Effect of fluid viscosity and density on proppant transport in complex slot systems, The

Abstract

The main functions of hydraulic fracturing fluids are to create a fracture network and to carry and place the proppant into the created fractures networks, thus, keeping the fractures open and allowing hydrocarbons to flow from the reservoir into the wellbore. Many studies have been performed over the years to develop an ideal fracturing fluid system. Development focus has generally been on optimization of a fluid rheology that can carry and place the proppant into the created fractures with less damage to the formation and at a lower cost. The main goal of this research is to continue building the understanding and optimization of proppant transport in a complex hydraulic fracture network. Specifically for this research focus is placed on two different fluids, water-glycerin solution and water-sodium chloride solution, representing varying density and viscosity. The effects of changing fluid viscosities, densities, proppant densities, proppant sizes, proppant concentrations, and slurry injection rates on proppant transport were then investigated. This experimental work was performed using a laboratory size slot apparatus at the Colorado School of Mines. Four different proppants were tested, 100 mesh sand (2.65 sp.gr), 40/70 mesh sand (2.65 sp.gr), 40/70 mesh ceramic (2.08 sp.gr), and 40/70 mesh ceramic (2.71 sp.gr), at two slurry injection rates (1 and 2 gal/min) and two proppant concentrations (1 ppg (Cv=0.043) and 2 ppg (Cv=0.086)). Also, this study reviews the published correlations for proppant dune height and compares the calculated results with the lab results. The experimental results show that a water glycerin solution (viscosity of 4.3 cp) has a better capability of carrying the proppant into farther locations and to greater distances, which was ascertained by the weight of the collected proppant out of the sampling points. On the other hand, the results show that water-sodium chloride solution of 9.24 ppg density has less capability to carry the proppant deep into the fractures. The results also show that the proppant covered area inside all of the farther fractures (T-1, S-3, and T-2) from the injection point was greater using the water-glycerin solution that the water-sodium chloride solution, due to the ability of the high viscosity to keep more proppant suspended. For all tested proppants, the results show that increasing fluid density had no significant effect on the equilibrium dune height and proppant travel distance. Moreover, the results show that increasing the number of injected particles has a similar effect as that of increasing proppant concentration, both situations leading to quick dune build-up rates and increased proppant covered areas. For both tested fluids, the results show that increasing the injection rate has a significant impact on proppant transport. As the injection rate increases, the proppant dune height decreases and both the proppant covered area and the proppant collected from sampling points increase as well. It was found that smaller proppant sizes such as (100 mesh sand) transported deep into the slot system with lower dune height, as compared with larger proppant sizes (40/70 mesh). Also the results show that proppant density has a significant impact on proppant transport. Decreasing the proppant density resulted in increasing the proppant covered area inside the slots. This experimental work shows that viscosity has a greater impact on the proppant transport than fluid density does, thus implying a larger impact on the resulting fracture conductivity. All of the lab data were compared to published correlations that can predict proppant dune height inside the main fracture by Wang et al. (2003) and Alotaibi and Miskimins (2015). Both correlations showed values close to the laboratory measured values with some minor (< 2.5%) average percent difference. Overall, it appears that these correlations can be used to predict the equilibrium dune height inside the main fracture with low error values and without adjusting any parameters for the tested systems.