Scholarlink

<p>To address CO<sub>2</sub> sequestration and solid waste disposal in coal mining, this study proposes CO<sub>2</sub> mineralization sequestration storage in mine goafs via functional backfilling, thereby achieving safe sequestration, goaf utilization, and waste treatment. Meanwhile, it analyzes the fundamental mechanical properties of these storage units under thermal–hydro-mechanical (THM) multi-field coupling effects, as well as the influence patterns of storage parameters on key physical quantities. Additionally, machine learning is introduced to construct a surrogate model suitable for the THM coupling scenarios of these storage units, enabling effective prediction of their mechanical responses. The research findings indicate the following. (1) Stress in lateral backfill columns exhibits a serrated lateral distribution, with a stress gradient of 9 × 10<sup>6</sup> N/m<sup>2</sup> at lateral-vertical column junctions. The temperature and displacement of these columns show a parabolic lateral distribution, with minima in the column midsections. (2) Both the maximum storage stress and maximum displacement exhibit the same sensitivity order to the six parameters, from highest to lowest: burial depth, Young's modulus, Poisson's ratio, thermal expansion coefficient, porosity, and thermal conductivity. (3) The surrogate prediction model has a prediction error rate within 0.4% for storage stress, while those for displacement and temperature are both within 0.1%, demonstrating its reliable predictive capability. This study enables efficient, accurate prediction of THM-coupled mechanical responses in CO<sub>2</sub> mineralization sequestration storage, thereby supporting their mechanical performance evaluation and potential failure prediction.</p><p><b>Highlights</b><UnorderedList Mark="Bullet"> <ItemContent> <p>"CO2 sequestration + backfill storage" concept and technology was proposed.</p> </ItemContent> <ItemContent> <p>THM multi-field coupled mechanical properties of backfill storage were analyzed.</p> </ItemContent> <ItemContent> <p>An approximate multi-physics prediction model was developed via machine learning.</p> </ItemContent> </UnorderedList></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

Efficient Prediction of THM Multi-field Coupled Mechanical Properties of CO2-Filled Storage Based on Machine Learning

  • 期刊论文

  • 发表日期：

    2026年4月29日

查看全文
  • Bingbing Tu,
  • Haiqiang He,
  • Lang Liu,
  • Tingyue Wei

摘要

To address CO2 sequestration and solid waste disposal in coal mining, this study proposes CO2 mineralization sequestration storage in mine goafs via functional backfilling, thereby achieving safe sequestration, goaf utilization, and waste treatment. Meanwhile, it analyzes the fundamental mechanical properties of these storage units under thermal–hydro-mechanical (THM) multi-field coupling effects, as well as the influence patterns of storage parameters on key physical quantities. Additionally, machine learning is introduced to construct a surrogate model suitable for the THM coupling scenarios of these storage units, enabling effective prediction of their mechanical responses. The research findings indicate the following. (1) Stress in lateral backfill columns exhibits a serrated lateral distribution, with a stress gradient of 9 × 106 N/m2 at lateral-vertical column junctions. The temperature and displacement of these columns show a parabolic lateral distribution, with minima in the column midsections. (2) Both the maximum storage stress and maximum displacement exhibit the same sensitivity order to the six parameters, from highest to lowest: burial depth, Young's modulus, Poisson's ratio, thermal expansion coefficient, porosity, and thermal conductivity. (3) The surrogate prediction model has a prediction error rate within 0.4% for storage stress, while those for displacement and temperature are both within 0.1%, demonstrating its reliable predictive capability. This study enables efficient, accurate prediction of THM-coupled mechanical responses in CO2 mineralization sequestration storage, thereby supporting their mechanical performance evaluation and potential failure prediction.

Highlights

"CO2 sequestration + backfill storage" concept and technology was proposed.

THM multi-field coupled mechanical properties of backfill storage were analyzed.

An approximate multi-physics prediction model was developed via machine learning.