An improved cell-based smoothed radial point interpolation method using condensed shape functions for 3D interior acoustic problems

摘要

The accuracy of the finite element method (FEM) for interior acoustic problems deteriorates gradually as the wave number increases due to the "dispersion error". In essence, the "dispersion error" is caused by the "overly-stiff" property of a FE discrete model. To alleviate this problem in this study, a new type of virtual node selection scheme for the condensed shape functions is created first to soften the stiffness of a discrete model. Then, an improved cell-based smoothed radial point interpolation method (CS-RPIM-T5-Cd2) with the "close-to-exact-stiffness" property is developed based on the new selection scheme and the traditional method CS-RPIM-T5 Numerical examples, including a benchmark case and a practical engineering problem, have been studied, which demonstrates that the present method possesses better accuracy, efficiency, convergence, robustness, and error control performance than the FEM and the similar methods for interior acoustic problems. (C) 2021 ElsevierB.V. All rights reserved.