A Differentiable Framework for Inverse Identification of Heterogeneous Cohesive Laws
Abstract
This work develops a differentiable finite element framework based on an intrinsic cohesive-zone formulation for identifying traction-separation laws from global force-displacement histories and spatially sparse strain measurements. The framework employs physics-constrained neural networks to represent flexible, thermodynamically structured cohesive laws, with sensitivities computed through automatic differentiation and a discrete adjoint that accounts for the history-dependent evolution of interfacial damage. To enable identification of spatially heterogeneous cohesive behavior, we extend learned constitutive priors to irreversible interfacial damage by constructing a shared, low-dimensional family of cohesive laws from representative traction-separation histories. This representation exploits common constitutive structure and parameter sharing to reduce the number of unknowns and regularize the inverse problem, while allowing local cohesive behavior to vary spatially through latent coordinates. Numerical investigations involving double-cantilever beams and fiber-matrix composites demonstrate that accurate reproduction of global force-displacement responses does not necessarily imply accurate recovery of the underlying cohesive laws. Incorporating spatially resolved strain measurements substantially improves constitutive identification, even when the admissible responses are constrained by a learned prior. Comparisons with conventional bilinear cohesive models and Bayesian optimization further illustrate the benefits of the learned constitutive representation and adjoint-based identification for high-dimensional inverse problems.
Related
- Fully differentiable framework for inverse identification of geometry and material parameters with application to determining stress-free configuration of soft tissues
- A differentiable PCA-GPR surrogate framework for full-transient Bayesian inverse uncertainty quantification of black-box nuclear thermal-hydraulic systems
- Efficient angle-aware differentiable Mask3D modeling for inverse lithography
- Differentiable Inverse Design of Two-Body Interactions for the Moore-Read State
- Benchmarking Optimizers to Solve Inverse Problems with Differentiable Physics Simulators
BibTeX
@misc{hao2026differentiable,
title = {{A Differentiable Framework for Inverse Identification of Heterogeneous Cohesive Laws}},
author = {Sida Hao and Jinkyo Han and Bahador Bahmani},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.33711},
eprint = {2610.33711},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.33711},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - A Differentiable Framework for Inverse Identification of Heterogeneous Cohesive Laws
AU - Sida Hao
AU - Jinkyo Han
AU - Bahador Bahmani
PY - 2026
DA - 2026/10/24
PB - arXiv
DO - 10.48550/arxiv.2610.33711
UR - https://arxiv.org/abs/2610.33711
AB - This work develops a differentiable finite element framework based on an intrinsic cohesive-zone formulation for identifying traction-separation laws from global force-displacement histories and spatially sparse strain measurements. The framework employs physics-constrained neural networks to represent flexible, thermodynamically structured cohesive laws, with sensitivities computed through automatic differentiation and a discrete adjoint that accounts for the history-dependent evolution of interfacial damage. To enable identification of spatially heterogeneous cohesive behavior, we extend learned constitutive priors to irreversible interfacial damage by constructing a shared, low-dimensional family of cohesive laws from representative traction-separation histories. This representation exploits common constitutive structure and parameter sharing to reduce the number of unknowns and regularize the inverse problem, while allowing local cohesive behavior to vary spatially through latent coordinates. Numerical investigations involving double-cantilever beams and fiber-matrix composites demonstrate that accurate reproduction of global force-displacement responses does not necessarily imply accurate recovery of the underlying cohesive laws. Incorporating spatially resolved strain measurements substantially improves constitutive identification, even when the admissible responses are constrained by a learned prior. Comparisons with conventional bilinear cohesive models and Bayesian optimization further illustrate the benefits of the learned constitutive representation and adjoint-based identification for high-dimensional inverse problems.
ER -