Modeling and Prediction of Ground Shock Induced by Penetrating Weapons in Spatially Random Geologic Media

We propose a stochastic-finite-element based methodology to address these requirements: (i) Characterize material variability; (ii) Quantify corresponding ground-shock variability; and (iii) Design instrumentation plan. The site is discretized into N finite-elements; rock properties over this domain are characterized as the components of an N-vector stochastic field. The probabilistic characteristics of each component are estimated from analysis of field data and encapsulated as random parameters of the Cap model; spatial variability is captured by the correlation function. Explicit joints are sampled from their distributions, and their effects modeled by reduced strength of the affected elements. Monte-Carlo simulations of the ground-shock process are performed using the finite-element site model, with each simulation corresponding to a digitally-generated (realized) random sample function of the stochastic field. Realizations are obtained using two complementary techniques: Spectral and geostatistical. An adaptive, hierarchical procedure is proposed to accommodate the wide range of data (in)completeness in practice. The statistics of the ground-shock response and their spatial variability are used to support an instrumentation plan for given test objectives/constraints. It is suggested that the same stochastic techniques used to encapsulate and realize spatial material variability may be used to encapsulate response variability, leading to a systematical instrumentation methodology.