Deep material networks for efficient scale-bridging in thermomechanical simulations of solids

• Tác giả: Sebastian Gajek
• Ngôn ngữ: eng
• ISBN: KSP/1000155688
• Thông tin xuất bản: KIT Scientific Publishing 2023
• Mô tả vật lý: 1 electronic resource (326 p.)
• Bộ sưu tập: Tài liệu truy cập mở
• ID: 373738
• Chủ đề: 1. deep material networks; 2. data-driven modeling; 3. Two-scale simulations; 4. Deep Material Networks; 5. Datengetriebene Modellierung; 6. Zweiskalensimulationen; 7. micromechanics; 8. Mikromechanik; 9. machine learning; 10. Maschinelles Lernen; 11. Mechanical engineering & materials

Tóm tắt

We investigate deep material networks (DMN). We lay the mathematical foundation of DMNs and present a novel DMN formulation, which is characterized by a reduced number of degrees of freedom. We present a efficient solution technique for nonlinear DMNs to accelerate complex two-scale simulations with minimal computational effort. A new interpolation technique is presented enabling the consideration of fluctuating microstructure characteristics in macroscopic simulations.