Investigating the structure of turbulent flows in porous media: an endoscopic particle image velocimetry (E-PIV) approach

Flow in porous media is of central interest in a diverse range of environmental and industrial applications such as hydrocarbon extraction, groundwater flow, packed-bed chemical reactors and heat exchangers. Most flows in natural river channels occur over porous media, such as poorly sorted gravels and sands, but, paradoxically, river beds are traditionally modelled as impermeable surfaces. As a result, questions relating to the nature of fluid exchange between the free flow and the river bed have been largely ignored. However, preliminary experimental data show that for the high Reynolds numbers typically found within natural rivers, the resulting flow within the macro-pores of the subsurface may be highly turbulent, with significant exchange into and out of the bed. Consequently, since large-scale coherent flow structures can be generated in these permeable beds, and turbulence plays an important role in the dynamics of the flow, the standard assumptions of laminar flow within the subsurface of these river beds are often erroneous. However, due to the challenging physical environment of river beds, very little data has been collected from the pore spaces within such sediments. This dearth of data has then restricted efforts to numerically model such flows, as baseline and calibration data are lacking, and thus progress in properly exploring the nature of pore flow in river beds has proved elusive. In the last decade efforts to quantify pore flow have been made using a range of non-invasive techniques, such as MRI (Magnetic Resonance Imaging) and RIM (Refractive Index Matching). However, the low spatio-temporal resolution of these techniques makes them unsuitable for the higher Reynolds numbers that are typical of river flows. In order to overcome these problems, we present details here of a new endoscopic technique that is able to visualise and quantify the hydrodynamic processes within the pore spaces of a fixed permeable flume-bed. Most importantly, the approach has the ability to characterise the spatial structure of the instantaneous flow together with its temporal evolution. Initial results reveal the presence of large coherent vortical structures that can be tracked over long time periods. Strong jet flows occur between the interconnected pores and appear to drive vortex formation and evolution. Local interstitial pressure gradients, driven by free-stream vortex structures, trigger the observed jet flow. We will illustrate the interaction between the jet flow and the structure of the bed, which appears to dictate the characteristics of the vortices that are generated, and that can be quantified using topological methods. Preliminary work will also be detailed on the acquisition of simultaneous flow data from above and within the permeable bed, which allows investigation of the links between the generation of turbulence across this interface.