Recently the world has seen tremendous growth in oil and gas production from unconventional reservoirs, such as tight-gas sands, gas and oil shales, and heavy oil and tar sands. This growth can be attributed in part to technological developments in fracturing and stimulating well production, and in part to the availability of oil and gas resources in shale. Around 75% of the sedimentary rocks on earth are clastic nanoporous rocks, which are often referred to as "shales". These shales contain most of the world's oil and probably gas resources as well, but only a small fraction can be recovered using the latest technology – typically around 5% for oil and 20% for gas. What seems to be missing is an understanding of the fundamental interactions between fluids and rock surfaces, and of the dynamics of fluids in pores with sizes ranging from a few nanometers (10−9m) to micrometers (10−6m). Filling these gaps in our knowledge will help to develop improved recovery strategies and better prediction of production.
Here we present a multiscale simulation approach meant to bridge these gaps. All Atom Classical Molecular Dynamics (AAMD), a simulation method at the scale of atoms and molecules, is used to tune parameters used in Many-Body Dissipative Particle Dynamics (MDPD), which is a mesoscale simulation method. In nanopores, a majority of the fluid interacts with the rock surface and therefore cannot be treated as bulk fluid. To capture resulting phenomena, a low level of coarse-graining was used for MDPD. This, however, limits the domain of two-fluid MDPD simulations to a single capillary. To reach the scale of true porous media, we extract parameters describing the dynamic effects of fluid-fluid interface shapes on capillary pressures, and these are used in a dynamic capillary network simulation of fluids moving through pore spaces.
Results from MDPD simulations show the extraction of dynamic variables, such as dynamic contact angle and capillary pressure as a function of speed and pore diameter, to feed a capillary network model. The dynamic network model results then reveal how modifying the molecular interactions manifests at the pore scale.
With our hierarchical multiscale approach, we can bridge the gaps from molecular-level, chemical phenomena up to flow in porous media in a physically and thermodynamically consistent manner and with a minimum of unknown parameters. We can thus investigate not only if but why and how EOR strategies for shale succeed or fail.