This paper successfully simulated the hydraulic fracturing process in lab-scale coal samples and induced seismicity using the self-developed code in particle flow code (PFC). Numerical simulations of fluid injection, fluid transport and seismic response are achieved simultaneously. The code has been applied to simulate water injection into an intact coal sample and coal samples with one pre-existing fracture. Model results satisfy observations from theoretical assumptions and laboratory experiments, which justifies the reliability of the proposed self-developed code. They indicate that hydraulic fractures on the side of the sample with a pre-existing fracture present different cracking propagation pattern compared to the other side with no pre-existing fracture. The pattern can be used to predict the impact of pre-fractures on the seismic extension-to-compression ratios and the aperture of hydraulic fractures.
Hydraulic fracturing is widely used in the energy resource sector, not only to enhance oil and gas production in tight reservoirs but also to facilitate caving and prevent dynamic disasters in mining (Huang & Liu 2017). However, due to the geological complexities, although there are novel failure mechanism theories (Fischer & Guest 2011), it is still challenging to determine the in-situ conditions for various modes of rock failure at target depths and fluid pressure regimes in a reservoir. Seismic monitoring is an important tool to infer subsurface rock failure behaviour reservoir properties (e.g., stress, permeability, connectivity) adjacent to induced fractures (Rutledge et al. 2004 and Guo et al. 2021). However, the conclusions drawn cannot be easily validated and widely transferred due to the geological complexity and uniqueness of specific regions.
Numerical modelling has provided a unique strength to simulate rock failure under various stress and geological conditions. The discrete element method (DEM) has become increasingly popular in recent years because it allows a better understanding of crack initiation and growth in rocks with pre-fractures or flaws (Potyondy 2010). In addition, a systematic set of methods and theories have been proposed in DEM to not only numerically reproduce seismicity but also to model the seismicity induced by fluid injection in naturally fractured rocks (Hazzard & Young 2000 and Al-Busaidi et al. 2005). For instance, Zhao & Young (2011) simulated fluid injection via PFC to study different modes of interaction between pre-fractures of different orientations and hydraulic fractures under various stress boundaries. The corresponding seismic events and their moment tensors were calculated and visualised on simulated rock samples.
