TL;DR
This paper investigates how high temperatures and pressures affect rock strength, a critical factor for the safety of deep underground infrastructure like geothermal systems. By evaluating six common rock strength models, the study finds that the Hoek-Brown (HB) criterion is significantly more accurate than others in predicting rock behavior under these extreme conditions, leading to improved wellbore stability analysis.
What Problem Does This Paper Address?
Geothermal energy development and other deep underground projects (like waste repositories) operate in environments with significantly elevated temperatures and pressures. These conditions drastically alter the mechanical behavior of rock formations, making standard rock strength models, developed for ambient conditions, unreliable. This unreliability poses a major risk to the structural integrity and operational safety of subsurface infrastructure. The paper aims to address this by evaluating and identifying the most accurate rock strength criteria for these high-temperature, high-pressure scenarios.
Methods
The researchers employed a systematic evaluation of six commonly used rock strength criteria: Mohr-Coulomb (MC), Mogi-Coulomb (MGC), modified Lade (ML), Hoek-Brown (HB), modified Wiebols–Cook (MWC), and Zhang-Zhu (ZZ). They used the least absolute deviation (LAD) method to fit experimental rock strength data across various temperatures and confining pressures. This method is preferred over the least squares method as it is less sensitive to outliers, providing a more robust fit to the majority of the data. The study then compared the predictive accuracy of these criteria based on their fitting errors. Finally, the most promising criterion (HB) was applied to a real-world case study of wellbore stability in a challenging geothermal drilling operation.
Datasets
The study utilized experimental data from six independent sources, encompassing five different lithologies: granite (Strathbogie and Tak), shale (Tournemire and Changning), slate, and sandstone (Crab Orchard). These datasets covered a range of temperatures from ambient (25°C) up to 500°C and various confining pressures. The data was presented in appendices, detailing rock strength values under different temperature and pressure conditions for each lithology.
Key Insights and Ideas
- Temperature is a major strength degrader: Rock strength significantly decreases (18-32% at 500°C) under high temperatures, a phenomenon driven by thermally activated microcrack propagation and mineralogical changes.
- Hoek-Brown reigns supreme: The Hoek-Brown (HB) criterion consistently outperformed other models, showing 12-18% better accuracy in predicting temperature-dependent rock strength. This is attributed to its ability to capture both nonlinear confinement effects and thermal damage mechanisms.
- Linear models fall short: Simpler, linear models like Mohr-Coulomb (MC) and Zhao-Zheng (ZZ) showed significant deviations, highlighting their inadequacy for complex thermo-mechanical conditions.
- Lithology matters: Different rock types (granite, shale, slate, etc.) exhibit unique thermal responses, meaning a one-size-fits-all approach to strength modeling is insufficient. Some rocks show counterintuitive strength behaviors (e.g., initial strengthening before weakening).
- Real-world validation: The HB criterion's successful application to a challenging wellbore stability problem in the Karamay Oilfield demonstrated its practical utility, achieving 89% alignment with field observations.
Key Results
- The Hoek-Brown (HB) criterion demonstrated superior predictive accuracy, outperforming MGC, ML, and MWC by 12-18%.
- MC and ZZ criteria showed statistically significant deviations, indicating their limitations.
- In the field application at Well X, the HB criterion-based analysis yielded 89% alignment with observed wellbore stability and kept principal stress variations within 7% of critical failure thresholds.
- Rock strength reductions of 18-32% were observed across lithologies at 500°C compared to ambient conditions.
Limitations and Future Work
The authors acknowledge that laboratory simulations of in-situ thermal conditions are challenging and scarce, leading to a reliance on existing experimental data. They also note that while the HB criterion is excellent for tensile regimes, other models like ML/MWC might offer computational advantages in such scenarios, albeit with slightly reduced accuracy.
Future work should focus on:
- Investigating cyclic thermal loading effects to better understand reversible thermoelastic responses versus irreversible thermochemical alterations.
- Integrating multi-scale damage mechanisms to enhance predictive capabilities for deep energy extraction and storage systems.
- Developing lithology-specific thermo-mechanical coupling frameworks to account for the diverse responses of different rock types.
Summary updated