基于非开挖泥浆性能检测系统与弱监督学习的地层岩性识别

徐晗; 程丹仪; 徐永华; 姚孔轩; 邱峰; 乌效鸣; 林朋皓

doi:10.19509/j.cnki.dzkq.2021.0629

基于非开挖泥浆性能检测系统与弱监督学习的地层岩性识别

doi: 10.19509/j.cnki.dzkq.2021.0629

基金项目:

国家自然科学基金项目 41731284

详细信息

作者简介:
徐晗(1994-), 男, 正在攻读地质工程专业博士学位, 主要从事岩土钻掘与工程浆液方向的研究。E-mail: xuhancug@163.com

通讯作者:
乌效鸣(1956-), 男, 教授, 博士生导师, 主要从事科学钻探与非开挖方面的教学与研究工作。E-mail: xmwu5610@163.com

中图分类号: P588
计量
- 文章访问数: 617
- PDF下载量: 214
- 被引次数: 0
出版历程
- 收稿日期: 2021-02-05

Stratigraphic lithology identification based on no-dig mud property detection system and weakly-supervised learning

摘要

摘要: 针对非开挖工程工勘资料缺乏，掘进地层岩性难以判别的问题，提出一种基于非开挖泥浆性能检测与弱监督机器学习结合的典型非开挖地层岩性识别方法。结合自主设计研发的非开挖泥浆性能检测系统工程现场应用，获取非开挖掘进导向段泥浆流变性能参数和密度等敏感参数的训练样本。利用部分有标签数据与K近邻（K Nearest Neighbors，KNN）算法对所有泥浆参数训练样本进行特征标签，采用核函数映射到高维空间支持向量机（Support Vector Machines，SVM）进行分类处理，建立了以上海地区典型非开挖地层分类为目标的分类模型。将该地层识别模型应用于上海地区非开挖工程，验证其有效性。结果表明，该方法能在非开挖实时钻进条件下快速识别钻进地层，识别正确率高达96%。研究成果通过采集导向段泥浆性能参数，识别非开挖掘进段地层岩性，为非开挖扩孔阶段钻具选型、泥浆设计等提供了重要地质信息保障。
- 非开挖 /
- 泥浆性能检测系统 /
- 支持向量机 /
- 地层岩性识别
Abstract: In view of the lack of geological investigation information data in no-dig and the difficulty in distinguishing the lithology of tunneling stratum, a typical no-dig formation lithology identification method based on support vector machines(SVM)algorithm of no-dig mud property data is proposed.Combined with the field application of the self-designed no-dig mud property detection system, the training samples of rheological parameters, density and other sensitive mud parameters were obtained.The obtained mud parameters training samples were learned by SVM algorithm, and the mud parameters sample space was constructed.The kernel function was used to map to the high-dimensional space for classification, a classification model is established for the classification of typical no-dig strata in Shanghai.The model is applied to the no-dig engineering in Shanghai to verify its effectiveness.The results show that the method can quickly identify the drilling stratigraphic lithology under the condition of no-dig real-time drilling, and the recognition accuracy is as high as 96%.The research provide important geological information for drilling tool selection and mud design in no-dig reaming stage by collecting the mud property parameters of the guide section and identifying the formation lithology of the no-dig advance section.
- no-dig /
- mud property detection system /
- support vector machines /
- strata lithology recognition

HTML全文

图 1 泥浆性能检测系统示意图及框架图

Figure 1. Schematic diagram and frame diagram of no-dig mud property detection system

下载: 全尺寸图片幻灯片

图 2 泥浆流变性能检测模块

Figure 2. Mud rheological property testing module

下载: 全尺寸图片幻灯片

图 3 KNN-SVM弱监督机器学习地层识别模型框架

C.惩罚系数；ε.RBF函数作kernel后自带参数

Figure 3. KNN-SVM weak supervised machine learning model framework for stratum recognition

下载: 全尺寸图片幻灯片

图 4 KNN标签传递示意图

绿色圆为K=5范围；红色圆为K=35范围

Figure 4. Schematic diagram of KNN label transfer

下载: 全尺寸图片幻灯片

图 5 各K值标签传递模型准确率示意图

Figure 5. Accuracy of label transfer model at each K value

下载: 全尺寸图片幻灯片

图 6 4类基于SVM地层分类器

Figure 6. Four classes of stratum classifiers based on SVM

下载: 全尺寸图片幻灯片

图 7 砾石地层分类器SVM原理示意图

d为支持向量到超平面距离

Figure 7. Schematic diagram of SVM principle of gravel stratum classifier

下载: 全尺寸图片幻灯片

图 8 未使用核函数处理的杂填土地层分类器

Figure 8. Classification of miscellaneous soil layer without kernel function processing

下载: 全尺寸图片幻灯片

表 1 导向段标准比对泥浆性能参数

Table 1. Property of standard comparison slurry in guide section

表观黏度AV/ (mPa·s)	塑性黏度PV/ (mPa·s)	动切力 YP/Pa	滤失量 FL_API/ mL	润滑系数	比重/ (g·cm^-3)
15.0	14.0	1.0	4.0	18	1.04

下载: 导出CSV

表 2 返排泥浆性能数据举例

Table 2. Example of property data of flowback slurry

分类	密度ρ_B/ (g·cm^-3)	表观黏度AV/ (mPa·s)	塑性黏度PV/ (mPa·s)
杂填土	1.21	17.10	14.80
	1.24	16.80	14.70
	1.28	16.20	15.10
	1.21	17.00	14.60
	1.23	16.70	14.30
黏土	1.13	20.80	12.70
	1.09	18.40	11.60
	1.12	20.10	11.90
	1.11	19.60	11.80
	1.10	18.80	11.50
粉细砂	1.19	14.50	14.30
	1.21	13.80	13.50
	1.25	13.40	13.20
	1.23	14.20	13.80
	1.18	14.70	14.50
砾石	1.27	6.80	6.80
	1.32	6.20	6.30
	1.29	6.30	6.20
	1.24	7.10	6.90
	1.26	6.90	6.80
淤泥	1.06	23.70	16.20
	1.05	24.10	16.40
	1.07	26.00	16.80
	1.04	21.80	15.70
	1.05	25.30	16.50
未知地层标签数据	1.22	16.90	14.40
	1.05	25.00	16.30
	1.30	6.30	6.20
	1.20	13.70	13.50
	1.11	19.80	11.80

下载: 导出CSV

表 3 部分模型精确度验证数据

Table 3. Examples of model accuracy validation data

序号	密度 ρ_B/(g·cm^-3)	表观黏度 AV/(mPa·s)	塑性黏度 PV/(mPa·s)	岩性识别	工勘岩性	识别结果
1	1.09	18.2	10.9	黏土	粉质黏土	正确
2	1.29	6.8	7.2	砾石	弱胶结砾岩	正确
3	1.06	24.6	16.8	淤泥	淤泥	正确
4	1.22	16.1	14.9	粉细砂	杂填土	错误
5	1.23	14.1	13.8	粉细砂	粉砂	正确
6	1.13	20.4	13.0	黏土	粉质黏土	正确
7	1.32	5.9	6.1	砾石	中胶结砾岩	正确
8	1.04	22.6	15.3	淤泥	淤泥	正确
9	1.22	16.8	14.5	杂填土	杂填土	正确
10	1.19	14.1	14.2	粉细砂	粉细砂	正确

下载: 导出CSV

参考文献(20)

[1]	颜纯文. 我国非开挖行业现状与展望[J]. 探矿工程: 岩土钻掘工程, 2010, 37(10): 56-60. https://www.cnki.com.cn/Article/CJFDTOTAL-TKGC201010021.htm Yan C W. Current status of trenchless industry in China and the prospect[J]. Exploration Engineering(Rock & Soil Drilling and Tunneling), 2010, 37(10): 56-60(in Chinese with English abstract). https://www.cnki.com.cn/Article/CJFDTOTAL-TKGC201010021.htm
[2]	江文, 蒋宏业, 姚安林. 水平定向钻穿越施工的失效可能性研究[J]. 中国安全生产科学技术, 2015, 11(5): 111-116. https://www.cnki.com.cn/Article/CJFDTOTAL-LDBK201505021.htm Jiang W, Jiang H Y, Yao A L. Study on failure possibility of crossing construction by horizontal directional drilling[J]. Journal of Safety Science and Technology, 2015, 11(5): 111-116(in Chinese with English abstract). https://www.cnki.com.cn/Article/CJFDTOTAL-LDBK201505021.htm
[3]	谷宇峰, 张道勇, 鲍志东, 等. 利用GS-LightGBM机器学习模型识别致密砂岩地层岩性[J]. 地质科技通报, 2021, 40(4): 224-234. doi: 10.19509/j.cnki.dzkq.2021.0416 Gu Y F, Zhang D Y, Bao Z D, et al. Lithology prediction of tight sandstone formation using GS-LightGBM hybrid machine learning model[J]. Bulletin of Geological Science and Technology, 2021, 40(4): 224-234(in Chinese with English abstract). doi: 10.19509/j.cnki.dzkq.2021.0416
[4]	张幼振, 张宁, 邵俊杰, 等. 基于钻进参数聚类的含煤地层岩性模糊识别[J]. 煤炭学报, 2019, 44(8): 2328-2335. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201908007.htm Zhang Y Z, Zhang N, Shao J J, et al. Fuzzy identification of coal-bearing strata lithology based on drilling parameter clustering[J]. Journal of Coal Industry, 2019, 44(8): 2328-2335(in Chinese with English abstract). https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201908007.htm
[5]	翟淑花, 高谦. 基于进化算法的岩体参数智能识别[J]. 煤炭学报, 2011, 36(1): 34-38. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201101009.htm Zhai S H, Gao Q. Intelligent identification of rock mass parameters based on evolutionary algorithm[J]. Journal of Coal Society, 2011, 36(1): 34-38(in Chinese with English abstract). https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201101009.htm
[6]	Sair K, Jamai R, Ali N. Review of ground characterization by using instrumented drills for underground mining and construction[J]. Rock Mechanics and Rock Engineering, 2016, 49(2): 585-602. doi: 10.1007/s00603-015-0756-4
[7]	刘彦锋, 张文彪, 段太忠, 等. 深度学习油气藏地质建模研究进展[J]. 地质科技通报, 2021, 40(4): 235-241. doi: 10.19509/j.cnki.dzkq.2021.0417 Liu Y F, Zhang W B, Duan T Z, et al. Progress of deep learning in oil and gas reservoir geological modeling[J]. Bulletin of Geological Science and Technology, 2021, 40(4): 235-241(in Chinese with English abstract). doi: 10.19509/j.cnki.dzkq.2021.0417
[8]	房昱纬, 吴振君, 盛谦, 等. 基于超前钻探测试的隧道地层智能识别方法[J]. 岩土力学, 2020, 41(7): 2494-2503. https://www.cnki.com.cn/Article/CJFDTOTAL-YTLX202007036.htm Fang Y W, Wu Z J, Sheng Q, et al. Intelligent recognition of tunnel stratum based on advanced drilling tests[J]. Geotechnical Mechanics, 2020, 41(7): 2494-2503(in Chinese with English abstract). https://www.cnki.com.cn/Article/CJFDTOTAL-YTLX202007036.htm
[9]	乌效鸣, 胡郁乐, 贺冰新, 等. 钻井液与岩土工程浆液[M]. 武汉: 中国地质大学出版社, 2002. Wu X M, Hu Y L, He B X, et al. Drilling fluid and geotechnical slurry[M]. Wuhan: China University of Geosciences Press, 2002(in Chinese).
[10]	鄢捷年, 李健鹰. 钻井液工艺学[M]. 北京: 中国石油大学出版社, 2016. Yan J N, Li J Y. Drilling fluid technology[M]. Beijing: China University of Petroleum Press, 2016(in Chinese).
[11]	刘钊, 杜威, 闫冬梅, 等. 基于K近邻算法和支持向量回归组合的短时交通流预测[J]. 公路交通科技, 2017, 34(5): 122-128, 158. https://www.cnki.com.cn/Article/CJFDTOTAL-GLJK201705017.htm Liu Z, Du W, Yan D M, et al. Short-term traffic flow forecast based on combination of K-nearest neighbor algorithm and support vector regression[J]. Journal of Highway and Transportation Research and Development, 2017, 34(5): 122-128, 158(in Chinese with English abstract). https://www.cnki.com.cn/Article/CJFDTOTAL-GLJK201705017.htm
[12]	Feng K, González A, Casero M. A KNN algorithm for locating and quantifying stiffness loss in a bridge from the forced vibration due to a truck crossing at low speed[J]. Mechanical Systems and Signal Processing, 2021, DOI: 10.1016/J.YMSSP.2020.107599.
[13]	Tsalera E, Papadakis A, Samarakou M. Monitoring, profiling and classification of urban environmental noise using sound characteristics and the KNN algorithm[J]. Energy Reports, 2020, 6(S6): 223-230. http://www.sciencedirect.com/science/article/pii/S2352484720313007
[14]	Giulio V, Marco N. Enhancement of a short-term forecasting method based on clustering and kNN: Application to an industrial facility powered by a cogenerator[J]. Energies, 2019, 12(23): 4407-4423. http://www.researchgate.net/publication/337407243_Enhancement_of_a_Short-Term_Forecasting_Method_Based_on_Clustering_and_kNN_Application_to_an_Industrial_Facility_Powered_by_a_Cogenerator
[15]	Vapnik V N. The nature of statistical learning theory[M]. New York: Springer Verlag, 2000.
[16]	Ahmadi M A, Chen Z. Comparison of machine learning methods for estimating permeability and porosity of oil reservoirs via petro-physical logs[J]. Petroleum, 5(3): 271-284. http://qikan.cqvip.com/Qikan/Article/Detail?id=7103867149
[17]	周志华. 机器学习[M]. 北京: 清华大学出版社, 2016. Zhou Z H. Machine learning[M]. Beijing: Tsinghua University Press, 2016(in Chinese).
[18]	Chen H, Zhang C, Jia N H, et al. A machine learning model for predicting the minimum miscibility pressure of CO₂ and crude oil system based on a support vector machine algorithm approach[J]. Fuel, 2021, 290(1): 120048. http://www.sciencedirect.com/science/article/pii/S0016236120330453
[19]	王万良. 人工智能导论[M]. 北京: 高等教育出版社, 2017. Wang W L. Introduction to artificial intelligence[M]. Beijing: Higher Education Press, 2017(in Chinese).
[20]	范昕炜. 支持向量机算法的研究及其应用[D]. 杭州: 浙江大学, 2003. Fan X W. Supportvector machine and its applications[D]. Hangzhou: Zhejiang University, 2003(in Chinese with English abstract).