Evolution of Proppant Embedment and its Impact on Shale Permeability

ABSTRACT: Hydraulic fracturing is important for modern gas extraction. However, the proppant embedment affects gas recovery significantly. This study investigates the impact of proppant embedment on permeability over time in shale reservoirs. The study utilized a series of experiments to determine the impact of proppant embedment on permeability over time and designed new methods to induce long-term proppant embedment in proppant-supported fractures. The results showed that the permeability loss after 8 weeks of compression ranges from 4.38 to 6.44 mD, and the percentage ranges from 6.74 to 8.89%. The study also found that the average proppant embedment depth reaches 87.96 μm after 8 weeks compression. The correlation between embedment depth and compression day can be suitably modeled with a logarithmic function. These findings can help optimize production in shale reservoirs by providing a better understanding of proppant embedment and its impact on gas recovery. 1. INTRODUCTION Hydraulic fracturing is a vital technique to increase the gas production from unconventional reservoirs, which are characterized by low permeability and porosity such as shale and coal seam. This technique creates a network of fractures in the formation, which increases the production rate of natural gas. Due to the high closing stress underground, proppants, typically sands and ceramic grains, are also mixed with the fracking fluid and pumped into the fracture to prevent the fracture from closing and allow gas to pass through the fracture continually. However, the fracture still trends to close over time due to some behaviors in the formation such as proppant flow back (Chuprakov et al., 2020; Hu et al., 2014; Murphey et al., 1989; Parker et al., 1999), proppant fines generation and migration (Kong et al., 2023; Pan et al., 2018; Wang et al., 2020), proppant crush resistance (Liang et al., 2015; Liang et al., 2016; Stephens et al., 2006; Zheng et al., 2018) and proppant embedment (Ghanizadeh et al., 2016; Huang et al., 2019; Liu et al., 2021).