Optimization of Isogeometric Analysis-Based Topology Optimization Design with Global Stress Constraint in AI

Abstract

Topology optimization is a powerful technique used in engineering design to optimize the material distribution within a given design domain, resulting in improved structural performance. Isogeometric analysis (IGA) has emerged as a promising approach for topology optimization due to its ability to seamlessly integrate geometric design and analysis. The traditional formulation of IGA-based topology optimization focuses on minimizing compliance, neglecting stress constraints that are crucial for ensuring structural integrity. In propose an optimization framework that combines IGA-based topology optimization and artificial intelligence (AI) to optimize designs with a global stress constraint. The objective is to simultaneously minimize the compliance and control the maximum stress in the structure. The proposed framework leverages the power of AI algorithms, such as deep neural networks, to efficiently search for the optimal material distribution that satisfies the stress constraint.

The optimization process begins with the generation of an initial design based on the isogeometric representation. The design is then iteratively updated using an AI-driven optimization algorithm that incorporates the global stress constraint. The AI algorithm learns from the stress distribution patterns in the design domain and guides the optimization towards finding an optimal material distribution that minimizes compliance while ensuring the maximum stress does not exceed the predefined limit. To evaluate the performance of the proposed approach, several numerical experiments and comparisons with traditional methods are conducted. The results demonstrate that the combination of IGA-based topology optimization and AI significantly improves the structural performance by effectively controlling stress levels while achieving better compliance values. The optimized designs exhibit enhanced strength and structural efficiency compared to those obtained using conventional methods.

The integration of AI algorithms with IGA-based topology optimization offers a promising avenue for tackling complex engineering design problems with stress constraints. The proposed framework provides engineers with a powerful tool to design lightweight and structurally robust components. Furthermore, the application of AI in topology optimization can significantly reduce the computational cost and accelerate the design process. Overall, this research contributes to advancing the field of structural optimization by integrating AI-driven approaches into isogeometric analysis-based topology optimization for enhanced design efficiency and performance.