Generative Design Combined with Topology Optimization for Lightweight Product Structure: An Implementation Path

Advances in digital manufacturing and computational design have transformed product development. Technologies such as 3D printing, topology optimization (TO), and generative design (GD) allow the production of lightweight, high-performance structures. However, traditional design workflows often rely on manual iterations and cannot efficiently explore diverse design options. The development of digital manufacturing and computational design has transformed the way products are developed, and it is now possible to produce lightweight and high-performance structures. Conventional design processes tend to be manual in nature, and they do not allow exploration of a wide range of design variants. The hybrid method introduced in this paper is a combination of Topology Optimization (TO) and Generative Design (GD) to create automatically lightweight, structurally efficient, and manufacturable designs. The approach uses the Solid Isotropic Material with Penalization (SIMP) technique of TO and the state-of-the-art Fruit Fly Optimized Generative Gradient Networks (F2O-GGN) to reduce the amount of material used in the process and still retain the structural integrity of the product, offering an automated design exploration and fabrication tool. The TO Designs dataset from Kaggle defines the target structure's design space, loads, and material properties. Use TO and the Solid Isotropic Material with Penalization (SIMP) approach to identify important load routes and remove unnecessary material. Use Fruit Fly Optimized Generative Gradient Networks (F2O-GGN), which combine Generative Adversarial Network (GAN), Gradient Boosting (GB), and Fruit Fly Optimization (FFO), to create various design options that meet structural restrictions while optimizing performance metrics. Export the chosen designs in formats suitable for additive manufacturing to ensure fabrication viability. The suggested F2O-GGN approach successfully generated diversified, lightweight, and structurally resilient designs for benchmark TO challenges. Designs showed significant material reductions while maintaining performance. The Python-based implementation evaluated results using RMSE = 0.07 and R² = 0.98, demonstrating high accuracy in predicting structurally efficient designs. Combining with F2O-GGN generative design creates an efficient, automated pathway for producing lightweight and manufacturable product structures, improving both design flexibility and engineering performance.