Coupling image analysis and thermo-mechanical simulation results to produce a model of the fracture network in a nuclear glass canister

Highlights: • Nuclear glass fracture network is described by morphological mathematics. • Results of thermo-mechanical simulations are used as the proxy model. • Equiprobable realizations of equivalent fracture network are realized for glass canister. - Abstract: Understanding nuclear glass alteration in a fracture network of a block of nuclear glass is critical in predicting the release of the radionuclides in the subsurface and is relevant for the security of nuclear waste disposal. Here, we present an image processing-based fracture network characterization workflow that enables us to characterize a glass block fracture network, establish a link between a physical parameter representing an internal state of glass structural relaxation and an internal structure of the block fracture network, and generate multiple realizations of equivalent fracture networks by considering the variability of the fracture network parameters such as fracture distribution and apertures. The proposed workflow includes the following stages: (i) fracture network extraction, (ii) measurement of apertures of all distinct ramifications of the fracture network, (iii) image segmentation, (iv) anisotropy characterization, (v) passage to segment representation with the intention of modeling block equivalent fracture networks by using an algorithm of mesh growth, (vi) geostatistical modeling of average apertures and variables responsible for fracture spatial distribution, and (vii) simulation of equiprobable realizations of fracture network equivalent tessellations. Given this processing, the workflow is able to provide data for further steps including upscaling of diffusivity and permeability, and it facilitates reactive transport modeling at the glass block scale. As an illustrative example, we apply the proposed sequence of actions to a two-dimensional (2D) high-resolution image of two blocks of nuclear glass. This application brings into service both hard data obtained by direct measurement of the fracture network and soft physics-based explanatory data and considers the variability in the fracture network. The results show that implementation of the devised procedure presents an opportunity for better understanding of the impact of fracturing on aqueous alteration of glass.