Comparison of flow and transport experiments on 3D printed 'rocks' with direct numerical simulations

3D printing technology has the potential to revolutionise modelling of fluid flow and mass transport in porous media. Using 3D printing to replicate pore geometries from real rocks quickly and cheaply, and being able to assign specific properties to the samples, would enable us to repeat experiments where things such as permeability, porosity, reactivity, or wettability are known a priori. Destructive tests, such as reactive transport experiments, could then be performed using the same, realistic, initial geometry, in a repeatable fashion; in addition, specific properties of the porous media (e.g. the reactivity of individual grains) could be altered in a controlled way. Similarly, two-phase flow experiments could be carried out where relevant properties (e.g. individual grain wettability) are modified between experiments. Results from such experiments would shed new light on key physio-chemical processes occurring at the pore-scale during multi-phase reactive flow and would allow us to validate the suite of emerging direct numerical simulation techniques (e.g. Lattice Boltzman, Volume of Fluid) currently used to model pore-scale flow and transport. We have developed a novel experimental setup to investigate single-phase flow and transport through translucent 3D printed and Perspex samples and to visualise the experiments for comparison with numerical simulations. Dyed fluid is injected at one end of the sample and the behaviour of the system is recorded using a digital camera situated directly above it. Image processing techniques are employed to quantify dye concentration and location through time. We use the finite volume method to simulate the flow experiments in OpenFOAM, using the same input geometry that was used to print the sample. Comparison of experimental results with simulations enables us to identify similarities and differences between flow and transport observed in the 3D printed samples and behaviour expected from the numerical simulations.