School of Mechanics and Civil Engineering, China University of Mining & Technology(Beijing)

针对深埋复杂地质条件下煤巷支护设计难题,基于巷道支护、计算机开发、数值模拟二次开发等开展了交叉融合技术的研究。通过文献收集、问卷调查及现场调研的方法,建立了煤巷支护知识库,通过库内数据查找和规则推理机制为用户提供岩石力学参数和支护方案参数;构建能满足多种地质条件、多种巷道类型、多种支护方案的数值模拟脚本,基于FLAC3D内置语言开展二次开发技术的研究,建立了煤巷支护效果预测模型;运用计算机高级编程语言开发煤巷支护效果预测系统,实现了煤巷支护效果预测的可视化、智能化;利用该系统对经典案例进行测试,并用决定系数R2和平均绝对百分比误差MAPE对该系统预测效果进行了评价,结果显示其各指标的R2均在0.8以上,MAPE值均在20%以下,都在合理范围之内,该系统在巷道位移预测方面效果良好;系统在信湖煤矿818外风巷的工程实例的验证结果显示,其预测精度达到84%以上,能够为现场施工提供准确指导。

Addressing the challenge of designing coal roadway support under deeply buried and complex geologi-cal conditions, this study integrates roadway support, computer development, and numerical simulation into a co-hesive cross-fusion technological research approach. By compiling literature, conducting questionnaires, and un-dertaking field investigations, a comprehensive coal roadway support knowledge base was established. This base aids in providing users with essential rock mechanics and support scheme parameters through data retrieval and a rule-based reasoning mechanism. A versatile numerical simulation script, adaptable to varied geological condi-tions, diverse roadway types, and multiple support schemes, was developed utilizing FLAC3D's built-in language for secondary development. This facilitated the creation of a predictive model for assessing the efficacy of coal roadway support systems. Advanced computer programming was employed to develop a predictive system for coal roadway support outcomes, enabling the visualization and intelligent forecasting of support effects. The sys-tem's efficacy was demonstrated through application to a classic case, with its predictive accuracy evaluated using the coefficient of determination (R2) and the Mean Absolute Percentage Error (MAPE), both yielding satisfactory results within acceptable ranges-R2 values above 0.8 and MAPE values below 20%. Further validation in an engi-neering application, the 818 external air roadway at Xinhu Coal Mine, confirmed the system's high predictive ac-curacy of over 84%, proving its capability to provide accurate guidance for onsite construction operations.

国家自然科学基金资助项目(52074301)

马鑫民,向俊杰,翟中华,等.煤巷支护效果智能预测系统及应用[J].采矿与岩层控制工程学报,2024,6(2):023538.

MA Xinmin, XIANG Junjie, ZHAI Zhonghua, et al. Intelligent prediction system and application of coal roadway support effect[J]. Journal of Mining and Strata Control Engineering, 2024, 6(2): 023538.