Where is the Water? - A Physical Analysis of Hydraulic Fracturing Processes

Abstract

In hydraulic fracturing (HF) operations, large quantities of expensive water are injected into a rock mass to cause breakage, intended to create new flow pathways to allow fluid extraction. It is common to observe a significant discrepancy between the water volumes injected and those recovered, posing a question as to where the ‘lost’ water is located (the void volume of a notional HF is much smaller than that of the total water volume injected). Various explanations exist for the fate of the lost water, such as: it goes into the matrix pores; it goes into pre-existing natural fractures; it moves into the overburden (and represents a pollution risk). All of these ideas lead to physical consequences that can be assessed via energy-budget considerations, in which the input energy of the HF injection process (1/2 x volume x pressure, of the injected water) provides an upper energy bound to the set of processes that are imagined to occur in any hypothesized explanation. For the water to go into already-water-filled pore space, it is simple to calculate the resulting pressure based on the fixed (or expandable) pore (and/or natural-fracture) volume, and the compressibility of water. For a small HF injection volume, and considering the entirety of the rock volume between fracture stages as the possible host rock, the resulting pore/fracture pressure calculates out at 2-4 (or more) times the injection pressure - so that idea is readily rejected (with implications for ‘leak-off’ concepts). The much-larger pore-volume of the overburden could, in total, allow the lost water to move there and ultimately to the surface, but the rates do not allow this to be a feasible answer during the time when the well is pressurized. Therefore, new space must be created to contain the injection water during the time of the HF operation. This likely occurs via creation of new fractures, or opening old ones, across a large rock volume. The energy costs to make these new spaces are, again, limited by the energy budget, and the calculations show that the process must be much more efficient than simply making enough new fractures. The only known deformation process that meets these requirements is the non-uniform straining of a pre-fractured rock mass, a topic that has been examined in fractured-reservoir studies. By formulating an optimization problem, limited by the total energy, the feasible outcomes can be well-constrained, and hypothetical answers can be eliminated if they require more energy than is available. All of the physically-possible processes (even as partial solutions to the water question) demand that the stimulated volume will experience a volumetric increase, with important consequences for changes to bulk properties (eg seismic) and to assessment of the integrity of caprocks. HF analysis needs to adopt this type of whole-system approach that has proven valuable in other geomechanics-dominated problems.

AAPG Datapages/Search and Discovery Article #90350 © 2019 AAPG Annual Convention and Exhibition, San Antonio, Texas, May 19-22, 2019