A thermodynamically consistent physics-informed deep learning material model for short fiber/polymer nanocomposites
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914730299359232 |
|---|---|
| author | Bahtiri, Betim Arash, Behrouz Scheffler, Sven Jux, Maximilian Rolfes, Raimund |
| author_facet | Bahtiri, Betim Arash, Behrouz Scheffler, Sven Jux, Maximilian Rolfes, Raimund |
| contents | This work proposes a physics-informed deep learning (PIDL)-based constitutive model for investigating the viscoelastic-viscoplastic behavior of short fiber-reinforced nanoparticle-filled epoxies under various ambient conditions. The deep-learning model is trained to enforce thermodynamic principles, leading to a thermodynamically consistent constitutive model. To accomplish this, a long short-term memory network is combined with a feed-forward neural network to predict internal variables required for characterizing the internal dissipation of the nanocomposite materials. In addition, another feed-forward neural network is used to indicate the free-energy function, which enables defining the thermodynamic state of the entire system. The PIDL model is initially developed for the three-dimensional case by generating synthetic data from a classical constitutive model. The model is then trained by extracting the data directly from cyclic loading-unloading experimental tests. Numerical examples show that the PIDL model can accurately predict the mechanical behavior of epoxy-based nanocomposites for different volume fractions of fibers and nanoparticles under various hygrothermal conditions. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_18310 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | A thermodynamically consistent physics-informed deep learning material model for short fiber/polymer nanocomposites Bahtiri, Betim Arash, Behrouz Scheffler, Sven Jux, Maximilian Rolfes, Raimund Machine Learning Artificial Intelligence Computational Engineering, Finance, and Science Numerical Analysis This work proposes a physics-informed deep learning (PIDL)-based constitutive model for investigating the viscoelastic-viscoplastic behavior of short fiber-reinforced nanoparticle-filled epoxies under various ambient conditions. The deep-learning model is trained to enforce thermodynamic principles, leading to a thermodynamically consistent constitutive model. To accomplish this, a long short-term memory network is combined with a feed-forward neural network to predict internal variables required for characterizing the internal dissipation of the nanocomposite materials. In addition, another feed-forward neural network is used to indicate the free-energy function, which enables defining the thermodynamic state of the entire system. The PIDL model is initially developed for the three-dimensional case by generating synthetic data from a classical constitutive model. The model is then trained by extracting the data directly from cyclic loading-unloading experimental tests. Numerical examples show that the PIDL model can accurately predict the mechanical behavior of epoxy-based nanocomposites for different volume fractions of fibers and nanoparticles under various hygrothermal conditions. |
| title | A thermodynamically consistent physics-informed deep learning material model for short fiber/polymer nanocomposites |
| topic | Machine Learning Artificial Intelligence Computational Engineering, Finance, and Science Numerical Analysis |
| url | https://arxiv.org/abs/2403.18310 |