Multi-Objective Optimization of Modular Component Based on Geometric Digital Twin and NSGA-II

With the increasing demand for high-precision assembly in aerospace and precision manufacturing, the limitations of traditional manual assembly methods in terms of efficiency and accuracy have become increasingly crucial. Therefore, a multi-objective optimization framework for modular component assembly, integrating NSGA-II with geometric digital twin modeling is proposed in this paper. First, three-dimensional point cloud data of modular components are scanned using a handheld 3D laser scanner, followed by preprocessing steps including noise filtering, rigid transformation, prior segmentation, and analytical surface modeling. Then, geometric parameters extracted from the point cloud are incorporated into an arc-based dimensional chain model to evaluate circumferential consistency and gaps between modules and shells. Finally, NSGA-II is used to optimize assembly grouping while balancing multiple objectives and maintaining solution diversity. The experimental results demonstrate that the proposed approach reduces intra-group geometric variation by up to 57.35 % compared with random grouping and outperforms single-objective genetic algorithms by 8.64 %. This approach provides an effective tool for precise virtual assembly and digital twin assembly.