ThermoChemoMechanics in SPH for modelling of cement grouting

Natural materials, such as e.g. wood, stone and soil, exhibit a distinct pore space with its characteristics being directly linked to the mode of material genesis, such as e.g. growth, compaction, sedimentation, etc. In fact, these characteristics strongly affect the macroscopic properties of the material. On the one hand, higher porosity results in lower weight and enhances e.g. heat insulation as well as permeability. On the other hand, it may significantly reduce stiffness and strength of the material. One way to enhance the mechanical properties of porous materials is the injection of hardening fluids, such as cement slurry in case of cement grouting. The performance of the strengthened material (in size and mechanical properties) strongly depends on the applied pressure, the fluid and pore-space properties. Accordingly, in order to model and finally simulate the underlying injection process during e.g. cement grouting, detailed information about (i) the pore space and (ii) reactive fluid as well as the effect of reaction extent and temperature on the fluids viscosity are required. Within the present research project, these two challenges are addressed by employing a simulation technique allowing the simultaneous treatment of chemical reactions, heat transport and flow of fluids exhibiting evolving properties. For this purpose, the so-called smoothed particle hydrodynamics (SPH) method, e.g., applied in fluid mechanics, computer graphics, astrophysics, and more recently material technology, shall be extended for considering the evolution of properties of cement-based materials as a function of reaction extent and temperature. The foundation of the anticipated modeling and simulation approach is a thoroughly experimental characterization of both, the geometry of the injected porous material as well as the thermo-chemo-mechanical properties of the reactive fluid. For characterization of the pore space, high-resolution computer-tomography (CT) imaging shall be applied on selected materials, accompanied by topological data analysis (DTA) giving access to so-called persistence bar-codes (the fingerprint) of the respective pore space. For the simulation-based study of the injection process, various sets of artificial pore spaces shall be generated, with each statistical realization exhibiting the same distinct persistence bar-code. The combined variation of pore-space characteristics and other parameters, such as injection pressure, thermal conditions and material properties, shall reveal a better understanding of mechanism behind the transport of reactive fluids and, finally, provide basic relations between pore-space characteristics and injection behavior. The simulation framework shall be validated by comparing the numerically-obtained results with results from grouting experiments on two types of materials (i.e., granular material and cohesive material exhibiting an open pore space). In the course of this validation procedure, chemical-reaction and flow properties of the cement grout and the respective persistence bar-code of each tested material enter the simulation framework as input parameters.

X-ray microscopy and pore-space/solid-phase analysis As part of the project, a wide variety of materials were examined using an X-ray microscope available at the NanoLab at the University of Innsbruck. The spectrum of materials included various porous materials, as well as granular and ceramic materials, in order to subsequently create a broad pool of pore spaces for subsequent topological data analyses. A Python-based tool was developed within the project to characterize the pore space and the spatial distribution of the solid phases. In addition to the pixel-based data from X-ray microscopy, numerical methods were implemented to generate virtual pore spaces. The latter utilized so-called "annealing" and "machine learning" algorithms. An essential component of this program is segmentation and thus the differentiation of the various material phases. A novel algorithm that utilizes the distribution of the intensity gradient was developed within the project. Extension of the SPH method As part of the project, an in-house SPH program was developed using C++, which was extended towards consideration of the fluid-solid contact angle, the fluid-fluid interface, and coupled phenomena such as those arising during heat transfer and chemical reactions. For this purpose, the "Moving Least Squares" (MLS) correction was applied to the SPH method, which compensates for any irregularities due to particle discretization. This allows for the compensation of missing particles at the edge of the simulation domain and the exact reproduction of polynomials, improving the accuracy and convergence of the method. With regard to the simulation of coupled systems, the implemented MLS correction enables the application of stabilization methods known from FEA, since derivatives can be determined uniformly from the functional approximation, unlike conventional SPH. Biochar concrete as a four-phase material system Finally, a four-phase material system consisting of aggregate, cement paste, biochar, and air was investigated. The described phases exhibit a density difference, which allows for the proper segmentation of the phases. Furthermore, the amount of biochar present in biochar concrete can be reduced by thermal loading, increasing the pore space and thus the permeability. Within the scope of the project, sample material with variable porosity/permeability was created by the allowance of biochar (2.5 to 10 wt.%) and subsequent thermal loading (up to 450C). For this sample material, CT measurements were carried out to determine the microstructure and pore structure, as well as mechanical tests and permeability measurements (air permeability).