Search

Hypervelocity impact (HVI) on carbon fiber reinforced plastics (CFRPs) is associated with extreme impulse response of material as well as complex characteristics of the composites. It is a challenging task to predict the physical process of CFRP-HVI problems accurately. We apply a GPU-accelerated smoothed particle hydrodynamics (SPH) method which incorporates the composite material model and the decoupled finite particle method to simulate the CFRP-HVI processes. The presented SPH model is validated by simulating the processes of aluminum spheres impacting metallic plates. The simulation results agree well with available experimental and numerical results, and the GPU parallelization technique significantly improves the simulation efficiency (350 times faster than an equivalent serial SPH model). With such computational accuracy and efficiency, the model is extended to simulate CFRP laminate impact problems, and a particle convergence analysis is performed. It is shown that the model can obtain convergent results when modeling each CFRP layer with three SPH particles in the thickness direction and correspondingly using about twenty million particles in total. The simulation can be completed within several hours, and the dominant mechanisms of the CFRP-HVI can be captured quantitatively. To further investigate the protective performance of CFRP structures, HVI on a well-known shielding structure Whipple bumper is investigated numerically. The dynamic response of the structure is well reproduced, and the results show that the CFRP pressure wall is more effective than the metallic one. These simulation results demonstrate that the presented SPH model can model the CFRP-HVI problems accurately and efficiently. [ABSTRACT FROM AUTHOR]

Free-surface flows, especially those associated with fluid-structure interactions (FSIs), pose challenging problems in numerical simulations. The authors of this work recently developed a smoothed particle element method (SPEM) to simulate FSIs. In this method, both the fluid and solid regions are initially modeled using a smoothed finite element method (S-FEM) in a Lagrangian frame, whereas the fluid regions undergoing large deformations are adaptively converted into particles and modeled with an improved smoothed particle hydrodynamics (SPH) method. This approach greatly improves computational accuracy and efficiency because of the advantages of the S-FEM in efficiently treating solid/fluid regions showing small deformations and the SPH method in effectively modeling moving interfaces. In this work, we further enhance the efficiency of the SPEM while effectively capturing local fluid information by introducing a multi-resolution technique to the SPEM and developing an effective approach to treat multi-resolution element-particle interfaces. Various numerical examples demonstrate that the multiresolution SPEM can significantly reduce the computational cost relative to the original version with a constant resolution. Moreover, the novel approach is effective in modeling various incompressible flow problems involving FSIs. [ABSTRACT FROM AUTHOR]

As a popular meshfree particle method, the smoothed particle hydrodynamics (SPH) has suffered from not being able to directly implement the solid boundary conditions. This influences the SPH approximation accuracy and hinders its further development and application to engineering and scientific problems. In this paper, a coupled dynamic solid boundary treatment (SBT) algorithm has been proposed, after investigating the features of existing SPH SBT algorithms. The novelty of the coupled dynamic SBT algorithm includes a new repulsive force between approaching fluid and solid particles, and a new numerical approximation scheme for estimating field functions of virtual solid particles. The new SBT algorithm has been examined with three numerical examples including a typical dam-break flow, a dam-break flow with a sharp-edged obstacle, and a water entry problem. It is demonstrated that SPH with this coupled dynamic boundary algorithm can lead to accurate results with smooth pressure field, and that the new SBT algorithm is also suitable for complex and even moving solid boundaries. [ABSTRACT FROM AUTHOR]