Scholarlink

<p>Evaluation of mechanical parameters in reservoir rocks is critical for assessing unconventional reservoirs, yet standard testing methods are often hindered by the unavailability of suitable rock samples and challenges in characterizing complex micro-structures. In this study, we propose a multi-scale simulation approach for determining reservoir rock mechanical parameters. Digital core models were constructed from CT scans of four core samples from the Shengli oilfield, and mineral phases were classified using thin slice observations and machine learning techniques. Crystal lattice models were developed at the molecular level, and mechanical parameter simulations were conducted based on molecular modeling. Through finite difference methods, stress was calculated under various strain states to determine Young’s modulus and Poisson’s ratio for each mineral component. Finite element simulations of uniaxial compression tests at the core scale were performed using the mechanical parameters derived from molecular simulations. Comparison with uniaxial compression test results demonstrated the effective prediction of rock mechanical properties, saving experimental time and costs while upholding a notable level of accuracy with discrepancies less than 20%. </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

From Molecules to Core: A Multi-scale Simulation Approach for Determining Reservoir Rock Mechanical Parameters

  • 期刊论文

  • 发表日期：

    2025年10月14日

查看全文
  • Zhixue Sun,
  • Yifan Yin,
  • Ki-Bok Min,
  • Zhilei Sun,
  • Kwang Yeom Kim,
  • Yongfei Yang,
  • Xilin Zhang

摘要

Evaluation of mechanical parameters in reservoir rocks is critical for assessing unconventional reservoirs, yet standard testing methods are often hindered by the unavailability of suitable rock samples and challenges in characterizing complex micro-structures. In this study, we propose a multi-scale simulation approach for determining reservoir rock mechanical parameters. Digital core models were constructed from CT scans of four core samples from the Shengli oilfield, and mineral phases were classified using thin slice observations and machine learning techniques. Crystal lattice models were developed at the molecular level, and mechanical parameter simulations were conducted based on molecular modeling. Through finite difference methods, stress was calculated under various strain states to determine Young’s modulus and Poisson’s ratio for each mineral component. Finite element simulations of uniaxial compression tests at the core scale were performed using the mechanical parameters derived from molecular simulations. Comparison with uniaxial compression test results demonstrated the effective prediction of rock mechanical properties, saving experimental time and costs while upholding a notable level of accuracy with discrepancies less than 20%.