A naukri.com initiative
Arxiv
2d
309
Image Credit: Arxiv
History-Aware Neural Operator: Robust Data-Driven Constitutive Modeling of Path-Dependent Materials
- An end-to-end learning framework for data-driven modeling of path-dependent inelastic materials using neural operators is presented in a recent study.
- The History-Aware Neural Operator (HANO) is introduced as an autoregressive model that predicts path-dependent material responses without hidden state variables, addressing issues in recurrent neural network models.
- HANO, built on a Fourier-based neural operator backbone, allows for discretization-invariant learning and incorporates a hierarchical self-attention mechanism for multiscale feature extraction.
- By modeling stress-strain evolution as a continuous operator, HANO can adapt to varying path discretizations, making it robust in handling complex conditions like irregular sampling, multi-cycle loading, noisy data, and pre-stressed states.
- The study evaluates HANO on elastoplasticity and progressive anisotropic damage problems, demonstrating superior predictive accuracy, generalization, and robustness compared to baseline models.
- With its capabilities, HANO serves as an effective data-driven tool for simulating inelastic materials and can be integrated with classical numerical solvers.
Read Full Article
18 Likes
For uninterrupted reading, download the app