Asha Viswanath ^1* Kamran A. Khan ¹ Diab Abueidda ² Seid Koric ² Mohamad Modrek ¹ Rashid K. Abu Al-Rub ¹

• ¹ Khalifa University, Abu Dhabi, United Arab Emirates
• ² University of Illinois at Urbana-Champaign, Champaign, Illinois, United States

Data drivenData-driven models acting that act as surrogates to for computationally costly 3D topology optimization techniques are very popular as because they help alleviate the multiple time-consuming 3D finite element analyses during optimization. In this paper, a one such 3D CNN basedCNN-based surrogate model for the topology optimization of a Schoen's Gyroid Triply triply periodic minimal surface unit cell is investigated. Gyroid-like unit cells are designed using a voxel algorithm and homogenization-based topology optimization codes in MATLAB. Few such optimization data are used as input-output for supervised learning of the topology optimization process by via the 3D CNN model in Python code. These models could then be used to instantaneously predict the optimized unit cell geometry for any topology parameters. The high accuracy of the model was demonstrated by a low mean square error metric and a high dice Dice coefficient metric. The model showed ahas the major disadvantage of running numerous costly topology optimization runs, but has an advantage that the trained model can be reused for different cases of TO and the methodology of accelerated design of 3D metamaterials can be extended for Formatted: Highlight designing any complex, computationally costly problems involving complex geometriesy of metamaterials with multi--objective properties or multi-scale applications. Towards this,The main purpose of this paper is to provide the complete associated MATLAB and PYTHON codes, for optimizing topology of any cellular structure and predicting new topologies using Deep Learning process, made available for educational purposes.