A new computer-controlled laboratory technique has been developed to measure propped fracture width and embedment in weakly consolidated cores or unconsolidated sands under simulated down-hole conditions. Previously, laboratory studies on cores had determined embedment in hard rocks where embedment was normally limited to 50% of the proppant grain diameter. Several studies also indicated the importance of embedment with one monolayer or less of proppant coverage. The effects of water saturation and fracture-fluid filtrate on formation softness and embedment have not been previously published. Consequently, the objectives of the current paper are to extend previous research results to include soft, weakly consolidated cores and unconsolidated sands with multiple proppant layers. The influence of water saturation and dynamic fluid leakoff on embedment are also shown to be important.
The current investigations indicate that embedment becomes a problem when the Brinell hardness (BH) of the formation is less than about 20 kg/mm2 or when the static Young's modulus of formations cores is less than about 2 million psi (13 GPa). Embedment has been determined for cores with static Young's moduli between 0.1 and 1.4 million psi (0.7 to 9.6 GPa). In soft and wet sandstone, embedment can reduce fracture width up to 60% or more for proppant concentrations of 2 lbm/ft2. For unconsolidated sands, embedment is influenced by fracturing-fluid type, water saturation, and downhole conditions. Cyclic loading conditions associated with well shut-ins also increase embedment in unconsolidated sands. This paper reviews and discusses test data on formation cores from south Texas, New Mexico, the North Sea, and the Gulf of Mexico. Most of the commercial fracture-design programs neglect embedment problems in calculating fracture width, while other fracture simulators contain allowances for embedment.
Proppant embedment is a more significant problem today than in the past because of fracturing-stimulation treatments in weakly consolidated formations. Unlike previously published investigations on hard rocks, embedment can be as high as several proppant-grain diameters in some weakly consolidated sandstones. Proppant embedment can reduce fracture width from 10 to 60% with subsequent reduction of productivity from oil and gas wells in weakly consolidated sandstone. Simple estimations with a parallel-plate model indicates that fluid flow to the well will be reduced proportionally to the cube of the fracture width. Consequently, a 20% reduction in fracture width might restrict fluid flow and recovery by 50 to 60%.
To evaluate embedment in soft rocks and unconsolidated sands, a new computer-controlled apparatus was designed to determine propped fracture width and embedment as a function of closure stress. The effects of fluid type and variable proppant densities were also investigated. The dynamics of fluid flow and proppant settling can be evaluated as the fracture face meets the suspended proppant grains that support the closure stress. This paper also reviews the effects of cyclic loading conditions caused by simulating production and well shut-ins.
Yarbrough, McGlothlin, and Muirhead1 studied the effects of proppant embedment for the Lost Hills field, California. Embedment was considered to be a problem for low-formation-hardness values and proppant coverage less than 0.75 monolayers. Volk et al.2 studied embedment of sintered bauxite into Berea sandstone and shale with closure stresses up to 10,000 psi using one monolayer. The percent of fracture closure depended on the roughness of the fracture face and proppant diameter. Much and Penny3 reported on a study of fracture conductivity using an API proppant conductivity cell with Ohio sandstone and several proppants for closure stresses of 5,000 and 8,000 psi. Some of the proppant conductivity loss was associated with embedment of 0.008 in. (0.32-grain diameters) and crushing of sand. They also demonstrated that the filter-cake buildup can reduce fracture conductivity by 50%.
Snow and Hough4 determined embedment in chalk formations in the North Sea with BH values between 10 and 35 kg/mm2. In a more recent laboratory study, Milton-Taylor et al.5 determined that proppant pack stability in flowback tests depended on proppant size and rock hardness. Martins, Leung, and Jackson6 indicated lost fracture width caused by embedment for proppant concentrations between 0.5 to 1 lbm/ft2. Harley and Bosma7 reviewed embedment problems studied in a non-API conductivity cell with rock-hardness values between 50 and 1,000 MPa (5 and 100 kg/mm2) in a chalk reservoir.
Recently, Park8 measured embedment of 20/40 quartz gravel for confining stress of 1,000 psi in unconsolidated sands. Park noted the lack of embedment at high-flow-rate conditions and at low confining stress for nonspherical gravel. Additional papers6–12 have discussed the importance of embedment on field performance. The current work has recently (in 1997) been extended to include unconsolidated sands,12 where the formation hardness is only 0.1 kg/mm2 (140 psi), and different fluids and proppant sizes are investigated.
Currently, some fracture designs ignore embedment in calculating the propped fracture width, while others contain allowances for embedment at 2 lbm/ft2. If commercial software programs for fracture designs assume no embedment, this corresponds to assuming that all rock formations have the hardness of steel. Formation-rock-hardness values, as measured on a ball penetrometer (i.e., BH), can vary from about 0 to about 400 kg/mm2 (5.69×105 psi) . The formation-hardness value is a more important factor today because of recent industry trends to fracture softer, weakly consolidated, and higher-porosity formations where BH values are generally less than about 20 kg/mm2 (28,400 psi).
The current laboratory tests were performed on unconsolidated sands and soft, weakly consolidated sandstone with static Young's modulus values between 0.1 and 1.4 million psi, closure stresses up to 10,000 psi (70 MPa), and proppant concentrations of 2 to 4 lbm/ft2. Both dry and water-saturated cores were evaluated. The following two sections describe the testing technique and formation properties. Test data on cores from south Texas, the Gulf of Mexico (GOM), the North Sea, and New Mexico are reviewed in the section on laboratory test cases.