基于现代原位测试CPTU的土体液化势统一评价方法

段伟; 蔡国军; 刘松玉; 赵泽宁; 董晓强; 陈瑞锋

doi:10.11779/CJGE202203005

基于现代原位测试CPTU的土体液化势统一评价方法 English Version

段伟^{1, 2,},
蔡国军^{2, 3, ,},
刘松玉²,
赵泽宁²,
董晓强¹,
陈瑞锋²

1.
太原理工大学土木工程学院, 山西太原 030024
2.
东南大学岩土工程研究所, 江苏南京 211189
3.
安徽建筑大学土木工程学院, 安徽合肥 230601

基金项目:

国家自然科学基金项目 52108332

国家自然科学基金项目 41877231

国家自然科学基金项目 42072299

中国博士后科学基金项目 2021M702421

详细信息

作者简介:
段伟(1989—)，男，山西太原人，博士，博士后，从事现代原位测试技术等方面研究。E-mail: zbdxdw@163.com

通讯作者:
蔡国军, E-mail: focuscai@163.com

中图分类号: TU413
计量
- 文章访问数: 094
- HTML全文浏览量: 0
- PDF下载量: 0132
出版历程
- 收稿日期: 2021-03-25
- 网络出版日期: 2022-09-22
- 刊出日期: 2022-02-28

Unified evaluation method for soil liquefaction potential based on modern in-situ piezocone penetration tests

1.
College of Civil Engineering, Taiyuan University of Technology, Taiyuan 030024, China
2.
Institute of Geotechnical Engineering, Southeast University, Nanjing 211189, China
3.
School of Civil Engineering, Anhui Jianzhu University, Hefei 230601, China

摘要

摘要: 静力触探(CPT)测试是国际上常用的砂质土液化判别技术，但无法反映孔隙水压力(孔压)对液化行为的影响。能够测试孔压的现代原位测试技术孔压静力触探(CPTU)对砂质土、粉质土具有更高的辨识度和灵敏度，因此在砂质土、粉质土液化评价方面具有独特的优势。通过编译的CPTU液化案例数据库，在简化应力框架下，通过粒子群优化核极限学习机算法、稳健搜索技术和非线性拟合优化分析建立了基于CPTU测试参数的液化阻力比(CRR_7.5)模型，该模型包括修正的锥尖阻力(q_t1N)和土类指数(I_{c, BJ})；其中I_{c, BJ}采用Jefferies和Davies提出的形式，包含了孔压参数比(B_q)。这样所提模型中土壤类型信息直接包含在液化判别公式中，无需根据土壤细粒含量进行贯入阻力的修正，更直接、更符合力学框架，适用于砂质土、粉质土等较广范围的土体；并通过内含B_q的土类指数I_{c, BJ}将砂质土和粉质土液化判别统一起来。最后用中国唐山地震液化案例证明所提模型的准确性和优越性。
- 孔压静力触探 /
- 粉质土 /
- 液化 /
- 土类指数 /
- 孔压参数比
Abstract: The cone penetration test (CPT) is a commonly used technique to identify the liquefaction of sandy soil in the world but it cannot reflect the influences of pore water pressure (pore pressure) on the liquefaction behavior. The piezocone penetration test (CPTU), a modern in-situ test technique, can measure pore pressure, and has high identification and sensitivity to sandy soil and silty soil, so it has a unique advantage in liquefaction evaluation of sandy soil and silty soil. In this study, a liquefaction resistance ratio (CRR_7.5) model based on CPTU test parameters is established under the simplified stress framework through the particle swarm optimization (PSO)-kernel limit learning machine (KLEM) algorithm, robust search technology and nonlinear fitting optimization analysis. The model includes modified cone resistance (q_t1N) and soil behavior type index (I_{c, BJ}) as per Jefferies and Davies, including the pore pressure ratio (B_q). In this way, the soil type information in the proposed model is directly included in the liquefaction discrimination formula, and it does not need to modify the penetration resistance according to the content of soil fine content, which is more direct and more consistent with the mechanical framework, and is suitable for a wide range of soil, such as sandy soil and silty soil. The liquefaction discrimination of sandy soil and silty soil is unified by the I_{c, BJ} with the inclusion of B_q in the formulation. Finally, the accuracy and superiority of the proposed model are proved by a case study of Tangshan Earthquake liquefaction in China.
- piezocone penetration test /
- silty soil /
- liquefaction /
- soil behavior type index /
- pore pressure ratio

HTML全文

图 1 PSO-KELM算法流程图

Figure 1. Flow chart of PSO-KELM algorithm

下载: 全尺寸图片幻灯片

图 2 PSO-KELM模型迭代寻优过程

Figure 2. Iterative optimization process of PSO-KELM model

下载: 全尺寸图片幻灯片

图 3 PSO-KELM模型液化判别结果

Figure 3. Results of liquefaction discrimination of PSO-KELM model

下载: 全尺寸图片幻灯片

图 4 参数敏感性分析

Figure 4. Sensitivity analysis

下载: 全尺寸图片幻灯片

图 5 极限状态边界搜索点模型示意图

Figure 5. Schematic diagram of model for searching points on limit state boundary

下载: 全尺寸图片幻灯片

图 6 所提模型的CRR_7.5预测散点图

Figure 6. Scattering of CRR_7.5 predictions by proposed model

下载: 全尺寸图片幻灯片

图 7 所提模型的二维图：CRR_7.5(CSR_7.5)与q_t1N

Figure 7. 2-D graphs of proposed model: CRR_7.5 (or CSR_7.5) versus q_t1N

下载: 全尺寸图片幻灯片

图 8 忽略B_q采用I_{c, RW}对所提CRR_7.5模型的影响

Figure 8. Percentage change in CRR_7.5 as a result of ignoring B_q in proposed model

下载: 全尺寸图片幻灯片

图 9 唐山场地T1的CPTU剖面

Figure 9. CPTU sounding profiles at location T1 in Tangshan sites

下载: 全尺寸图片幻灯片

图 10 场地T1的液化结果分析

Figure 10. Liquefaction results of analysis of Tangshan T1

下载: 全尺寸图片幻灯片

表 1 混淆矩阵

Table 1 Confusion matrix

模型预测	实际
模型预测	液化	非液化
液化	TP	FP
非液化	FN	TN

下载: 导出CSV

表 2 历史案例各个参数值的范围

Table 2 Ranges of parameter values in historical case

指标	q_t/MPa	f_s/kPa	u₂/kPa	B_q	σ_v/kPa	σ′_v/kPa	q_t1N	I_{c, BJ}	I_{c, RW}	a_max/g	M_w	CSR_7.5
最小	0.2	0.2	-104.1	-0.26	26.0	18.2	2.5	0.59	0.98	0.09	5.9	0.04
最大	28.9	358.6	1195.8	0.82	314.5	159.5	304.7	3.39	3.44	0.79	7.8	0.85
注：q_t，f_s，u₂分别为CPTU锥尖阻力、侧壁摩阻力、孔压水压力；σ_v，σ′_v分别为总竖向应力、有效竖向应力；a_max为地震加速度；M_w为震级。

下载: 导出CSV

表 3 敏感性分析结果

Table 3 Results of sensitivity analysis

模型	训练集OA	测试集OA
LI = f(q_t1N, I_{c, BJ}, ${\sigma '_{{\text{vo}}}}$ , CSR_7.5)	97.1%	89.1%
LI = f(q_t1N, I_{c, BJ}, CSR_7.5)	88.1%	81.5%
LI = f(q_t1N, ${\sigma '_{{\text{vo}}}}$ , CSR_7.5)	87.7%	80.6%
LI = f(q_t1N, I_{c, BJ}, ${\sigma '_{{\text{vo}}}}$ )	78.3%	73.5%

下载: 导出CSV

表 4 不同模型液化判别训练集结果对比

Table 4 Comparison among different machine learning models in training dataset

模型	训练集				测试集
模型	OA	P	R	F	OA	P	R	F
PSO-KELM	97.1%	97.2%	98.9%	0.980	89.1%	91.7%	94.3%	0.930
ELM	93.9%	92.6%	99.4%	0.959	87.3%	87.5%	91.4%	0.894
LS-SVM	96.1%	95.6%	99.4%	0.975	85.5%	86.9%	91.2%	0.890
LM-BP	89.4%	88.5%	96.1%	0.922	83.6%	86.5%	90.9%	0.886

下载: 导出CSV

表 5 不同水平下液化案例正确率

Table 5 Accuracy rates of liquefaction cases at different levels

I_{c, BJ}范围	土类	液化数	液化正判率
I_{c, BJ} < 1.25	砾质砂	12	83.3%
1.25 < I_{c, BJ} < 1.80	砂	101	87.1%
1.80 < I_{c, BJ} < 2.40	砂质混合物：纯净砂-粉砂	62	93.5%
2.40 < I_{c, BJ} < 2.76	粉质土：黏质粉土-粉质黏土	23	100%
I_{c, BJ} > 2.76	黏土	31	100%

下载: 导出CSV

参考文献(22)

[1]	陈国兴, 孔梦云, 李小军, 等. 以标贯试验为依据的砂土液化确定性及概率判别法[J]. 岩土力学, 2015, 36(1): 9–27. https://www.cnki.com.cn/Article/CJFDTOTAL-YTLX201501002.htm CHEN Guo-xing, KONG Meng-yun, LI Xiao-jun, et al. Deterministic and probabilistic triggering correlations for assessment of seismic soil liquefaction at nuclear power plant[J]. Rock and Soil Mechanics, 2015, 36(1): 9–27. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YTLX201501002.htm
[2]	蔡国军, 刘松玉, 童立元, 等. 基于静力触探测试的国内外砂土液化判别方法[J]. 岩石力学与工程学报, 2008, 27(5): 1019–1027. doi: 10.3321/j.issn:1000-6915.2008.05.018 CAI Guo-jun, LIU Song-yu, TONG Li-yuan, et al. Evaluation of liquefaction of sandy soils based on cone penetration test[J]. Chinese Journal of Rock Mechanics and Engineering, 2008, 27(5): 1019–1027. (in Chinese) doi: 10.3321/j.issn:1000-6915.2008.05.018
[3]	SEED H B, IDRISS I M. Simplified procedure for evaluating soil liquefaction potential[J]. Journal of the Soil Mechanics and Foundations Division, ASCE, 1971, 97(9): 1249–1273. doi: 10.1061/JSFEAQ.0001662
[4]	JUANG C H, CHEN C H, MAYNE P W. CPTU simplified stress-based model for evaluating soil liquefaction potential[J]. Soils and Foundations, 2008, 48(6): 755–770. doi: 10.3208/sandf.48.755
[5]	ROBERTSON P K, WRIDE C E. Evaluating cyclic liquefaction potential using the cone penetration test[J]. Canadian Geotechnical Journal, 1998, 35(3): 442–459. doi: 10.1139/t98-017
[6]	MOSS R E, SEED R B, KAYEN R E, et al. CPT-based probabilistic and deterministic assessment of in situ seismic soil liquefaction potential[J]. Journal of Geotechnical and Geoenvironmental Engineering, 2006, 132(8): 1032–1051. doi: 10.1061/(ASCE)1090-0241(2006)132:8(1032)
[7]	YOUD T L, IDRISS I M. Liquefaction resistance of soils: summary report from the 1996 NCEER and 1998 NCEER/NSF workshops on evaluation of liquefaction resistance of soils[J]. Journal of Geotechnical and Geoenvironmental Engineering, 2001, 127(4): 297–313. doi: 10.1061/(ASCE)1090-0241(2001)127:4(297)
[8]	袁晓铭, 曹振中. 基于土层常规参数的液化发生概率计算公式及其可靠性研究[J]. 土木工程学报, 2014, 47(4): 99–108. https://www.cnki.com.cn/Article/CJFDTOTAL-TMGC201404015.htm YUAN Xiao-ming, CAO Zhen-zhong. Conventional soils parameters-based liquefaction probabilistic evaluation formula and its reliability analysis[J]. China Civil Engineering Journal, 2014, 47(4): 99–108. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-TMGC201404015.htm
[9]	段伟, 蔡国军, 刘松玉, 等. 多功能参数静力触探在地震液化判别方法中的应用进展研究[J]. 地震工程学报, 2020, 42(3): 764–776. doi: 10.3969/j.issn.1000-0844.2020.03.764 DUAN Wei, CAI Guo-jun, LIU Song-yu, et al. Progress of the application of the multifunctional cone penetration test in the seismic liquefaction discrimination method[J]. China Earthquake Engineering Journal, 2020, 42(3): 764–776. (in Chinese) doi: 10.3969/j.issn.1000-0844.2020.03.764
[10]	AHMAD M, TANG X W, QIU J N, et al. Application of machine learning algorithms for the evaluation of seismic soil liquefaction potential[J]. Frontiers of Structural and Civil Engineering, 2021, 15(2): 490–505. doi: 10.1007/s11709-020-0669-5
[11]	胡记磊, 唐小微, 裘江南. 基于贝叶斯网络的地震液化概率预测分析[J]. 岩土力学, 2016, 37(6): 1745–1752. https://www.cnki.com.cn/Article/CJFDTOTAL-YTLX201606027.htm HU Ji-lei, TANG Xiao-wei, QIU Jiang-nan. Prediction of probability of seismic-induced liquefaction based on Bayesian network[J]. Rock and Soil Mechanics, 2016, 37(6): 1745–1752. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YTLX201606027.htm
[12]	ZHAO Z N, DUAN W, CAI G J. A novel PSO-KELM based soil liquefaction evaluation system using CPT and V_s measurements[J]. Soil Dynamics and Earthquake Engineering, 2021. online.
[13]	BRAY J D, SANCIO R B, DURGUNOGLU T, et al. Subsurface characterization at ground failure sites in Adapazari, Turkey[J]. Journal of Geotechnical and Geoenvironmental Engineering, 2004, 130(7): 673–685. doi: 10.1061/(ASCE)1090-0241(2004)130:7(673)
[14]	HUANG G B, ZHU Q Y, SIEW C K. Extreme learning machine: theory and applications[J]. Neurocomputing, 2006, 70(1/3): 489–501.
[15]	杨锡运, 关文渊, 刘玉奇, 等. 基于粒子群优化的核极限学习机模型的风电功率区间预测方法[J]. 中国电机工程学报, 2015, 35(增刊1): 146–153. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGDC2015S1023.htm YANG Xi-yun, GUAN Wen-yuan, LIU Yu-qi, et al. Prediction intervals forecasts of wind power based on PSO-KELM[J]. Proceedings of the CSEE, 2015, 35(S1): 146–153. (In Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZGDC2015S1023.htm
[16]	HUANG G B. An insight into extreme learning machines: random neurons, random features and kernels[J]. Cognitive Computation, 2014, 6(3): 376–390. doi: 10.1007/s12559-014-9255-2
[17]	KENNEDY J, EBERHART R. Particle swarm optimization [C]// Proceedings of ICNN'95-International Conference on Neural Networks, IEEE, 1995, Perth.
[18]	JEFFERIES M G, DAVIES M P. Use of CPTU to estimate equivalent SPT N₆₀[J]. Geotechnical Testing Journal, 1993, 16(4): 458–468. doi: 10.1520/GTJ10286J
[19]	JUANG C H, CHING J, KU C S, et al. Unified CPTu-based probabilistic model for assessing probability of liquefaction of sand and clay[J]. Géotechnique, 2012, 62(10): 877–892. doi: 10.1680/geot.9.P.025
[20]	JUANG C H, CHEN C J, TANG W H, et al. CPT-based liquefaction analysis, Part 1: determination of limit state function[J]. Géotechnique, 2000, 50(5): 583–592. doi: 10.1680/geot.2000.50.5.583
[21]	CAI G J, LIU S Y, PUPPALA A J. Liquefaction assessment using seismic piezocone penetration (SCPTU) test investigations in Tangshan region in China[J]. Soil Dynamics and Earthquake Engineering, 2012, 41: 141–150. doi: 10.1016/j.soildyn.2012.05.008
[22]	KU C S, JUANG C H, CHANG C W, et al. Probabilistic version of the Robertson and Wride method for liquefaction evaluation: development and application[J]. Canadian Geotechnical Journal, 2012, 49(1): 27–44. doi: 10.1139/t11-085