Physical Mathematics

Neural Network-based Isogeometric Topology Optimization of Multi-material Structures under Thermal and Mechanical Loads

Abstract

Wenyi Morae*

The increasing complexity of engineering problems, especially in structural design, has led to the development of advanced optimization methods aimed at improving the performance of multi-material structures under various loading conditions. Topology Optimization (TO) is a powerful technique that allows for the design of structures with optimal material distribution to achieve desired mechanical, thermal, and other performance criteria. In recent years, Isogeometric Analysis (IGA) has gained popularity as a method for integrating design and analysis, bridging the gap between Computer-Aided Design (CAD) and Finite Element Analysis (FEA). Combining isogeometric analysis with topology optimization presents a promising approach for designing multimaterial structures, especially when subject to complex thermal and mechanical loadings. The use of Neural Networks (NNs) further enhances this optimization process by improving efficiency and enabling real-time design adjustments. This report explores the integration of these advanced techniques and presents a numerical investigation into the isogeometric topology optimization of multi-material structures under thermal and mechanical loadings using neural networks. The fundamental goal of topology optimization is to find the optimal material distribution within a given design space, subjected to various constraints and loading conditions.