Graded multiscale topology optimization using neural networks. (January 2023)

• Record Type: Journal Article
• Authors: Chandrasekhar, Aaditya; Sridhara, Saketh; Suresh, Krishnan
• Is Part Of: Advances in engineering software. Volume 175(2023)
• Publication Date: 2023-01
• Subjects: Multiscale topology optimization -- Graded microstructure -- Neural networks -- Automatic differentiation; Computer-aided engineering -- Periodicals; Engineering -- Computer programs -- Periodicals; Engineering -- Software -- Periodicals; Periodicals; 620.0028553
• DOI: 10.1016/j.advengsoft.2022.103359
• Languages: English
• ISSNs: 0965-9978

Abstract

Abstract: In this paper, we propose a novel graded multiscale topology optimization framework by exploiting the unique classification capacity of neural networks. The salient features of this framework include: (1) the number of design variables is only weakly dependent on the number of pre-selected microstructures, (2) it guarantees partition of unity while discouraging microstructure mixing, (3) it supports automatic differentiation, thereby eliminating manual sensitivity analysis, and (4) it supports high-resolution re-sampling, leading to smoother variation of microstructure topologies. The proposed framework is illustrated through several examples. Highlights: A novel neural network representation for multiscale topology optimization. It guarantees partition of unity and discourages microstructure mixing. It efficiently handles numerous pre-selected microstructures. Number of design variables is only weakly dependent on the number of microstructures.