Yığın zekası algoritması ile anahtar ve kesit geometrileri için ağ rafinasyonu ve optimizasyonu

Özet (abstract)

Finite element analysis is used to compute the stress and deformations of machine components. Meshing is a crucial step for analysis. The accuracy of results in the analysis depends on meshing. Mesh refinement is achieved by maintaining the position of elements with the help of geometric constraints without disrupting their connectivity. The quality of the mesh is improved by mesh optimization, maintaining the integrity of the geometry. The condition number of each hexahedral element serves as the quality measure for evaluating the element distortion and guiding the optimization process. The Cohort Intelligence (CI) based optimization algorithm iteratively updates the disturbed node positions, minimizing the global condition number while preserving the geometric features and boundary constraints. The distorted hexahedral meshes for different types of keys, such as Gib headed key, Flat key, Square key and different types of sections, such as T Section, L Section, I Section, which are categorised as mechanical complex geometries are considered for study. Statistical analysis, 95% confidence interval analysis, and paired t-tests confirmed the robustness of the proposed CI algorithm. Results indicate CI optimization effectively restores the noisy hexahedral meshes and improves the mesh quality. Thereby enhances the mesh suitable for finite element analysis. Performance of CI optimization for restoration of mesh was compared with Genetic Algorithm (GA) and the Particle Swarm Optimization (PSO) using the same parameters. Results indicate CI optimization effectively restores the noisy hexahedral meshes and improves the mesh quality. Thereby enhances the mesh suitable for finite element analysis.

Yazarların özeti; kaynağından alınmıştır. Discover Mechanical Engineering, 2026 · DOI ↗