Computational methods for the dynamic compaction of porous solids

Abstract

A 78% dense porous aluminum material was chosen and run through a variety of simulations to predict its compaction under dynamic loadings. The results of these computations were compared to experimental data collected by Butcher in 1974 on exactly the same material [3]. The results show that the p-alpha method is extremely accurate and reliable for predicting the quasistatic compaction of a porous material, but fails to predict the subtle differences when the material is subjected to dynamic loadings. At stresses below 2kbar, the continuum mixture theory model correctly predicted dynamic compaction effects on the material density. However, at higher stresses, the model incorrectly predicted the onset of complete compaction and the results were skewed from the experimental data presented by Butcher. The models and methods Kodiak provides are constantly under development, and show promise toward reliably providing more advanced methods for simulation of shock physics problems and their applications.

Sandia National Laboratories has begun development on a new breed of computational and theoretical methods for solving high-stress, high-velocity shock physics problems. The Kodiak code is a current research and development tool in use at the Labs for solving exactly these types of problems. Of particular interest is the compaction behavior of porous materials when subjected to dynamic shock loadings. In the past, the p-alpha method developed by Herrmann has been the tried-and-true model for predicting the compaction of porous materials [1]. Implemented in Kodiak is a model known as continuum mixture theory, as presented by Baer in A Two-phase Mixture Theory for the Deflagration-to-Detonation Transition (DDT) in Reactive Granular Materials [2]. The model uses a rate-dependent volume evolution equation to track changes in the volume fraction of the materials in time.