Deep material networks for efficient scale-bridging in thermomechanical simulations of solids
- 326 páginas
- 12 horas de lectura
Focusing on deep material networks (DMN), this work establishes a mathematical foundation and introduces a novel formulation that reduces degrees of freedom. It offers an efficient solution for nonlinear DMNs, enhancing complex two-scale simulations while minimizing computational demands. Additionally, a new interpolation technique is introduced, allowing for the integration of fluctuating microstructure characteristics into macroscopic simulations, thus improving accuracy and efficiency in material modeling.
