On the selection of submodeling zones for honeycomb sandwich elements
摘要整理
A problem of selecting rational submodeling zones in finite element analysis of sandwich structures with a honeycomb core is considered. The use of numerical methods in modern design inevitably faces a trade-off between analysis accuracy and computational cost. Simple mesh refinement across the entire structure is inefficient; therefore, the submodeling technique is a promising approach, allowing detailed examination of local zones only, with boundary conditions taken from the global analysis. A sandwich panel with aluminum face sheets and a honeycomb core, bolted and subjected to a uniformly distributed load, is taken as the object of study. The aim of the work is to select submodeling zones and determine their optimal dimensions, accounting for the loading pattern, boundary conditions, and underlying physics. In the theoretical part, two key zones are identified: the panel center (the region of maximum displacement) and the bolted joint zones (regions of elevated local stresses). A method for determining the dimensions of these zones is proposed. For the center zone, the criterion is suppression of edge effects (inclusion of at least 3–5 periodic unit cells). For the attachment zones, the in-plane dimension is determined by the half-wavelength of the deformation (from the bolt to the midspan), and the through-thickness dimension is determined by the decay length of local stresses according to the Saint-Venant principle (1,5–2,5 of the characteristic body dimension). Practical calculations were performed in two stages: global analysis with an orthotropic equivalent core and subsequent submodel analysis with explicitly modeled discrete cell geometry. Comparison of the results confirmed the effectiveness of the approach. The discrepancy in displacements between the global model and the submodels did not exceed 10%, indicating correct transfer of boundary conditions. At the same time, the maximum von Mises stresses in the attachment zone of the long side and at the panel center, obtained from the submodels, exceeded the global analysis results by 138 % and 695 %, respectively. This demonstrates that accounting for the actual core geometry is essential for strength assessment, as it reveals stress redistribution from the face sheets to the honeycomb core members. The developed recommendations for selecting submodeling zone dimensions allow efficient analysis models to be constructed, ensuring high fidelity of local strength analysis at acceptable computational cost, which is of particular importance for aerospace applications.