Numerical Computation for Acoustic Problems with a Cell-Based Smoothed Radial Point Interpolation Method

YAO Ling-yun,YU De-jie,ZANG Xian-guo

DOI: https://doi.org/10.3969/j.issn.1000-3835.2011.10.036

2011-01-01

Abstract:It is well known that the standard finite element method(FEM) is unreliable to compute approximate solutions to Helmholtz equation for high wave numbers due to numerical dispersion.The cell-based smoothed radial point interpolation method(CS-RPIM) was extended to solve 2D acoustic problems and the formulation of CS-RPIM was presented for two-dimensional acoustic problems.With this method,the acoustic domain was discretized using triangular background cells,and each cell was further divided into several smoothing cells,the acoustic pressure gradient smoothing technique was implemented to each smooth cell.The system equations were derived using the smoothed Galerkin weak form,and the necessary boundary conditions were imposed directly according to the finite element method(FEM).The CS-RPIM greatly reduced the numerical dispersion error and obtained accurate results for acoustic problems because it provided a properly softened model stiffness.Numerical examples including a tube and a 2D acoustic problem of a car were analyzed and the results showed that the CS-RPIM achieves more accurate solutions and higher convergence rates as compared with the corresponding finite element method,especially,for higher wave numbers.