.png){kind=icon}
.png){kind=icon}
Return
On the synthesis of the measurements and structural models using statistical finite element method
.png){kind=icon}
OA
AI
A
DOI：10.1016/j.mechrescom.2026.104638.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In this paper, we present a statistical finite element method (statFEM) for the synthesis of measurement data and computational models of architected metamaterials under material and model uncertainty. The proposed framework combines a probabilistic forward model with Bayesian inference to update displacement fields using sparse, noisy observations. Synthetic measurements are generated using a high-fidelity Timoshenko beam model, while inference is performed with a reduced Euler-Bernoulli model, thereby introducing controlled model discrepancy. Material uncertainty is modelled as a Gaussian random field defined over the connectivity graph of the lattice, with a covariance structure derived from a discretised stochastic partial differential equation (SPDE). This formulation yields sparse precision matrices that enable efficient Bayesian updating for large-scale lattice systems. Numerical studies on body-centred cubic (BCC) and octet lattice structures demonstrate that the proposed approach can recover accurate solution fields from limited measurements, reduce predictive uncertainty, and identify regions most affected by model mismatch and connectivity defects. Our results show that the SPDE-based statFEM framework provides a computationally tractable and statistically rigorous approach for fusing data and models of metamaterials. It offers a promising extension to uncertainty-aware digital twins of architected materials and structures.
Keywords:
Random fields
Lattice structures
Digital twinning
Gaussian processes
Finite element method
Bayesian inference
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
M
IF:
2.3
Papers:
115
Citations:
3.9K
Organization
Cited Papers