An accelerated, convergent, and stable nodal integration in Galerkin meshfree methods for linear and nonlinear mechanics

Abstract

Summary Convergent and stable domain integration that is also computationally efficient remains a challenge for Galerkin meshfree methods. High order quadrature can achieve stability and optimal convergence, but it is prohibitively expensive for practical use. On the other hand, low order quadrature consumes much less CPU but can yield non‐convergent, unstable solutions. In this work, an accelerated, convergent, and stable nodal integration is developed for the reproducing kernel particle method. A stabilization scheme for nodal integration is proposed based on implicit gradients of the strains at the nodes that offers a computational cost similar to direct nodal integration. The method is also formulated in a variationally consistent manner, so that optimal convergence is achieved. A significant efficiency enhancement over a comparable stable and convergent nodal integration scheme is demonstrated in a complexity analysis and in CPU time studies. A stability analysis is also given, and several examples are provided to demonstrate the effectiveness of the proposed method for both linear and nonlinear problems. Copyright © 2015 John Wiley & Sons, Ltd.

A new Jameson–Schmidt–Turkel Smooth Particle Hydrodynamics algorithm for large strain explicit fast dynamics

“…Within the community of mesh free methods, two well-established families are available for the consideration of numerical stabilisation, namely gradient based stabilisation [23] and the strain smoothing procedure [65,66]. Unfortunately, to the best of the authors' knowledge, most of the available stabilisation algorithms still suffer from non-physical pressure instabilities in highly nonlinear nearly incompressible scenarios [17,67].…”

Penetration modeling of ultra‐high performance concrete using multiscale meshfree methods

“…NMAP naturally models large deformations that include material fragmentation of the experiments described in the previous section because the RKPM framework has no need for ad hoc failure models such as element erosion that are necessary for the finite element method to model perforation (Guan et al and Yreux and Chen). For this class of problems, naturally stabilized nodal integration (Hillman and Chen,) is introduced to reduce spatial instabilities that govern the effects of debris and fragmentation experienced in these dynamic events. Quasi‐linear kernels on the support are proposed by Yreux and Chen, to ensure the stability of the model.…”

A SPH framework for dynamic interaction between soil and rigid body system with hybrid contact method

“…The lack of flexibility in numerical resolution may result in high computational cost in large‐scale problems with fine grain size. On the other hand, continuum‐based Lagrangian particle meshless methods, for example, smoothed particle hydrodynamics (SPH), reproducing kernel particle method (RKPM), and material point method (MPM), can easily model large soil deformations and moving boundaries. This is because they are based on Lagrangian formulations so there is no need to track the interfaces and moving boundaries, and owing to the meshless nature, they do not suffer from the instabilities induced by mesh distortion.…”