Abstract
Rock squeezing has a large influence on tunnel construction safety; thus, when designing and constructing tunnels it is highly important to use a reliable method for predicting tunnel squeezing from incomplete data. In this study, a combination SVM-BP (support vector machine-back-propagation) model is proposed to classify the deformation caused by surrounding rock squeezing. We design different characteristic parameters and three types of classifiers (a SVM model, a BP model, and the proposed SVM-BP model) for the tunnel-squeezing prediction experiments and analyse the accuracy of predictions by different models and the influences of characteristic parameters on the prediction results. In contrast to other prediction methods, the proposed SVM-BP model is verified to be reliable. The results show that four characteristics: tunnel diameter (D), tunnel buried depth (H), rock quality index (Q) and support stiffness (K) reflect the effect of rock squeezing sufficiently for classification. The SVM-BP model combines the advantages of both an SVM and a BP neural network. It possesses flexible nonlinear modelling ability and the ability to perform parallel processing of large-scale information. Therefore, the SVM-BP model achieves better classification performance than do the SVM or BP models separately. Moreover, coupling D, H, and K has a significant impact on the predicted results of tunnel squeezing.
Similar content being viewed by others
References
Aydan Ö, Akagi T, Kawamoto T (1996) The squeezing potential of rock around tunnels: theory and prediction with examples taken from japan. Rock Mech Rock Eng 26:137–163. https://doi.org/10.1007/BF01032650
Chang CC, Lin CJ (2011) Libsvm: a library for support vector machines. ACM Trans Intel Sys Technol 2(3):1–27. https://doi.org/10.1145/1961189.1961199
Chen Y, Li TB, Zeng P, Ma JJ, Patelli E, Edwards B (2020) Dynamic and probabilistic multi-class prediction of tunnel squeezing intensity. Rock Mech Rock Eng 53:3521–3542. https://doi.org/10.1007/s00603-020-02138-8
Debernardi D, Barla G (2009) New viscoplastic model for design analysis of tunnels in squeezing conditions. Rock Mech Rock Eng 42(2):259. https://doi.org/10.1007/s00603-009-0174-6
Dhakshina K, Esakkirajan S, Bama S, Keerthiveena B (2020) A microcontroller based machine vision approach for tomato grading and sorting using SVM classifier. Microprocess Microsyst 76:103090. https://doi.org/10.1016/j.micpro.2020.103090
Ding S, Su C, Yu J (2011) An optimizing BP neural network algorithm based on genetic algorithm. Artif Intell Rev 36(2):153–162. https://doi.org/10.1007/s10462-011-9208-z
Dwivedi RD, Singh M, Viladkar MN, Goel RK (2013) Prediction of tunnel deformation in squeezing grounds. Eng Geol 161:55–64. https://doi.org/10.1016/j.enggeo.2013.04.005
Feng XD, Jimenez R (2015) Predicting tunnel squeezing with incomplete data using bayesian networks. Eng Geol 195:214–224. https://doi.org/10.1016/j.enggeo.2015.06.017
Fritz P (2010) An analytical solution for axisymmetric tunnel problems in elasto-viscoplastic media. Int J Numer Anal Meth Geomech 8(4):325–342. https://doi.org/10.1002/nag.1610080403
Gao F, Stead D, Kang H (2015) Numerical simulation of squeezing failure in a coal mine roadway due to mining-induced stresses. Rock Mech Rock Eng 48(4):1635–1645. https://doi.org/10.1007/s00603-014-0653-2
Ghasemi E, Gholizadeh H (2018) Development of two empirical correlations for tunnel squeezing prediction using binary logistic regression and linear discriminant analysis. Geotech Geol Eng 37(4):3435–3446. https://doi.org/10.1007/s10706-018-00758-0
Goel RK, Jethwa JL, Paithankar AG (1995) Tunnelling through the young Himalayas—A case history of the maneri-uttarkashi power tunnel. Eng Geol 39(1–2):31–44. https://doi.org/10.1016/0013-7952(94)00002-J
Hoek E, Marinos P (2000) Predicting tunnel squeezing problems in weak heterogeneous rock masses. Tunn Tunn Int 32(11):45–51
Hoek E (2001) Big tunnels in bad rock. J Geotech Geoenviron Eng 127(9):726–740. https://doi.org/10.1061/(ASCE)1090-0241(2001)127:9(726)
Jethwa JL, Singh B, Singh B (1984) Estimation of ultimate rock pressure for tunnel linings under squeezing rock conditions—a new approach. In: Brown ET, Hudson JA (eds) Design and performance of underground excavations, ISRM symposium. Cambridge, pp 231–238
Jimenez R, Recio D (2011) A linear classifier for probabilistic prediction of squeezing conditions in Himalayan tunnels. Eng Geol 121:101–109. https://doi.org/10.1016/j.enggeo.2011.05.006
Karush-Kuhn-Tucker, 2020, (Access online: http://apmonitor.com/me575/index.php/Main/KuhnTucker)
Mahdevari S, Haghighat HS, Torabi SR (2013) A dynamically approach based on svm algorithm for prediction of tunnel convergence during excavation. Tunn Undergr Space Technol 38(9):59–68. https://doi.org/10.1016/j.tust.2013.05.002
Samadzadegan F, Bigdeli B, Ramzi P (2010) A multiple classifier system for classification of LIDAR remote sensing data using multi-class SVM. In: 9th international workshop on multiple classifier systems, Cairo, EGYPT , 5997, 254-263.
Shafiei A, Parsaei H, Dusseault MB, (2012) Rock squeezing prediction by a support vector machine classifier. In: 46th US Rock Mechanics/Geomechanics Symposium. Chicago. USA.
Singh B, Jethwa JL, Dube AK, Singh B (1992) Correlation between observed support pressure and rock mass quality. Tunn Undergr Space Technol 7(1):59–74. https://doi.org/10.1016/0886-7798(92)90114-W
Singh B, Goel RK, Jethwa JL (1997) Support pressure assessment in arched underground openings through poor rock masses. Eng Geol 48(1–2):59–81. https://doi.org/10.1016/S0013-7952(97)81914-X
Sun Y, Feng X, Yang L (2018) Predicting tunnel squeezing using multiclass support vector machines. Adv in Civil Eng 2018:4543984. https://doi.org/10.1155/2018/4543984
Wang J (2020) The key way is to release the genuine rock pressure—discussion on problems of tunnelling in squeezing ground. Modern Tunn Technol 57(4):1–11
Wang XC, Shi F (2013) Matlab neural network analysis of 43 cases. Beijing University of Aeronautics and Astronautics Press, Beijing, China
Wen J, Li ZJ, Wei LS, Zhen H (2002) The improvements of BP neural network learning algorithm. In: IEEE: 16th world computer congress. Beijing. China, pp 1647–1649. https://doi.org/10.1109/ICOSP.2000.893417
Yassaghi A, Salari-Rad H (2005) Squeezing rock conditions at an igneous contact zone in the Taloun tunnels, Tehran-Shomal freeway, Iran: a case study. Int J Rock Mech Mining Sci 42(1):95–108. https://doi.org/10.1016/j.ijrmms.2004.07.002
Zhang J, Li D, Wang Y (2020) Predicting tunnel squeezing using a hybrid classifier ensemble with incomplete data. Bull Eng Geol Env 79(6):3245–3256. https://doi.org/10.1007/s10064-020-01747-5
Zhao HH, Liu H (2020) Multiple classifiers fusion and CNN feature extraction for handwritten digits recognition. Granular Comput 5(3):411–418. https://doi.org/10.1007/s41066-019-00158-6
Zhou G, Sun Y, Peng J (2018) Application of genetic algorithm based BP neural network to parameter inversion of surrounding rock and deformation prediction. Modern Tunn Technol 55(01):107–113. https://doi.org/10.13807/j.cnki.mtt.2018.01.015
Zhou ZH (2016) Machine learning. Tsinghua University Press, Beijing, China
Acknowledgements
This work was supported by the National Science Foundation of China (Grant Nos. 51978668, 51968005), and the Guangxi University Young and Middle-Aged Teachers’ Basic Scientific Research Ability Improvement Project (2020ky01011).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that there are no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix 1
Appendix 1
No | D(m) | H(m) | Q | K(Mpa) | Squeezing | Reference |
---|---|---|---|---|---|---|
1 | 6.00 | 150.00 | 0.40 | 26.19 | − 1 | Sun et al. (2018) |
2 | 6.00 | 200.00 | 0.40 | 20.00 | − 1 | |
3 | 5.80 | 350.00 | 0.50 | 2.53 | 1 | |
4 | 4.80 | 225.00 | 3.60 | 1000.00 | − 1 | |
5 | 4.80 | 340.00 | 1.80 | 500.00 | − 1 | |
6 | 4.80 | 550.00 | 5.10 | 1600.00 | − 1 | |
7 | 12.00 | 220.00 | 0.80 | 32.89 | − 1 | |
8 | 13.00 | 52.00 | 15.00 | 16.67 | − 1 | |
9 | 3.00 | 280.00 | 0.05 | 9.80 | 1 | |
10 | 3.00 | 280.00 | 0.02 | 5.96 | 1 | |
11 | 9.00 | 680.00 | 0.05 | 9.90 | 1 | |
12 | 9.00 | 280.00 | 0.02 | 48.56 | 1 | |
13 | 4.20 | 100.00 | 0.01 | 88.96 | 1 | |
14 | 4.00 | 112.00 | 0.01 | 71.28 | 1 | |
15 | 4.30 | 111.00 | 0.01 | 1936.00 | − 1 | |
16 | 4.00 | 112.00 | 0.01 | 936.00 | − 1 | |
17 | 4.00 | 112.00 | 0.01 | 651.00 | − 1 | |
18 | 4.00 | 140.00 | 0.01 | 430.00 | − 1 | |
19 | 4.20 | 100.00 | 0.01 | 31.72 | 1 | |
20 | 4.00 | 138.00 | 0.01 | 1934.00 | − 1 | |
21 | 4.40 | 212.00 | 0.04 | 5324.00 | − 1 | |
22 | 5.00 | 300.00 | 0.05 | 1430.00 | − 1 | |
23 | 4.00 | 112.00 | 0.06 | 458.00 | − 1 | |
24 | 4.00 | 95.00 | 0.07 | 933.00 | − 1 | |
25 | 4.00 | 218.00 | 0.07 | 739.00 | − 1 | |
26 | 4.00 | 98.00 | 0.08 | 933.00 | − 1 | |
27 | 5.00 | 284.00 | 0.09 | 68.55 | 1 | |
28 | 5.00 | 300.00 | 0.09 | 664.29 | − 1 | |
29 | 4.00 | 261.00 | 0.10 | 931.00 | − 1 | |
30 | 4.00 | 198.00 | 0.14 | 934.00 | − 1 | |
31 | 4.00 | 225.00 | 0.14 | 1430.00 | − 1 | |
32 | 5.00 | 130.00 | 0.20 | 936.00 | − 1 | |
33 | 4.10 | 158.00 | 0.23 | 650.00 | − 1 | |
34 | 5.00 | 276.00 | 0.25 | 940.00 | − 1 | |
35 | 5.00 | 276.00 | 0.28 | 652.00 | − 1 | |
36 | 4.00 | 126.00 | 0.30 | 461.00 | − 1 | |
37 | 4.00 | 114.00 | 0.47 | 648.00 | − 1 | |
38 | 4.00 | 114.00 | 0.60 | 556.00 | − 1 | |
39 | 4.60 | 300.00 | 0.02 | 7.71 | 1 | |
40 | 4.80 | 350.00 | 0.50 | 25.32 | 1 | |
41 | 4.80 | 800.00 | 2.50 | 48.99 | 1 | |
42 | 7.00 | 285.00 | 0.10 | 9.79 | 1 | |
43 | 7.00 | 410.00 | 0.30 | 9.79 | 1 |
No | D(m) | H(m) | Q | K(Mpa) | Squeezing | Reference |
---|---|---|---|---|---|---|
44 | 7.00 | 415.00 | 0.88 | 9.79 | 1 | Sun et al. (2018) |
45 | 2.50 | 480.00 | 0.80 | 9.84 | 1 | |
46 | 7.00 | 500.00 | 1.00 | 9.79 | 1 | |
47 | 2.50 | 510.00 | 0.88 | 9.84 | 1 | |
48 | 4.60 | 240.00 | 0.12 | 3.97 | 1 | |
49 | 4.60 | 440.00 | 0.05 | 3.97 | 1 | |
50 | 4.60 | 450.00 | 0.06 | 3.97 | 1 | |
51 | 4.60 | 400.00 | 0.03 | 3.98 | 1 | |
52 | 4.60 | 400.00 | 0.05 | 3.98 | 1 | |
53 | 4.60 | 200.00 | 0.02 | 2.98 | 1 | |
54 | 4.60 | 325.00 | 0.03 | 2.98 | 1 | |
55 | 4.60 | 400.00 | 0.51 | 2.98 | − 1 | |
56 | 5.80 | 700.00 | 0.30 | 9.81 | 1 | |
57 | 5.80 | 550.00 | 1.70 | 9.81 | 1 | |
58 | 5.80 | 635.00 | 4.00 | 9.81 | 1 | |
59 | 5.80 | 650.00 | 4.12 | 9.81 | 1 | |
60 | 5.80 | 450.00 | 0.31 | 5.10 | 1 | |
61 | 5.80 | 750.00 | 0.50 | 8.10 | 1 | |
62 | 7.00 | 450.00 | 0.59 | 9.67 | 1 | |
63 | 6.80 | 337.00 | 0.01 | 44.76 | 1 | |
64 | 6.80 | 337.00 | 0.01 | 16.05 | 1 | |
65 | 6.80 | 337.00 | 0.01 | 22.58 | 1 | |
66 | 6.80 | 337.00 | 0.01 | 36.36 | 1 | |
67 | 6.80 | 337.00 | 0.08 | 14.09 | 1 | |
68 | 8.70 | 550.00 | 0.03 | 39.13 | 1 | |
69 | 8.70 | 600.00 | 0.02 | 90.71 | 1 | |
70 | 8.70 | 600.00 | 0.03 | 34.48 | 1 | |
71 | 8.70 | 600.00 | 0.02 | 26.20 | 1 | |
72 | 8.70 | 600.00 | 0.02 | 28.48 | 1 | |
73 | 8.70 | 620.00 | 0.02 | 26.20 | 1 | |
74 | 8.70 | 620.00 | 0.01 | 14.67 | 1 | |
75 | 8.70 | 620.00 | 0.01 | 14.67 | 1 | |
76 | 8.70 | 620.00 | 0.01 | 14.67 | 1 | |
77 | 8.70 | 620.00 | 0.02 | 26.20 | 1 | |
78 | 8.70 | 620.00 | 0.02 | 26.10 | 1 | |
79 | 8.70 | 620.00 | 0.03 | 50.80 | 1 | |
80 | 8.70 | 580.00 | 0.02 | 26.20 | 1 | |
81 | 8.70 | 580.00 | 0.03 | 74.66 | 1 | |
82 | 8.70 | 550.00 | 0.03 | 39.87 | 1 | |
83 | 8.70 | 575.00 | 0.01 | 21.17 | 1 | |
84 | 11.00 | 700.00 | 0.42 | 7.43 | 1 |
No | D(m) | H(m) | Q | K(Mpa) | Squeezing | Reference |
---|---|---|---|---|---|---|
85 | 11.00 | 700.00 | 0.33 | 9.14 | 1 | Sun et al. (2018) |
86 | 11.00 | 750.00 | 0.33 | 9.14 | 1 | |
87 | 11.00 | 600.00 | 0.25 | 9.14 | 1 | |
88 | 11.00 | 850.00 | 0.06 | 20.40 | 1 | |
89 | 11.00 | 600.00 | 0.03 | 33.33 | 1 | |
90 | 11.00 | 300.00 | 0.00 | 16.50 | 1 | |
91 | 11.00 | 400.00 | 0.00 | 17.00 | 1 | |
92 | 11.00 | 800.00 | 0.19 | 17.14 | 1 | |
93 | 6.50 | 300.00 | 0.03 | 10.00 | 1 | |
94 | 6.50 | 312.00 | 0.09 | 34.67 | 1 | |
95 | 6.50 | 280.00 | 0.08 | 29.33 | 1 | |
96 | 6.50 | 270.00 | 0.13 | 15.91 | 1 | |
97 | 6.50 | 285.00 | 0.06 | 12.80 | 1 | |
98 | 6.50 | 280.00 | 0.03 | 11.54 | 1 | |
99 | 6.50 | 280.00 | 0.04 | 12.50 | 1 | |
100 | 6.00 | 727.00 | 2.29 | 5.88 | 1 | |
101 | 6.00 | 736.00 | 2.43 | 7.69 | 1 | |
102 | 6.00 | 733.00 | 2.90 | 6.25 | 1 | |
103 | 6.00 | 690.00 | 1.65 | 9.38 | 1 | |
104 | 13.00 | 577.00 | 1.52 | 11.11 | 1 | |
105 | 5.40 | 199.70 | 0.02 | 1217.16 | 1 | |
106 | 5.40 | 217.50 | 0.01 | 1217.16 | 1 | |
107 | 5.40 | 252.20 | 0.01 | 1523.07 | 1 | |
108 | 5.40 | 246.30 | 0.01 | 1523.07 | 1 | |
109 | 5.40 | 283.90 | 0.01 | 1645.38 | 1 | |
110 | 5.40 | 284.50 | 0.01 | 1828.98 | 1 | |
111 | 5.40 | 210.80 | 0.01 | 1575.72 | 1 | |
112 | 5.40 | 237.70 | 0.01 | 1575.72 | 1 | |
113 | 5.40 | 230.00 | 0.02 | 1217.16 | 1 | |
114 | 5.40 | 222.60 | 0.02 | 1217.16 | 1 | |
115 | 5.40 | 80.00 | 93.50 | 0.00 | − 1 | |
116 | 5.40 | 190.00 | 7.45 | 0.00 | − 1 | |
117 | 5.40 | 130.00 | 1.53 | 0.00 | − 1 | |
118 | 3.5 | 285 | 0.1 | 9.79 | 1 | |
119 | 3.5 | 410 | 0.3 | 9.79 | − 1 | |
120 | 3.5 | 415 | 0.88 | 9.79 | 1 | Dwiviedi, et al.(2013) |
121 | 1.25 | 480 | 0.8 | 9.84 | − 1 | |
122 | 3.5 | 500 | 1 | 9.79 | − 1 | |
123 | 1.25 | 510 | 0.88 | 9.84 | − 1 | |
124 | 2.3 | 240 | 0.12 | 3.97 | −1 | |
125 | 2.3 | 440 | 0.05 | 3.97 | 1 | |
126 | 2.3 | 450 | 0.06 | 3.97 | 1 | |
127 | 2.3 | 400 | 0.03 | 3.98 | − 1 |
No | D(m) | H(m) | Q | K(Mpa) | Squeezing | Reference |
---|---|---|---|---|---|---|
128 | 2.3 | 400 | 0.05 | 3.98 | 1 | Dwiviedi, et al.(2013) |
129 | 2.3 | 200 | 0.02 | 2.98 | 1 | |
130 | 2.3 | 325 | 0.03 | 2.98 | −1 | |
131 | 2.3 | 400 | 0.512 | 2.98 | 1 | |
132 | 1.5 | 280 | 0.05 | 9.8 | 1 | |
133 | 1.5 | 280 | 0.022 | 5.96 | 1 | |
134 | 4.5 | 680 | 0.05 | 9.9 | 1 | |
135 | 4.5 | 280 | 0.022 | 48.56 | −1 | |
136 | 2.9 | 700 | 0.3 | 9.81 | −1 | |
137 | 2.9 | 550 | 1.7 | 9.81 | −1 | |
138 | 2.9 | 635 | 4 | 9.81 | −1 | |
139 | 2.9 | 650 | 4.12 | 9.81 | −1 | |
140 | 2.9 | 450 | 0.31 | 5.1 | −1 | |
141 | 2.9 | 750 | 0.5 | 8.1 | −1 | |
142 | 3.5 | 450 | 0.59 | 9.67 | −1 | |
143 | 3.4 | 337 | 0.011 | 8.97 | 1 | |
144 | 4.35 | 600 | 0.015 | 34.52 | 1 | |
145 | 4.35 | 600 | 0.023 | 90.71 | 1 | |
146 | 4.35 | 600 | 0.025 | 34.17 | 1 | |
147 | 4.35 | 600 | 0.018 | 26.20 | 1 | |
148 | 4.35 | 600 | 0.023 | 28.48 | 1 | |
149 | 4.35 | 620 | 0.02 | 26.20 | −1 | |
150 | 4.35 | 620 | 0.008 | 14.67 | −1 | |
151 | 4.35 | 620 | 0.009 | 14.67 | −1 | |
152 | 4.35 | 620 | 0.01 | 26.20 | 1 | |
153 | 4.35 | 620 | 0.009 | 14.67 | −1 | |
154 | 4.35 | 620 | 0.016 | 26.20 | −1 | |
155 | 4.35 | 620 | 0.02 | 26.20 | 1 | |
156 | 4.35 | 620 | 0.025 | 56.96 | −1 | |
157 | 4.35 | 580 | 0.023 | 26.20 | 1 | |
158 | 4.35 | 580 | 0.025 | 74.66 | 1 | |
159 | 4.35 | 575 | 0.001 | 34.17 | 1 | |
160 | 4.35 | 550 | 0.025 | 39.87 | 1 | |
161 | 2 | 98 | 0.08 | 933.0 | −1 | |
162 | 2.15 | 111 | 0.008 | 1936.0 | −1 | |
163 | 2 | 112 | 0.06 | 458.0 | −1 | |
164 | 2 | 126 | 0.3 | 461.0 | −1 | |
165 | 2 | 138 | 0.013 | 1934.0 | −1 | |
166 | 2 | 198 | 0.14 | 934.0 | −1 | |
167 | 2 | 261 | 0.095 | 931.0 | −1 | |
168 | 2 | 95 | 0.065 | 933.0 | −1 | |
169 | 2.5 | 130 | 0.2 | 936.0 | −1 |
No | D(m) | H(m) | Q | K(Mpa) | Squeezing | Reference |
---|---|---|---|---|---|---|
170 | 2.05 | 158 | 0.23 | 650.0 | −1 | Dwiviedi et al.(2013) |
171 | 2.5 | 276 | 0.25 | 940.0 | −1 | |
172 | 2.5 | 276 | 0.28 | 652.0 | −1 | |
173 | 2 | 140 | 0.009 | 430.0 | −1 | |
174 | 2.5 | 300 | 0.05 | 1430.0 | −1 | |
175 | 2 | 225 | 0.14 | 1430.0 | −1 | |
176 | 2 | 218 | 0.07 | 739.0 | −1 | |
177 | 2 | 114 | 0.47 | 648.0 | −1 | |
178 | 2 | 114 | 0.6 | 556.0 | −1 | |
179 | 2 | 112 | 0.008 | 936.0 | −1 | |
180 | 2 | 112 | 0.008 | 651.0 | −1 |
Rights and permissions
About this article
Cite this article
Huang, Z., Liao, M., Zhang, H. et al. Predicting Tunnel Squeezing Using the SVM-BP Combination Model. Geotech Geol Eng 40, 1387–1405 (2022). https://doi.org/10.1007/s10706-021-01970-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10706-021-01970-1