煤炭工程 ›› 2025, Vol. 57 ›› Issue (9): 129-137.doi: 10. 11799/ ce202509018

• 设计技术 • 上一篇    下一篇

基于智能集成算法的煤矿TBM掘进参数多目标优化方法研究

沈远伟,朱昊,曹梦炫,冯昌如,朱梦圆,张超凡   

  1. 1. 陕西正通煤业有限责任公司,陕西 咸阳 713699

    2. 西安科技大学 能源学院,陕西 西安 710054

  • 收稿日期:2025-01-13 修回日期:2025-07-24 出版日期:2025-09-10 发布日期:2025-10-13
  • 通讯作者: 郭振桥 E-mail:guozhenqiao@longruan.com

Construction and application of transparent geological model in Shendong mining area based on drilling and geophysical data

  • Received:2025-01-13 Revised:2025-07-24 Online:2025-09-10 Published:2025-10-13
  • Contact: GUO Zhenqiao E-mail:guozhenqiao@longruan.com

摘要:

在煤矿巷道TBM掘进过程中,掘进参数对推进效率、能耗及刀具磨损的影响显著,且三者之间存在复杂的非线性耦合关系。传统的经验调参方式或单目标优化方法难以同时兼顾效率、成本与安全性。为此提出一种集成灰狼优化算法(GWO)、径向基神经网络(RBF)、非支配排序遗传算法Ⅱ(NSGA Ⅱ)与逼近理想解排序法(TOPSIS)的多层级智能优化方法,该方法融合了元启发式优化、非线性建模、多目标进化与决策排序,可实现掘进参数的全流程优化控制。首先,通过RBF神经网络构建掘进参数与推进速度、掘进比能、刀具磨损量之间的非线性映射模型,并采用GWO算法优化其超参数以提高预测精度;其次,采用NSGA Ⅱ实现多目标优化,获得Pareto最优解集;最后,通过TOPSIS-熵权法筛选出综合最优的掘进参数方案。以正通煤业TBM实测数据为案例开展验证,结果表明:在最优方案条件下,推进速度提高23.97%,掘进比能下降26.44%,刀具磨损量降低43.67%。该方法在提升掘进效率的同时,显著降低了能耗与刀具损耗,实现了效率、成本与安全性的协同优化,具备良好的工程适用性与推广价值。

关键词:

TBM , GWO , RBF神经网络 , NSGA Ⅱ , TOPSIS , 多目标优化 , 掘进参数优化 , 掘进比能

Abstract:

In order to solve the problems of complex multi-source data management, insufficient intelligent graphics rendering, low system integration and abstract results expression in coal mine underground drilling and geophysical exploration engineering, a "two-exploration" data fusion application system for transparent geological construction is constructed. A dual-engine architecture based on multi-dimensional cloud GIS platform and 3D model cloud rendering platform is proposed, and a two-3D integrated transparent geological support system including "digital-graphic-mode" triple linkage mechanism is developed. Through the establishment of a unified data center, the standardized filling and reporting of the "two-probe" data of 14 mines was realized, the geological attribute field was constructed by the improved anisotropic inverse distance weighted interpolation algorithm (AIDW), the high-precision modeling of geological bodies was realized by combining the Loop subdivision surface optimization technology, and the differential linkage model of drilling parameters and the three-dimensional picture space mapping algorithm were innovatively designed. The application shows that the system can increase the interpretation efficiency of geophysical exploration results by 40%, reduce the design time of water exploration and drainage holes by 65%, control the response time of 3D seismic data cutting analysis within 200ms, and the accuracy rate of hydrogeological anomaly identification reaches 92%. The triple linkage mechanism and spatial interpolation optimization algorithm proposed in this paper effectively solve the technical bottleneck of "two-probe" data fusion modeling in coal mines, and provide a scalable geological transparency solution for the construction of intelligent mines.

Key words: triple linkage of ", number-graph-module"

中图分类号: