems, but their direct use in structural dynamics is limited by high dimensionality and model-form errors. Existing PINN-ROM couplings often rely on fixed reduced subspaces or assume complete physics. This work proposes a Reduced-Order Physics-Informed Neural Network (RO-PINN) with adaptive basis refinement, embedding reduced governing equations into the PINN loss via projection and updating the basis during training to realign the latent space with evolving parameters or residual restoring forces. Validated on a four-story steel frame with nonlinear hysteretic braces under sparse and noisy measurements, RO-PINN achieves parameter identification comparable to or more accurate than Bayesian model updating at lower computational cost, recovers unmodeled nonlinear restoring forces under incomplete physics, and jointly identifies residual forces and parameters in a unified framework.