A hierarchical system to predict behavior of soil and cantilever sheet wall by data-driven models
Nang Duc BUI1, Hieu Chi PHAN2(), Tiep Duc PHAM1, Ashutosh Sutra DHAR2
1. Institute of Techniques for Special Engineering, Le Quy Don Technical University, Hanoi 100000, Vietnam 2. Faculty of Engineering and Applied Science, Memorial University of Newfoundland, St. John’s, NL A1B 3X5, Canada
The study proposes a framework combining machine learning (ML) models into a logical hierarchical system which evaluates the stability of the sheet wall before other predictions. The study uses the hardening soil (HS) model to develop a 200-sample finite element analysis (FEA) database, to develop the ML models. Consequently, a system containing three trained ML models is proposed to first predict the stability status (random forest classification, RFC) followed by 1) the cantilever top horizontal displacement of sheet wall (artificial neural network regression models, RANN1) and 2) vertical settlement of soil (RANN2). The uncertainty of this data-driven system is partially investigated by developing 1000 RFC models, based on the application of random sampling technique in the data splitting process. Investigation on the distribution of the evaluation metrics reveals negative skewed data toward the 1.0000 value. This implies a high performance of RFC on the database with medians of accuracy, precision, and recall, on test set are 1.0000, 1.0000, and 0.92857, respectively. The regression ANN models have coefficient of determinations on test set, as high as 0.9521 for RANN1, and 0.9988 for RANN2, respectively. The parametric study for these regressions is also provided to evaluate the relative insight influence of inputs to output.
. [J]. Frontiers of Structural and Civil Engineering, 2022, 16(6): 667-684.
Nang Duc BUI, Hieu Chi PHAN, Tiep Duc PHAM, Ashutosh Sutra DHAR. A hierarchical system to predict behavior of soil and cantilever sheet wall by data-driven models. Front. Struct. Civ. Eng., 2022, 16(6): 667-684.
control case (for parametric study in Subsection 3.4)
1
L
m
10
12
2
EI
kN·m2/m
6540.7
44982
3
Hw
m
0
2
4
H
m
3.05; 4.05; 5.05; 5.53; 5.83
4
5
q
kN/m2
0
5
6
B
m
0
5
7
γunsat
kN/m3
16
18.0
8
γsat
kN/m3
19.85
19.0
9
Eoed
kN/m2
10000
9958
10
c
kN/m2
0
1
11
phi
°
39.4
27
Tab.1
Fig.4
Fig.5
Fig.6
No.
L (m)
EI (kN·m2/m)
Hw (m)
H (m)
q (kN/m2)
B (m)
γunsat (kN/m3)
γsat (kN/m3)
Eoed (kN/m2)
c (kN/m2)
phi (° )
stable
Δ (mm)
not failure?
1
8
44982
40
5
0
0
18
19
23000
41
15
1
33.61
1
2
9
44982
40
5
0
0
18
19
23000
41
15
1
30.12
1
3
10
44982
40
5
0
0
18
19
23000
41
15
1
28.69
1
4
12
44982
40
5
0
0
18
19
23000
41
15
1
28.14
1
5
16
44982
40
5
0
0
18
19
23000
41
15
1
27.97
1
6
8
44982
40
5
0
0
19
20
9958
1
27
0
–
0*
7
9
44982
40
5
0
0
19
20
9958
1
27
0
–
0*
8
10
44982
40
5
0
0
19
20
9958
1
27
1
211.68
0*
9
12
44982
40
5
0
0
19
20
9958
1
27
1
164.33
1
10
16
44982
40
5
0
0
19
20
9958
1
27
1
155.18
1
198
10
44982
4
5
10
10
19
19.5
12000
32
22
1
40.67
1
199
12
44982
4
5
10
10
19
19.5
12000
32
22
1
39.1
1
200
16
44982
4
5
10
10
19
19.5
12000
32
22
1
39.14
1
Tab.2
sample #
1
2
…
199
200
x (m)
y(x) (m)
x (m)
y(x) (m)
…
x (m)
y(x) (m)
x (m)
y(x) (m)
1
0.0000
0.0082
0.0000
0.0061
…
0.0000
0.0140
0.0000
0.0134
2
0.2951
0.0113
0.2951
0.0092
…
0.2951
0.0232
0.2951
0.0234
3
0.2951
0.0113
0.2951
0.0092
…
0.2951
0.0232
0.2951
0.0234
4
0.6264
0.0123
0.6264
0.0104
…
0.6264
0.0251
0.6264
0.0254
5
0.6264
0.0123
0.6264
0.0104
…
0.6264
0.0251
0.6264
0.0254
53
55.2219
0.0009
55.2219
0.0009
…
55.2219
0.0019
55.2219
0.0020
54
57.6109
0.0009
57.6109
0.0009
…
57.6109
0.0019
57.6109
0.0019
55
60.0000
0.0009
60.0000
0.0009
…
60.0000
0.0019
60.0000
0.0019
Tab.3
No.
variable
unit
count
min
max
1
L
(m)
200
8
16
2
EI
(kN·m2/m)
200
44982
110250
3
Hw
(m)
200
1
40
4
H
(m)
200
2.5
5
5
q
(kN/m2)
200
0
15
6
B
(m)
200
0
15
7
γunsat
(kN/m3)
200
17
19
8
γsat
(kN/m3)
200
17.6
20
9
Eoed
(kN/m2)
200
5479
78000
10
c
(kN/m2)
200
1
41
11
°
200
15
34
Tab.4
Fig.7
No.
model
data ID
predicted variable
unit
valid count
min
max
samples for training
samples for testing
samples for graphical evaluation
1
RFC
data1
Statusa)
–
200
–1
1
160
40
–
2
RANN1
data2
Δb)
mm
135
8.2
200
108
27
–
3
RANN2
data3
y(x)c)
mm
135 × 31 = 4185d)
–13.1
854.7
104 × 31 = 3224
26 × 31 = 806
5e) × 31 = 155
Tab.5
variable
L
EI
Hw
H
q
B
γunsat
γsat
Eoed
c
x
stable
0.3674
–0.0642
0.1327
–0.0832
0.1362
0.1184
–0.0698
–0.1439
0.4293
0.4100
–0.0055
–
Δ
0.3235
–0.0447
0.0012
0.1555
–0.1327
–0.1248
0.0838
0.1510
–0.4126
–0.4753
0.0798
–
y(x)
0.0614
–0.0077
–0.0771
-0.0374
0.2938
0.2575
–0.2399
–0.2714
–0.3055
–0.3830
0.3683
–0.3926
Tab.6
Fig.8
Fig.9
layer
number of nodes
trainable weights
input layer
12 + 1 (bias)
–
hidden layer 1
64 + 1 (bias)
(12+1) × 64 = 832
hidden layer 2
512 + 1 (bias)
(64+1) × 512 = 33280
hidden layer 3
32 + 1 (bias)
(512+1) × 32 = 16416
output layer
1
(32+1) × 1 = 33
total
625
50561
Tab.7
layer
number of nodes
trainable weights
input layer
12 + 1(x) + 1 (bias)
–
hidden layer 1
64 + 1 (bias)
(13+1) × 64 = 896
hidden layer 2
64 + 1 (bias)
(64+1) × 64 = 4160
hidden layer 3
64 + 1 (bias)
(64+1) × 64 = 4160
hidden layer 4
64 + 1 (bias)
(64+1) × 64 = 4160
hidden layer 5
32 + 1 (bias)
(64+1) × 32 = 2080
hidden layer 6
8 + 1 (bias)
(32+1) × 8 = 264
output layer
1
(8+1) × 1 = 9
total
317
15729
Tab.8
metric
equation
RANN1
RANN2
on train set
on test set
on train set
on test set
mean absolute error
0.201697
4.531986
0.000288
0.00103
mean squared error
0.11290
49.2773
7.2152e−07
9.6209e−06
coefficient of determination
0.9999
0.95212
0.9999
0.9988
Tab.9
Fig.10
Fig.11
Fig.12
Fig.13
Fig.14
Fig.15
No.
stable
Δ (FEA) (mm)
x* (m)
y(x) (FEA) (mm)
not failure?
output 1 (Δ) (mm)
output 2 (y(x)) (mm)
1
1
33.610
2.5
12.700
1
30.939
13.200
2
1
30.120
3.5
10.100
1
29.084
98.800
3
1
28.690
4
9.100
1
27.399
9.900
4
1
28.140
0.5
8.700
1
26.209
8.300
5
1
27.970
10
4.400
1
27.932
4.700
6
0
–
–
–
0
–1
–1
7
0
–
–
–
0
–1
–1
8
1
211.680
2.5
22.840
0
–1
–1
9
1
164.330
6.5
29.500
1
149.577
28.900
10
1
155.180
1.5
107.100
1
130.018
112.900
Tab.10
1
J Kwon. Investigation of the influence of an excavation on adjacent excavations, using neural networks. Journal of the Southern African Institute of Mining and Metallurgy, 1998, 98( 3): 147– 156
2
M I Ramadan, E H Ramadan, M M Khashila. Cantilever contiguous pile wall for supporting excavation in clay. Geotechnical and Geological Engineering, 2018, 36( 3): 1545– 1558 https://doi.org/10.1007/s10706-017-0407-5
3
H Poulos, L Chen. Pile response due to unsupported excavation-induced lateral soil movement. Canadian Geotechnical Journal, 1996, 33( 4): 670– 677 https://doi.org/10.1139/t96-091-312
A P Singh, K Chatterjee. Ground settlement and deflection response of cantilever sheet pile wall subjected to surcharge loading. Indian Geotechnical Journal, 2020, 50( 4): 540– 549
6
M S Es-haghi, M Abbaspour, T Rabczuk. Factors and failure patterns analysis for undrained seismic bearing capacity of strip footing above void. International Journal of Geomechanics, 2021, 21( 10): 04021188 https://doi.org/10.1061/(ASCE)GM.1943-5622.0002166
7
H C Phan, A S Dhar. Predicting pipeline burst pressures with machine learning models. International Journal of Pressure Vessels and Piping, 2021, 191 : 104384 https://doi.org/10.1016/j.ijpvp.2021.104384
8
C Anitescu, E Atroshchenko, N Alajlan, T Rabczuk. Artificial neural network methods for the solution of second order boundary value problems. Computers, Materials and Continua, 2019, 59( 1): 345– 359 https://doi.org/10.32604/cmc.2019.06641
9
E Samaniego, C Anitescu, S Goswami, V M Nguyen-Thanh, H Guo, K Hamdia, X Zhuang, T Rabczuk. An energy approach to the solution of partial differential equations in computational mechanics via machine learning: Concepts, implementation and applications. Computer Methods in Applied Mechanics and Engineering, 2020, 362 : 112790 https://doi.org/10.1016/j.cma.2019.112790
10
A K Verma, T N Singh, N K Chauhan, K Sarkar. A hybrid FEM–ANN approach for slope instability prediction. Journal of The Institution of Engineers (India): Series A, 2016, 97( 3): 171– 180
11
C Peng, W Wu, B Zhang. Three-dimensional simulations of tensile cracks in geomaterials by coupling meshless and finite element method. International Journal for Numerical and Analytical Methods in Geomechanics, 2015, 39( 2): 135– 154 https://doi.org/10.1002/nag.2298
12
A Chakraborty, D Goswami. Prediction of slope stability using multiple linear regression (MLR) and artificial neural network (ANN). Arabian Journal of Geosciences, 2017, 10( 17): 385 https://doi.org/10.1007/s12517-017-3167-x
13
S Chern J H Tsai L K Chien C Y Huang. Predicting lateral wall deflection in top-down excavation by neural network. International Journal of Offshore and Polar Engineering, 2009, 19(2): 151− 157
14
H Moayedi, M Mosallanezhad, A S A Rashid, W A W Jusoh, M A Muazu. A systematic review and meta-analysis of artificial neural network application in geotechnical engineering: Theory and applications. Neural Computing & Applications, 2020, 32( 2): 495– 518
15
M S Es-haghi M Sarcheshmehpour. A novel strategy for tall building optimization via combination of asymmetric genetic algorithm and machine learning methods. In: The 1st Online Conference on Algorithms. MDPI, 2021
16
H T Duong H C Phan T T Le N D Bui. Optimization design of rectangular concrete-filled steel tube short columns with Balancing Composite Motion Optimization and data-driven model. Structures, 2020, 28: 757– 765
17
H C Phan, L Le-Thanh, H Nguyen-Xuan. A semi-empirical approach and uncertainty analysis to pipes under hydrogen embrittlement degradation. International Journal of Hydrogen Energy, 2022, 47( 8): 5677– 5691
18
H C Phan, N D Bui. Failure assessment of defected pipe under strike-slip fault with data-driven models accounting to the model uncertainty. Neural Computing & Applications, 2021, 34 : 1541– 1555
19
N Attoh-Okine, E S Fekpe. Strength characteristics modeling of lateritic soils using adaptive neural networks. Construction & Building Materials, 1996, 10( 8): 577– 582 https://doi.org/10.1016/S0950-0618(96)00021-9
20
M Pala, N Caglar, M Elmas, A Cevik, M Saribiyik. Dynamic soil–structure interaction analysis of buildings by neural networks. Construction & Building Materials, 2008, 22( 3): 330– 342 https://doi.org/10.1016/j.conbuildmat.2006.08.015
21
M D Nazzal, O Tatari. Evaluating the use of neural networks and genetic algorithms for prediction of subgrade resilient modulus. International Journal of Pavement Engineering, 2013, 14( 4): 364– 373 https://doi.org/10.1080/10298436.2012.671944
22
D R Groholski, Y M Hashash. Development of an inverse analysis framework for extracting dynamic soil behavior and pore pressure response from downhole array measurements. International Journal for Numerical and Analytical Methods in Geomechanics, 2013, 37( 12): 1867– 1890 https://doi.org/10.1002/nag.2172
R Nazir, H Moayedi, A Pratikso, M Mosallanezhad. The uplift load capacity of an enlarged base pier embedded in dry sand. Arabian Journal of Geosciences, 2015, 8( 9): 7285– 7296 https://doi.org/10.1007/s12517-014-1721-3
25
A Ismail, D S Jeng. Modelling load–settlement behaviour of piles using high-order neural network (HON-PILE model). Engineering Applications of Artificial Intelligence, 2011, 24( 5): 813– 821 https://doi.org/10.1016/j.engappai.2011.02.008
O Yilmaz, M Eser, M Berilgen. Applications of engineering seismology for site characterization. Journal of Earth Science, 2009, 20( 3): 546– 554 https://doi.org/10.1007/s12583-009-0045-9
28
Z Cao, Y Wang, D Li. Quantification of prior knowledge in geotechnical site characterization. Engineering Geology, 2016, 203 : 107– 116 https://doi.org/10.1016/j.enggeo.2015.08.018
29
V K Dwivedi, R K Dubey, S Thockhom, V Pancholi, S Chopra, B K Rastogi. Assessment of liquefaction potential of soil in Ahmedabad Region, Western India. Journal of Indian Geophysical Union, 2017, 21( 2): 116– 123
30
C Hsein Juang, C J Chen, Y M Tien. Appraising cone penetration test based liquefaction resistance evaluation methods: Artificial neural network approach. Canadian Geotechnical Journal, 1999, 36( 3): 443– 454 https://doi.org/10.1139/t99-011
31
A M Hanna, D Ural, G Saygili. Evaluation of liquefaction potential of soil deposits using artificial neural networks. Engineering Computations, 2007, 24( 1): 5– 16 https://doi.org/10.1108/02644400710718547
32
Z Liu, J Shao, W Xu, H Chen, Y Zhang. An extreme learning machine approach for slope stability evaluation and prediction. Natural Hazards, 2014, 73( 2): 787– 804 https://doi.org/10.1007/s11069-014-1106-7
33
B Gordan, D Jahed Armaghani, M Hajihassani, M Monjezi. Prediction of seismic slope stability through combination of particle swarm optimization and neural network. Engineering with Computers, 2016, 32( 1): 85– 97 https://doi.org/10.1007/s00366-015-0400-7
34
A Li, S Khoo, A V Lyamin, Y Wang. Rock slope stability analyses using extreme learning neural network and terminal steepest descent algorithm. Automation in Construction, 2016, 65 : 42– 50 https://doi.org/10.1016/j.autcon.2016.02.004
35
I Ilia, I Koumantakis, D Rozos, G Koukis, P Tsangaratos. A geographical information system (GIS) based probabilistic certainty factor approach in assessing landslide susceptibility: The case study of Kimi, Euboea, Greece. In: IAEG XII Congress: Engineering Geology for Society and Territory. Turin: Springer, 2015, 1199– 1204
36
F Souza, N Ebecken. A data mining approach to landslide prediction. WIT Transactions on Information and Communication Technologies, 2004, 33
37
C Melchiorre, M Matteucci, A Azzoni, A Zanchi. Artificial neural networks and cluster analysis in landslide susceptibility zonation. Geomorphology, 2008, 94( 3-4): 379– 400 https://doi.org/10.1016/j.geomorph.2006.10.035
38
F K Huang G S Wang. ANN-based reliability analysis for deep excavation. In: EUROCON 2007—The International Conference on “Computer as a Tool”. Warsaw: IEEE, 2007
39
A T Goh, K Wong, B Broms. Estimation of lateral wall movements in braced excavations using neural networks. Canadian Geotechnical Journal, 1995, 32( 6): 1059– 1064 https://doi.org/10.1139/t95-103
Y Jun C Haiming. Artificial neural network’s application in intelligent prediction of surface settlement induced by foundation pit excavation. In: 2009 Second International Conference on Intelligent Computation Technology and Automation. Zhangjiajie: IEEE, 2009
42
C Koy, C Y Yune. Numerical analysis on consolidation of soft clay by sand drain with heat injection. Journal of the Korean Geotechnical Society, 2017, 33( 11): 45– 57
43
J Zhou, X Shi, K Du, X Qiu, X Li, H S Mitri. Feasibility of random-forest approach for prediction of ground settlements induced by the construction of a shield-driven tunnel. International Journal of Geomechanics, 2017, 17( 6): 04016129 https://doi.org/10.1061/(ASCE)GM.1943-5622.0000817
44
S Nikbakht, C Anitescu, T Rabczuk. Optimizing the neural network hyperparameters utilizing genetic algorithm. Journal of Zhejiang University, Science A, 2021, 22( 6): 407– 426 https://doi.org/10.1631/jzus.A2000384
45
W Zhang, R Zhang, C Wu, A T C Goh, S Lacasse, Z Liu, H Liu. State-of-the-art review of soft computing applications in underground excavations. Geoscience Frontiers, 2020, 11( 4): 1095– 1106
46
T Schanz, P A Vermeer, P G Bonnier. Beyond 2000 in Computational Geotechnics. London: Routledge, 1999, 281– 296
47
C Y Ou, C H Lai. Finite-element analysis of deep excavation in layered sandy and clayey soil deposits. Canadian Geotechnical Journal, 1994, 31( 2): 204– 214 https://doi.org/10.1139/t94-026
48
M Mansour, A Rashed, A Farag. Adopting numerical models for prediction of ground movements induced by deep excavation. International Journal of Recent Technology and Engineering, 2020, 8( 6): 976– 988
49
R B J Brinkgreve W M Swolfs E Engin D Waterman A Chesaru P Bonnier V Galavi. PLAXIS 2D Reference Manual. Delft: Delft University of Technology and PLAXIS bv, 2011
50
T T Le, H C Phan. Prediction of ultimate load of rectangular CFST columns using interpretable machine learning method. Advances in Civil Engineering, 2020, 2020 : 8855069
51
T T Le. Prediction of tensile strength of polymer carbon nanotube composites using practical machine learning method. Journal of Composite Materials, 2021, 55( 6): 787– 811 https://doi.org/10.1177/0021998320953540
52
L Breiman L Breiman J H Friedman R A Olshen C J Stone. Classification and Regression Trees. Boca Raton: Chapman & Hall, 1984
53
A Géron. Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems. Sebastopol: O’Reilly Media, 2019
54
T D Pham, N D Bui, T T Nguyen, H C Phan. Predicting the reduction of embankment pressure on the surface of the soft ground reinforced by sand drain with random forest regression. IOP Conference Series: Materials Science and Engineering, 2020, 869( 7): 072027
55
W S McCulloch, W Pitts. A logical calculus of the ideas immanent in nervous activity. Bulletin of Mathematical Biophysics, 1943, 5( 4): 115– 133 https://doi.org/10.1007/BF02478259
56
Y LeCun, Y Bengio. Convolutional networks for images, speech, and time series. The Handbook of Brain Theory and Neural Networks, 1995, 3361( 10): 1– 14
57
M A Carreira-Perpinan, G E Hinton. On contrastive divergence learning. In: The Tenth International Workshop on Artificial Intelligence and Statistics. Barbados: PMLR, 2005, 33– 40
M A Ranzato F J Huang Y L Boureau Y LeCun. Unsupervised learning of invariant feature hierarchies with applications to object recognition. In: 2007 IEEE Conference on Computer Vision and Pattern Recognition. Minneapolis: IEEE, 2007
60
A Chogueur, Z Abdeldjalil, P Reiffsteck. Parametric and comparative study of a flexible retaining wall. Periodica Polytechnica. Civil Engineering, 2018, 62( 2): 295– 307
61
M A Shahin, M B Jaksa, H R Maier. Artificial neural network applications in geotechnical engineering. Australian Geomechanics, 2001, 36( 1): 49– 62
62
H C Phan, T T Le, N D Bui, H T Duong, T D Pham. An empirical model for bending capacity of defected pipe combined with axial load. International Journal of Pressure Vessels and Piping, 2021, 191 : 104368 https://doi.org/10.1016/j.ijpvp.2021.104368
63
H C Phan, H T Duong. Predicting burst pressure of defected pipeline with principal component analysis and adaptive neuro fuzzy inference system. International Journal of Pressure Vessels and Piping, 2021, 189 : 104274 https://doi.org/10.1016/j.ijpvp.2020.104274