Abstract
Using machine-learning models as surrogate models is a popular technique to increase the computational efficiency of stochastic analysis. In this technique, a smaller number of numerical simulations are conducted for a case, and obtained results are used to train machine-learning surrogate models specific for this case. This study presents a new framework using deep learning, where models are trained with a big dataset covering any soil properties, spatial variabilities, or load conditions encountered in practice. These models are very accurate for new data without re-training. So, the small number of numerical simulations and training process are not needed anymore, which further increases efficiency. The prediction of bearing capacity of shallow strip footings is taken as an example. We start with a simple scenario, and progressively consider more complex scenarios until the full problem is considered. More than 12,000 data are used in training. It is shown that one-hidden-layer fully connected networks can give reasonable results for simple problems, but they are ineffective for complex problems, where deep neural networks show a competitive edge, and a deep-learning model achieves a very high accuracy (the root-mean-square relative error is 3.1% for unseen data). In testing examples, this model is proven very accurate if the parameters of specific cases are well in the defined limits. Otherwise, the capability of deep-learning models can be extended by simply generating more data outside the current limits and re-training the models.
Similar content being viewed by others
Abbreviations
- AvgPool:
-
Average pooling
- CNN:
-
Convolutional neural networks
- Conv:
-
Convolutional
- COV:
-
Coefficient of variation
- FC:
-
Fully connected
- FDQS:
-
Finite-difference quasi-static
- FEQS:
-
Finite-element quasi-static
- FES:
-
Finite-element static
- LC:
-
Locally connected
- MaxPool:
-
Max pooling
- PDF:
-
Probability density function
- RMSRE:
-
Root-mean-square relative error
- \(A\) :
-
Extent of slip lines
- \(B\) :
-
Width of strip footing
- \(c\) :
-
Cohesion
- \(c^{\prime}\) :
-
Dimensionless cohesion, \(c^{\prime} = \frac{c}{{c_{r} }}\)
- \({\mathbf{C}}^{\prime}\) :
-
Finite field of the dimensionless cohesion, represented by a matrix
- \(c_{r}\) :
-
A reference strength
- \(f_{R} \left( R \right)\) :
-
Probability density function of a random variable/field \(R\)
- \(H_{i}\) :
-
A random field
- \({\varvec{H}}\) :
-
Vector of random fields
- \(l_{x}\) :
-
Horizontal scale of fluctuation for a random field
- \(l_{y}\) :
-
Vertical scale of fluctuation for a random field
- \(N_{MC}\) :
-
Number of Monte Carlo simulations
- \(q_{0}\) :
-
Overburden load
- \(q_{0}^{^{\prime}}\) :
-
Dimensionless overburden load \(\frac{{q_{0} }}{{c_{r} }}\)
- \(q_{u}\) :
-
Bearing capacity
- \(q_{u}^{^{\prime}}\) :
-
Dimensionless bearing capacity \(\frac{{q_{u} }}{{c_{r} }}\)
- \(R\) :
-
Structural response
- \(s_{u}\) :
-
Undrained strength
- \(v\) :
-
Coefficient of variation for a random field/variable
- \({\varvec{x}}\) :
-
Position vector
- \(\user2{x^{\prime}}\) :
-
Dimensionless position vector, \(\user2{x^{\prime}} = \frac{{\varvec{x}}}{B}\)
- \(\gamma\) :
-
Unit weight
- \(\gamma ^{\prime}\) :
-
Dimensionless unit weight, \(\gamma ^{\prime} = \frac{\gamma B}{{c_{r} }}\)
- \({{\varvec{\Gamma}}}^{\prime}\) :
-
Finite field of the dimensionless unit weight, represented by a matrix
- \(\phi\) :
-
Friction angle
- \({{\varvec{\Phi}}}\) :
-
Finite field of the friction angle, represented by a matrix
- \(\mu\) :
-
Mean of a random field/variable
References
Anderson JA (1995) An introduction to neural networks. MIT Press, London
Beck JL, Au S-K (2002) Bayesian updating of structural models and reliability using markov chain Monte Carlo simulation. J Eng Mech 128(4):380–391. https://doi.org/10.1061/(asce)0733-9399(2002)128:4(380)
Bolton M (1979) A guide to soil mechanics. Macmillan Education, London
Chen CT, Gu GX (2020) Generative deep neural networks for inverse materials design using backpropagation and active learning. Adv Sci. https://doi.org/10.1002/advs.201902607
Das BM (1990) Principles of foundation engineering
Fenton GA, Griffiths DV (2003) Bearing-capacity prediction of spatially random c: ø soils. Can Geotech J 40(1):54–65. https://doi.org/10.1139/t02-086
Gao X, Shi M, Song X, Zhang C, Zhang H (2019) Recurrent neural networks for real-time prediction of TBM operating parameters. Autom Constr 98:225–235. https://doi.org/10.1016/j.autcon.2018.11.013
Griffiths DV, Fenton GA (2001) Bearing capacity of spatially random soil: the undrained clay Prandtl problem revisited. Géotechnique 51(4):351–359. https://doi.org/10.1680/geot.51.4.351.39396
He X, Liang D, Bolton MD (2018) Run-out of cut-slope landslides: mesh-free simulations. Géotechnique 68(1):50–63. https://doi.org/10.1680/jgeot.16.P.221
He X, Xu H, Sabetamal H, Sheng D (2020) Machine learning aided stochastic reliability analysis of spatially variable slopes. Comput Geotech 126:103711. https://doi.org/10.1016/j.compgeo.2020.103711
He X, Wang F, Li W, Sheng D (2021) Efficient reliability analysis considering uncertainty in random field parameters: trained neural networks as surrogate models. Comput Geotech 136:104212. https://doi.org/10.1016/j.compgeo.2021.104212
Huang J, Fenton G, Griffiths DV, Li D, Zhou C (2017) On the efficient estimation of small failure probability in slopes. Landslides 14(2):491–498. https://doi.org/10.1007/s10346-016-0726-2
Jiang SH, Huang JS (2016) Efficient slope reliability analysis at low-probability levels in spatially variable soils. Comput Geotech 75:18–27. https://doi.org/10.1016/j.compgeo.2016.01.016
Kang F, Xu Q, Li J (2016) Slope reliability analysis using surrogate models via new support vector machines with swarm intelligence. Appl Math Model 40(11–12):6105–6120. https://doi.org/10.1016/j.apm.2016.01.050
Liu L, Zhang S, Cheng YM, Liang L (2019) Advanced reliability analysis of slopes in spatially variable soils using multivariate adaptive regression splines. Geosci Front 10(2):671–682. https://doi.org/10.1016/j.gsf.2018.03.013
Müller S, GSTools. https://geostat-framework.org/
Phoon K-K, Kulhawy FH (1999) Characterization of geotechnical variability. Can Geotech J 36(4):612–624. https://doi.org/10.1139/t99-038
Silver D et al (2016) Mastering the game of Go with deep neural networks and tree search. Nature 529(7585):484–489. https://doi.org/10.1038/nature16961
Szegedy C et al (2015) Going deeper with convolutions. Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit. https://doi.org/10.1109/CVPR.2015.7298594
Taigman Y, Yang M, Ranzato M, Wolf L (2014) DeepFace: closing the gap to human-level performance in face verification. In: Proceedings of the 2014 IEEE conference on computer vision and pattern recognition, pp 1701–1708. https://doi.org/10.1109/CVPR.2014.220.
Wang P, Yin ZY (2020) Micro-mechanical analysis of caisson foundation in sand using DEM: particle breakage effect. Ocean Eng 215:107921. https://doi.org/10.1016/j.oceaneng.2020.107921
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
He, X., Wang, F., Li, W. et al. Deep learning for efficient stochastic analysis with spatial variability. Acta Geotech. 17, 1031–1051 (2022). https://doi.org/10.1007/s11440-021-01335-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11440-021-01335-1