Efficient high-dimensional material reliability analysis with explicit voxel-level stochastic microstructure representation

A novel efficient methodology for probabilistic material reliability analysis considering fine-scale microstructure stochasticity is proposed in this paper. Integrated computational engineering requires multiscale capabilities to enable design and validation. Two critical challenges are identified: handling uncertainties from microstructures properties; the “curse of dimensionality” solvers. The study addresses these two challenges. First, an analytical hierarchical uncertainty quantification method explicit stochastic representation at voxel-level. hierarchy both phase maps within each modeled using Gaussian mixture random field. Analytical approximation arbitrary non-Gaussian field derived, which can facilitate computation gradient information optimization. Following this, solver adjoint first-order combining importance sampling derived by formulating as a constrained optimization problem. used efficiently evaluate responses exact gradients with help Several numerical examples calculation high-dimensional (voxel-level) fields subsequently employed demonstrate validate methodology. results quantitatively compared those obtained via classical method, direct Monte Carlo simulation, subset sequential method. comparisons indicate that possesses high efficiency problems.