Laboratory Imaging of Stimulation Fluid Displacement from Hydraulic Fractures

Abstract

Abstract Laboratory experiments were conducted to physically investigate the processes governing stimulation fluid displacement from hydraulic fractures. Experiments were performed on two scales: meter-scale in a 1500 cm2 sand pack and core-scale in a 65 cm2 API linear conductivity cell. High-resolution light transmission imaging was employed at the meter-scale to visualize and quantify processes governing liquid displacement. For comparison, complimentary tests were performed using an API conductivity cell under ambient test conditions and at elevated closure stress. In these experiments viscous fingering and gravity drainage were identified as the dominant processes governing liquid displacement. Fluid viscosity was found to dictate the relative importance of the competing displacement processes and ultimately determine the residual liquid saturation of the sand pack. The process by which fluid displacement occurs was seen to affect the shape of both the gas and liquid phase relative permeability functions. Knowledge of such viscosity/relative permeability relationships may prove useful in bounding predictions of post-stimulation recovery of gels from the fracture pack. Introduction Hydraulic fracturing is an accepted and well-documented technique to stimulate production from gas reservoirs; however, gas productivity and recoverable reserves are often impaired by the poor recovery of stimulation fluids from the reservoir and hydraulic fracture. Design of enhanced stimulation fluid recovery strategies is complicated by the complexities inherent to the fluid/reservoir system and processes governing fluid displacement. Factors influencing system response include: viscous fingering, capillary trapping, filter-cake build-up, gel leak-off, and proppant pack/reservoir heterogeneity. In efforts to better predict poststimulation behavior of hydraulically fractured gas wells, investigators have demonstrated the ability to physically model the fracture with proppant packs under simulated reservoir conditions of stress and/or temperature. Such experiments have identified gelling agent type, gel viscosity (i.e., extent of fluid breaking), insoluble polymer residue, non-Darcy flow, and closure stress as effecting poststimulation fracture permeability. The extent of fracture permeability damage has been reported to range from negligible to near complete plugging of the fracture. Further work has been done to elicit total fracture/reservoir system response by including dynamic fluid loss processes. One limitation common to all is that experiments have been routinely conducted on core-scale samples using methods that only provide information on integrated system response. From such data it is difficult to uniquely discern the basic processes governing fluid displacement. Numerical experimentation has also been employed to investigate the displacement of stimulation fluids from hydraulically fractured reservoirs. Numerical simulations offer the advantage of being able to investigate the coupled behavior between the hydraulic fracture and invaded reservoir at the scale of interest. A variety of simulators have been used to investigate a wide range of issues. van Poollen investigated the effects of fracture orientation and fracture length on gel recovery from hydraulic fractures, while Tannich evaluated the effects of relative permeability reduction, liquid holdup in the well tubulars and non-Darcy gas flow. Holditch extended this work by investigating the combined effects of formation permeability darn age and relative permeability damage in the invaded matrix. Montgomery et al. and Sherman et al. considered the effects of fracture conductivity, formation darn age to the fracture face, and relative permeability hyseresis in the invaded zone. Other studies have shown unbroken fracture fluids significantly lower gas reserves, reduce gas flow rates, and delay peak gas production by weeks or months. P. 793