Investigations of seismicity induced by hydraulic fracturing in naturally fractured reservoirs based on moment tensors

https://doi.org/10.1016/j.jngse.2020.103448Get rights and content

Highlights

  • The algorithm for evaluating seismicities was proposed by an independently developed subroutine.

  • The proposed modes accurately identify the four types of hydraulic-natural fracture interactions.

  • The characteristics of seismic events generated by the interaction modes can be determined by the moment tensor method.

  • The effects of stress difference, approaching angle and friction coefficient on seismicities was carried out.

Abstract

In oil and gas production, natural fracture (NF) activation and the seismicities induced by hydraulic fracturing can cause unpredictable damage and even catastrophic consequences to reservoirs, which seriously threatens the safety of construction personnel and equipment. In this paper, the seismicities induced by hydraulic fracturing in naturally fractured reservoirs were investigated based on the moment tensor. Firstly, an algorithm for evaluating seismicities was generated using an independently developed subroutine. Then, a systematic simulation was implemented to explore the characteristics of seismic events (SEs) produced by the interaction of hydraulic fractures (HFs) and natural fractures. Finally, a sensitivity analysis was performed to examine the effects of the stress difference, approaching angle (θ), and friction coefficient on seismicities under different working conditions. The results show that there are four modes of interaction between hydraulic fractures and natural fractures. In the Type II mode, an NF can combine with other NFs to form fracture networks, however SEs produced in this mode are the most dangerous. In the Type IV mode, SEs are fewer and smaller in magnitude compared to those in the Type II mode. Additionally, in the Type IV mode, HFs can also cross NFs and thus become interconnected with other NFs. When θ ≤ 30°, the stress difference becomes the main factor controlling the seismicities. Meanwhile, when 45°≤θ ≤ 60°, the stress difference and the approaching angle jointly affect the interaction modes. The NF friction coefficient also plays a role in SEs at this stage, however its effect is not prominent. When θ ≥ 75°, the modes produced by the HF–NF interactions belong to Type IV or are in a transition to Type IV.

Introduction

Hydraulic fracturing is a key technology that is used to increase the production of unconventional oil and gas (Zhu et al., 2014). The goal of hydraulic fracturing is to interconnect hydraulic fractures (HFs) and natural fractures (NFs) in reservoirs in order to form a complex fracture network and thus promote the permeability and conductivity of the reservoir (Osiptsov, 2017; Liu et al., 2020). However, hydraulic fracturing operations inevitably cause damage to the surrounding natural environment (Jackson et al., 2015; Liu et al., 2017). Injection-induced NF activation and seismicities are dangerous dynamic disasters facing unconventional energy exploitation, and continually threaten the safety of oil and gas fields (Bao and Eaton, 2016). It has been reported worldwide (e.g., Ohio, United States; Western Canada Sedimentary Basin; Southwestern China) that several seismicities with magnitude M>3 have been induced by hydraulic fracturing in recent years (Atkinson et al., 2016; Davies et al., 2013; Skoumal et al., 2015). Therefore, obtaining a better understanding of the characteristics of seismic events (SEs) induced by HF–NF interactions can provide some reference for the monitoring, forecasting, and prevention of dynamic disasters that occur during hydraulic fracturing.

The assessment of injection-induced seismicities, especially accurate locations and the prediction of the maximum magnitude of SEs during hydraulic fracturing, is essential for mitigating the hazards to people and infrastructure due to such events. McGarr (2014) pointed out that the maximum magnitude of injection-induced SEs was controlled by the total volume of fracturing fluid. However, Sumy et al. (2014) disagreed with this viewpoint and concluded that seismicities were caused by Coulomb stress transfer. Meanwhile, Petersen et al. (2016) suggested using a wide range of uncertainty to characterize the level of seismicity.

Seismicities can be induced both by the geological conditions of a reservoir and by the industrial operation parameters. For example, the fluid pressure changes greatly when a reservoir is subjected to large stress or develops faults (Rutqvist et al., 2013). Schultz et al. (2015) concluded that SEs produced in the Western Canada Sedimentary Basin were closely related to large-scale NFs (i.e., faults) generated during fracturing. SEs produced in medium- and low-permeability reservoirs are often induced by linear relaxation resulting from pore pressure perturbations (Shapiro et al., 2005). In most cases, the migration of induced seismicities is related to the propagation of HFs perpendicular to the fault (Xie and Min, 2016). Fluid pressure along the fault plane may increase during or immediately after fracturing, and if the fluid pressure exceeds a critical value the fault is activated, thus inducing seismicities (Flewelling et al., 2013; Li et al., 2015). This knowledge helps to understand the generation mechanisms of seismicities and plays a guiding role in identifying areas with higher seismogenic potential.

If pre-existing NFs are activated by fluid injection, larger-magnitude SEs will occur frequently; this is an important cause of dynamic disasters in unconventional oil and gas exploitation (Bommer et al., 2006; Hu et al., 2017). Furthermore, the generation of the complex fracture networks that are required to increase the capacity of unconventional oil and gas production is highly dependent on HF–NF interactions. However, NFs exist in various forms, e.g., open, closed, or sealed by precipitated minerals. Besides, the existing theoretical models of hydraulic fracturing-induced seismicities are essentially based on statistical data due to the complexity and invisibility of HF propagation. The results of such theoretical models depend primarily on the empirical analysis of the SE incidence observed in each wellbore. Therefore, current models for predicting the likelihood, incidence, and magnitude of injection-induced SEs are insufficient. Moreover, it is relatively difficult to quantitatively evaluate SEs generated by HF–NF interactions through laboratory experiments.

In recent years, numerical simulations have been widely used to study seismicities induced by hydraulic fracturing in reservoirs. For example, Zangeneh et al. (2013) used the Universal Distinct Element Code (UDEC) to study the effects of geometrical features of NFs and injection stress on seismic energy release during hydraulic fracturing. The hydro-mechanical coupling simulator (TOUGH-FLAC) was used to study the fault slip and seismicity caused by the direct injection of fluid into a fault zone during shale gas exploitation in the Northeastern United States (Rutqvist, 2011). Furthermore, Bommer et al. (2006) combined the finite element method (FEM) and the pyroclastic displacement discontinuity method to analyze the possibility of formation failure caused by hydraulic fracturing, and also validated the relationship between the stimulated volume and the permeability enhancement through calculation.

Although a variety of methods have been adopted to study seismicities induced by hydraulic fracturing during unconventional oil and gas exploitation, there are still some uncertainties and knowledge gaps in our understanding of injection-induced seismicities and their magnitude distribution. The implementation of hydraulic fracturing in naturally fractured reservoirs can produce different interaction modes. It is unclear what the differences are in the characteristics of SEs produced in different modes, and how engineering and geological factors affect injection-induced seismicities. Answering the above questions is especially important to inform the design of fracturing in unconventional oil and gas exploitation. In this paper, an explicitly coupled hydro-mechanical model (ECHM) was established based on the discrete element method (DEM). Then, a moment tensor analysis was conducted on the calculation results of different modes, and the reliability of microscopic parameters and the moment tensor method were both verified. Furthermore, the characteristics of SEs in different interaction modes were analyzed, including the relationship between the magnitude and frequency of SEs, the cumulative frequency of SEs, and the number of HFs contained in SEs. Additionally, different influential factors that influence seismicities, i.e., stress difference, approaching angle, and friction coefficient, are also discussed.

Section snippets

Numerical simulation procedure

The bonded-particle model (BPM), a typical DEM, was first proposed by Cundall and was subsequently rapidly developed by Potion and Cundall (2004), who detailed the motion principle of particles. The properties of rock materials in BPM depend primarily on the stiffness and strength of particles and bonds. There are two types of BPM, namely the contact bond model (CBM) and the parallel bond model (PBM). In the CBM, only normal and shear forces can be transmitted at the contact point, while in the

Model validation for interaction between HFs and NFs

Three sets of parameters of the proposed ECHM need to be calibrated, namely the reservoir matrix parameters, the NF parameters, and the fluid flow parameters. In our previous study, the microscopic parameters of the BPM were validated through a confining-pressure experiment of specimens. At the macro level, the sandstones from the Shanxi Formation were fabricated into standard specimens (φ50 × 100 mm) with a uniaxial compressive strength of 17.3 MPa, a Young's modulus of 21.34 GPa, a cohesive

Results and analysis

All working conditions were calculated based on the proposed method. The microscopic parameters of the naturally fractured reservoirs and the fluid flow are given in Table 1, Table 2 and the numerical simulation conditions are listed in Table 3. The HF–NF interaction is a comparatively complex process [Fig. 8(a)], and was divided into two steps by Gu et al. (2012). First, NFs are affected by the stress field generated by NFs. At this time, the fluid pressure of NFs is regarded as 0. There are

Effect of stress difference (Δh)

Fig. 17 displays the effects of Δh on the HF propagation and seismicities in the naturally fractured reservoirs under the conditions of constant θ (θ=45° and 75°) and fn (fn=0.58). When θ = 45°, the magnitude of SEs produced by the HF–NF interactions shows a tendency of increasing first and then decreasing with increasing Δh. As Δh increases to 7 MPa (Case 10) and 10 MPa (Case 11), SEs produced by the HF–NF interactions turn into Type II. As the magnitude of SEs increases strongly, the b-value

Conclusions

In order to investigate the seismicities induced by hydraulic fracturing in naturally fractured reservoirs, an ECHM was established based on a DEM. Moment tensor analysis was carried out on the fractures generated in the model, and an algorithm for evaluating seismicities was generated using an independently developed subroutine. The validity and reliability of the model parameters were tested by comparing the numerical simulation results of HF–NF interactions with the results of the

CRediT authorship contribution statement

Zhaohui Chong: Writing - original draft, Conceptualization, Methodology, Software, Data curation, Investigation, Formal analysis. Qiangling Yao: Writing - original draft, Supervision, Methodology, Conceptualization, Data curation. Xuehua Li: Writing - review & editing, Supervision, Conceptualization. Jia Liu: Software, Validation, Visualization.

Declaration of competing interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This research is supported by the National Natural Science Foundation of China (51904289 and 51874285) and the China Postdoctoral Science Foundation (2018M640535).

References (34)

  • G.M. Atkinson et al.

    Hydraulic fracturing and seismicity in the western Canada Sedimentary Basin

    Seismol Res. Lett.

    (2016)
  • X. Bao et al.

    Fault activation by hydraulic fracturing in western Canada

    Science

    (2016)
  • T.L. Blanton

    Propagation of hydraulically and dynamically induced fractures in naturally fractured reservoirs

  • Z. Chong et al.

    Numerical investigation of acoustic emission events of argillaceous sandstones under confining pressure

    Math. Probl Eng.

    (2017)
  • S.A. Flewelling et al.

    Hydraulic fracture height limits and fault interactions in tight oil and gas formations

    Geophys. Res. Lett.

    (2013)
  • H. Gu et al.

    Criterion for fractures crossing frictional interfaces at non-orthogonal angles

  • H. Gu et al.

    Hydraulic fracture crossing natural fracture at nonorthogonal angles: a criterion and its validation

    SPE Prod. Oper.

    (2012)
  • Cited by (13)

    • Studying the direction of hydraulic fracture in carbonate reservoirs: Using machine learning to determine reservoir pressure

      2023, Petroleum Research
      Citation Excerpt :

      Thus, the Rastegaev et al. (2019) indicate that within large oil production regions, there are areas with unfavorable seismogeological conditions in which the practical application of the microseismic monitoring method of hydraulic fracturing is impractical. Other researchers speak about uncertainty in the interpretation of data from seismic methods for monitoring hydraulic fracturing is (Birialtcev et al., 2017; Li et al., 2019, 2020; Chong et al., 2020). The authors point out that the influence of engineering and geological factors on seismicity is insufficiently studied.

    • Crack propagation mechanism in rock-like specimens containing intermittent flaws under shear loading

      2022, Theoretical and Applied Fracture Mechanics
      Citation Excerpt :

      The simulation of AE events could not only enhance the understanding of failure mechanism of jointed rock masses, but also establish a solid foundation to investigate the injection-induced seismicity due to its capability of simulating micro seismicity. While recent work mainly focuses on the distribution of AE event magnitude and the effect of AE evolution on crack propagation [39–44], the relationships of crack propagation and the component of moment tensor are rarely mentioned. Therefore, further component analysis of moment tensor is necessary to facilitate the understanding of the source mechanism of rock shear failure.

    • Investigation of the effect of different injection schemes on fracture network patterns in hot dry rocks - A numerical case study of the FORGE EGS site in Utah

      2022, Journal of Natural Gas Science and Engineering
      Citation Excerpt :

      Creating a high-density fracture network without short-circuiting between stages, and to provide highly conductive flow pathways for fluid to exchange heat are key factors for EGS construction (Hofmann et al., 2014; Wu et al., 2016). However, hydraulic stimulation aimed at creating commercial reservoirs may generate seismic events, resulting in a negative public perception of EGS and termination (preliminary) of the project (Chong et al., 2020; Grigoli et al., 2018; Lee et al., 2019; Alghannam and Juanes, 2020; Tdr et al., 2020). The goal of stimulation is to implement advanced injection strategies to control the propagation of fractures, while reducing the risk of unnecessary seismic events (Zang et al., 2019; Hui et al., 2021).

    • Effect of heterogeneity and injection borehole location on hydraulic fracture initiation and propagation in shale gas reservoirs

      2021, Journal of Natural Gas Science and Engineering
      Citation Excerpt :

      Numerous DEM simulations have been conducted to investigate hydraulic fracturing in different rock reservoirs (Tomac and Gutierrez, 2020; Liu et al., 2020a; Tian et al., 2020). The influence of the distance between the borehole and the preexisting fracture on the seismicity has been studied (Chong et al., 2020), but rare attention has been paid to the influence of injection borehole location in the process of hydraulic fracturing in shale reservoir. Most of simulations on shale reservoirs only considered the heterogeneity induced by joints, the influence of particle heterogeneity is rarely studied (Dou et al., 2021; Zhang et al., 2021; Zhou et al., 2017).

    View all citing articles on Scopus
    View full text