Scholarlink

<p>Quantitatively separating the coupled expansion–seepage contribution within non-blasting SC-CO<sub>2</sub> fracturing mechanisms is essential for advancing CCUS-enabled rock-breaking technology. In this paper, SC-CO<sub>2</sub> fracturing tests were conducted on granite under different injection rates (5, 10, 15, 20, and 25&#xa0;ml/min), drilling depths (5, 10, and 15&#xa0;mm), and fluids (SC-CO<sub>2</sub> vs. H<sub>2</sub>O) conditions. Then, the multiple descriptors, including the difference of 3D and 2D fractal dimension (Δ<i>D</i>), skeleton connectivity ratio (<i>R</i><sub><i>m</i></sub>), surface roughness (<i>S</i><sub><i>q</i></sub>), directional anisotropy (<i>I</i><sub><i>ani</i></sub>), and peak break pressure (PBP), were used to quantitatively characterize the reconstructed 3D fracture network and loading response. Moreover, a multiscale complex fracture index (MCFI) was formulated to integrate Δ<i>D</i> and <i>R</i><sub><i>m</i></sub> as backbone measures of the fracture network and to incorporate seepage-related roughness (<i>S</i><sub><i>q</i></sub>) and expansion-related directionality (<i>I</i><sub><i>ani</i></sub>) through two weighting coefficients (<i>α, β</i>). The coefficients were inferred using a Bayesian inversion framework coupled with a logarithmic MCFI–PBP mapping, enabling probabilistic calibration and mechanistic decoupling of seepage- and expansion-dominated contributions. The results showed the MCFI-PBP mapping with high fidelity (<i>R</i><sup>2</sup> = 0.91, RMSE = 0.56&#xa0;MPa) and narrow 95% credible intervals (± 0.30–0.50&#xa0;MPa), outperforming single-descriptor models. Posterior estimates yielded <i>α</i> = 2.49 and <i>β</i> = 2.60 indicate a near-balanced contribution of seepage- and expansion-related effects within the SC-CO<sub>2</sub> fracturing granite tests, while the inferred Log-elasticities coefficient suggest that expansion-driven processes act as the primary trigger that enables subsequent seepage effect development. Within the present experimental design, an MCFI value exceeding 5.50e-3 is associated with the most developed fracture network response observed at a drilling depth of 10&#xa0;mm and an injection rate of 20&#xa0;ml/min. This study proposes a novel Bayesian-calibrated framework for SC-CO<sub>2</sub> fracturing tests, improving the 3D fracture network characterization and mechanism resolution.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Quantification of 3D Fracture Network Characterization and the Coupling Mechanisms in Non-blasting SC-CO2 Fracturing: A Bayesian Calibrated Multiscale Complex Fracture Index

  • 期刊论文

  • 发表日期：

    2026年4月9日

查看全文
  • Chao Zhu,
  • Jie Chen,
  • Zhonghao Luo,
  • Kunyong Guo,
  • Wenhao Shi,
  • Yuanyuan Pu,
  • Yichao Rui,
  • Deping Guo

摘要

Quantitatively separating the coupled expansion–seepage contribution within non-blasting SC-CO2 fracturing mechanisms is essential for advancing CCUS-enabled rock-breaking technology. In this paper, SC-CO2 fracturing tests were conducted on granite under different injection rates (5, 10, 15, 20, and 25 ml/min), drilling depths (5, 10, and 15 mm), and fluids (SC-CO2 vs. H2O) conditions. Then, the multiple descriptors, including the difference of 3D and 2D fractal dimension (ΔD), skeleton connectivity ratio (Rm), surface roughness (Sq), directional anisotropy (Iani), and peak break pressure (PBP), were used to quantitatively characterize the reconstructed 3D fracture network and loading response. Moreover, a multiscale complex fracture index (MCFI) was formulated to integrate ΔD and Rm as backbone measures of the fracture network and to incorporate seepage-related roughness (Sq) and expansion-related directionality (Iani) through two weighting coefficients (α, β). The coefficients were inferred using a Bayesian inversion framework coupled with a logarithmic MCFI–PBP mapping, enabling probabilistic calibration and mechanistic decoupling of seepage- and expansion-dominated contributions. The results showed the MCFI-PBP mapping with high fidelity (R2 = 0.91, RMSE = 0.56 MPa) and narrow 95% credible intervals (± 0.30–0.50 MPa), outperforming single-descriptor models. Posterior estimates yielded α = 2.49 and β = 2.60 indicate a near-balanced contribution of seepage- and expansion-related effects within the SC-CO2 fracturing granite tests, while the inferred Log-elasticities coefficient suggest that expansion-driven processes act as the primary trigger that enables subsequent seepage effect development. Within the present experimental design, an MCFI value exceeding 5.50e-3 is associated with the most developed fracture network response observed at a drilling depth of 10 mm and an injection rate of 20 ml/min. This study proposes a novel Bayesian-calibrated framework for SC-CO2 fracturing tests, improving the 3D fracture network characterization and mechanism resolution.