Research on Axial Force Prediction of Bolt in TBM Headrace Tunnel Based on ACO-SVM
-
摘要: 预测隧道锚杆轴力变化趋势，不仅能掌握隧道结构的安全状况，而且对隧道风险预警和应急响应至关重要。基于新疆YEGS输水工程喀双隧洞锚杆轴力监测数据，通过蚁群算法（ACO）和粒子群算法（PSO）优化支持向量机（SVM）模型，预测分析锚杆轴力的变化趋势，研究表明：相比PSO-SVM和传统的SVM预测模型，ACO-SVM预测模型在充分考虑隧道埋深、温度及作用时间等多项非线性影响因素后，求解的预测值与实测值更加接近，相对误差基本在15%以内，平均绝对百分误差仅为5.92，具有较好的鲁棒性，模型的稳定性和泛化能力更强，更加适合TBM隧道锚杆轴力变化趋势的预测分析，具有一定的工程应用和推广价值。Abstract: Based on the monitoring data of anchor axial force in Ka-Shuang tunnel of Xinjiang YEGS water conveyance project, the change trend of anchor axial force were forecasted and analyzed through ant colony algorithm (ACO) and particle swarm optimization (PSO) to optimize the support vector machine (SVM) model. The research shows that ACO-SVM prediction model fully considers the tunnel buried depth compared with PSO-SVM and traditional SVM prediction model. After a number of nonlinear influencing factors such as temperature and action time, the predicted value is closer to the measured value, the relative error is basically within 15%, and the average absolute percentage error is only 5.92. The model has better robustness, stability and generalization ability, and is more suitable for the prediction and analysis of the variation trend of bolt axial force in TBM tunnel. It has certain engineering application and popularization value.
-
表 1 锚杆轴力实测数据汇总
样本号 时间
（月-日 时）温度/℃ 埋深/m 历时/h 轴力/kN 拱顶 拱肩 拱腰 1 05-28 0:00 23.2 432 4.4 0.06 0.03 0.02 2 05-28 3:00 23.1 432 7.4 0.10 0.05 0.04 3 05-28 6:00 23.0 432 10.4 0.14 0.07 0.05 4 05-28 9:00 22.9 432 13.4 0.24 0.10 0.08 5 05-28 12:00 22.7 432 16.4 0.34 0.12 0.11 6 05-28 15:00 22.6 432 19.4 0.44 0.16 0.14 7 05-28 18:00 22.4 432 22.4 0.53 0.20 0.16 8 05-28 21:00 22.5 432 25.4 0.60 0.27 0.18 9 05-29 0:00 22.5 432 28.4 0.67 0.34 0.19 10 05-29 3:00 23.2 432 31.4 0.70 0.40 0.22 … … … … … … … … 291 07-03 6:00 25.4 432 874.4 2.70 2.06 1.52 292 07-03 9:00 25.1 432 877.4 2.74 2.04 1.52 293 07-03 12:00 24.7 432 880.4 2.78 2.01 1.51 294 07-03 15:00 24.8 432 883.4 2.79 2.00 1.58 295 07-03 18:00 24.8 432 886.4 2.79 1.99 1.65 296 07-03 21:00 25 432 889.4 2.76 2.06 1.59 297 07-04 0:00 25.1 432 892.4 2.73 2.12 1.52 298 07-04 3:00 24.9 432 895.4 2.73 2.11 1.55 299 07-04 6:00 24.6 432 898.4 2.73 2.10 1.58 300 07-04 9:00 12.3 432 901.4 2.71 2.12 1.61 
下载: 导出CSV
表 2 预测结果分析汇总表
预测模型 相对误差最大值/% MAPE MAPE
平均值样本号
11—30样本号
281—300样本号
11—30样本号
281—300SVM 39.72 25.70 18.35 15.92 17.14 ACO-SVM 14.58 10.58 6.12 5.71 5.92 PSO-SVM 29.86 16.94 12.41 10.71 11.56
下载: 导出CSV
-
[1] 王利明,张 兵,李凤远,等. TBM隧洞倾斜黏结锚杆受力特性分析及支护参数优化[J]. 河南科学,2021,39(1):84-90. [2] 黄明利,徐 飞,伍志勇. 城市环境下TBM施工对围岩稳定性影响的监测分析及支护参数优化[J]. 岩石力学与工程学报,2012,31(7):1325-1333. [3] SU Z M,CHEN J X,LUO Y B. Mechanical characteristic and length optimization of system anchor in loess tunnel based on field measurement and analytical solution[J]. Mathematical Problems in Engineering,2021:1-11. [4] 谢清忠,刘东东,傅鹤林,等. 公路隧道初期支护参数优化下安全储备分析[J]. 公路与汽运,2021,(3):143-147,155. [5] 陶永虎,饶军应,熊 鹏,等. 软岩隧道大变形预测模型及支护措施[J]. 矿业研究与开发,2021,41(5):59-66. [6] 刘亚鑫,邢明录,刘鹏程,等. 螺纹钢锚杆锚固岩石拉拔试验过程精细化数值模拟研究[J]. 煤矿安全,2022,53(3):66-74. [7] OLALUSI O B,SPYRIDIS P. Machine learning-based models for the concrete breakout capacity prediction of single anchors in shear[J]. Advances in Engineering Software,2020,147:102832. [8] 赵向波,王利明,王 军,等. 基于现场试验的TBM隧道支护结构力学性能研究[J]. 河南科学,2022,40(1):54-58. [9] KARMEN F B,BONUT P. Displacement analysis of tunnel support in soft rock around a shallow highway tunnel at Golovec[J]. Engineering Geology,2004,75(1):89-106. doi: 10.1016/j.enggeo.2004.05.003 [10] 李 慧. 高地应力软岩隧道锚杆支护技术及参数优化研究[D]. 西安: 西安科技大学, 2020. [11] VAPNIK V N. The Nature of Statistical Learning Theory [M]. NewYork: Spring — Verlag, 1995. [12] 欧阳玲. 基于遥感和 SVM 模型的松嫩平原南部耕地质量评价[D]. 长春: 中国科学院东北地理与农业生态研究所, 2017. [13] 孙 炀. 软岩隧道挤压大变形的SVM预测及其支护对策研究[D]. 济南: 济南大学, 2018. [14] 张颖芳,凌卫新. 基于动态调整的 GA - SVM 多分类二叉树的方法[J]. 科学技术与工程,2017,17(7):177-182. [15] DORIGO M. Optimization, Learning and Natural Algorithms[D]. Politecnico di Milano, Italy, 1992. -
图(8) / 表(2)
计量
- 文章访问数: 110
- HTML全文浏览量: 41
- PDF下载量: 17
- 被引次数: 0
