Digital Material characterization has become increasingly important for obtaining virtual experimental data that are needed in early stage product development [1]. Modern state of the art methods[2,3] use high quality µ-CT images to calibrate effective nonlinear material models on the component scale.

In recent times, DMN (Deep Material Network) [4-6] shows potential to replace the tedious calibration effort for the material model on the component scale. They combine concepts from laminate theory and machine learning allowing a more reliable prediction of nonlinear material behavior.

In this work the industrial applicability of DMN for short fiber reinforced plastics is investigated in comparison to direct numerical simulation results [2,3] for different nonlinear material behavior[7].

References

[1] Dey, A. P., Welschinger, F., Schneider, M., Gajek, S., & Böhlke, T. (2023). Rapid inverse calibration of a multiscale model for the viscoplastic and creep behavior of short fiber-reinforced thermoplastics based on Deep Material Networks. International Journal of Plasticity, 160, 103484.

[2] M. Kabel, D. Merkert, and M. Schneider, “Use of composite voxels in fft-based homogenization,” Computer Methods in Applied Mechanics and Engineering, vol. 294, pp. 168–188, 2015.

[3] M. Kabel, A. Fink, and M. Schneider, “The composite voxel technique for inelastic problems,” Computer Methods in Applied Mechanics and Engineering, vol. 322, pp. 396–418, 2017.

[4] Liu, C. Wu, and M. Koishi, “A deep material network for multiscale topology learning and accelerated nonlinear modeling of heterogeneous materials,” Computer Methods in Applied Mechanics and Engineering, vol. 345, pp. 1138–1168, 2019.

[5] Z. Liu and C. Wu, “Exploring the 3d architectures of deep material network in data-driven multiscale mechanics,” Journal of the Mechanics and Physics of Solids, vol. 127, pp. 20–46, 2019.

[6] S. Gajek, M. Schneider, and T. Böhlke, “On the micromechanics of deep material networks,” Journal of the Mechanics and Physics of Solids, vol. 142, p. 103984, 04 2020.

[7] Gajek, S., Schneider, M. and Böhlke, T. (2023), Material-informed training of viscoelastic deep material networks. Proc. Appl. Math. Mech., 22: e202200143.