Modeling One-Dimensional Nonlinear Consolidation Problems by Physics-Informed Neural Network with Layer-Wise Locally Adaptive Activation Functions
Source abstract
Abstract
The study on soil consolidation and settlement is of great importance in geotechnical engineering practice. Nowadays, physics-informed neural networks (PINN) are becoming more and more popular in solving geotechnical engineering problems thanks to their meshless, physically constrained, and data-driven nature. Although there have been some successful applications in one-dimensional (1D) consolidation problems in saturated soils, the ability and stability to deal with more complex boundary conditions remain to be tested. In this paper, the effects of activation function and random state on the PINN are investigated for solving two 1D consolidation problems in saturated soils, and the proposed method for inverse modeling of the two 1D consolidation problems. The results show that PINN with layer-wise locally adaptive activation functions improves the convergence speed and prediction accuracy of the PINN for solving the 1D nonlinear soil consolidation problems, and at the same time the robustness of the model to random states. Moreover, the proposed method still converges faster in the inverse modeling of 1D consolidation problems.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Model Reduction and Neural Networks · Dam Engineering and Safety · Drilling and Well Engineering
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Department of Civil Engineering, Shanghai University, Shanghai 200444, China
Department of Civil Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
Funding
National Natural Science Foundation of China
References (45)
Show source-provided references
1. Terzaghi, K. (1925). Erdbaumechanik auf Bodenphysikalischer Grundlage, Deuticke.
2. The theory of one-dimensional consolidation of saturated clays10.1680/geot.1967.17.3.261 ↗
3. Analytical solutions of one-dimensional large strain consolidation of saturated and homogeneous clays10.1016/j.compgeo.2004.02.006 ↗
4. A non-linear theory of consolidation10.1680/geot.1965.15.2.161 ↗
5. Research of one-dimensional consolidation theory considering nonlinear characteristics of soil
6. A study on one-dimensional nonlinear consolidation of double-layered soil10.1016/S0266-352X(01)00017-9 ↗
7. The consolidation of saturated clay with variable permeability and compressibility10.1680/geot.1969.19.2.234 ↗
8. Analysis of one dimensional nonlinear consolidation of layered soils under time-dependent loading
9. Locally adaptive activation functions with slope recovery for deep and physics-informed neural networks
10. A transfer learning-physics informed neural network (TL-PINN) for vortex-induced vibration10.1016/j.oceaneng.2022.113101 ↗
11. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations10.1016/j.jcp.2018.10.045 ↗
12. Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., and Antiga, L. (2019). Pytorch: An imperative style, high-performance deep learning library. arXiv.
13. Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., Corrado, G.S., Davis, A., Dean, J., and Devin, M. (2016). Tensorflow: Large-scale machine learning on heterogeneous distributed systems. arXiv.
14. Lu, L., Meng, X., Mao, Z., and Karniadakis, G.E. (2019). Deepxde: A deep learning library for solving differential equations. arXiv.
15. Sciann: A keras/tensorflow wrapper for scientific computations and physics-informed deep learning using artificial neural networks10.1016/j.cma.2020.113552 ↗
16. Nsfnets (navier-stokes flow nets): Physics-informed neural networks for the incompressible navier-stokes equations10.1016/j.jcp.2020.109951 ↗
17. Physics-informed machine learning approach for augmenting turbulence models: A comprehensive framework10.1103/PhysRevFluids.3.074602 ↗
18. A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics10.1016/j.cma.2021.113741 ↗
19. A nonlocal physics-informed deep learning framework using the peridynamic differential operator10.1016/j.cma.2021.114012 ↗
20. Physics-informed deep learning for computational elastodynamics without labeled data10.1061/(ASCE)EM.1943-7889.0001947 ↗
21. Reconstruction of nanoscale particles from single-shot wide-angle free-electron-laser diffraction patterns with physics-informed neural networks10.1103/PhysRevE.103.053312 ↗
22. Operator learning for predicting multiscale bubble growth dynamics10.1063/5.0041203 ↗
23. Physics-informed neural networks for inverse problems in nano-optics and metamaterials10.1364/OE.384875 ↗
24. Deep physical informed neural networks for metamaterial design10.1109/ACCESS.2019.2963375 ↗
25. Physics-informed deep learning for one-dimensional consolidation10.1016/j.jrmge.2020.09.005 ↗
26. Lu, Y., and Mei, G. (2022). A deep learning approach for predicting two-dimensional soil consolidation using physics-informed neural networks (pinn). Mathematics, 10.10.3390/math10162949 ↗
27. Physics-informed deep neural networks for learning parameters and constitutive relationships in subsurface flow problems10.1029/2019WR026731 ↗
28. Forward and inverse modeling of water flow in unsaturated soils with discontinuous hydraulic conductivities using physics-informed neural networks with domain decomposition10.5194/hess-26-4469-2022 ↗
29. Application of improved physics-informed neural networks for nonlinear consolidation problems with continuous drainage boundary conditions10.1007/s11440-023-01899-0 ↗
30. Scientific machine learning through physics–informed neural networks: Where we are and what’s next10.1007/s10915-022-01939-z ↗
31. Modeling water flow in unsaturated soils through physics-informed neural network with principled loss function10.1016/j.compgeo.2023.105546 ↗
32. Terzaghi’s one-dimensional consolidation equation and its solution based on asymmetric continuous drainage boundary
33. Physics-informed neural network simulation of multiphase poroelasticity using stress-split sequential training10.1016/j.cma.2022.115141 ↗
34. Understanding the difficulty of training deep feedforward neural networks
35. Latin hypercube sampling and the propagation of uncertainty in analyses of complex systems10.1016/S0951-8320(03)00058-9 ↗
36. Kingma, D.P., and Ba, J. (2014). Adam: A method for stochastic optimization. arXiv.
37. Bischof, R., and Kraus, M. (2021). Multi-objective loss balancing for physics-informed deep learning. arXiv.
38. Understanding and mitigating gradient flow pathologies in physics-informed neural networks10.1137/20M1318043 ↗
39. Kendall, A., Gal, Y., and Cipolla, R. (2018, January 18–22). Multi-task learning using uncertainty to weigh losses for scene geometry and semantics. Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA.
40. Self-adaptive loss balanced physics-informed neural networks10.1016/j.neucom.2022.05.015 ↗
41. Madhyastha, P.S., and Batra, D. (2019, January 3–4). On model stability as a function of random seed. Proceedings of the Conference on Computational Natural Language Learning, Hong Kong, China.10.18653/v1/K19-1087 ↗
42. Bengio, Y. (2012). Practical recommendations for gradient-based training of deep architectures. Neural Networks, Springer.10.1007/978-3-642-35289-8_26 ↗
43. Conservative physics-informed neural networks on discrete domains for conservation laws: Applications to forward and inverse problems10.1016/j.cma.2020.113028 ↗
44. Theory of consolidation for clays10.1061/AJGEB6.0000075 ↗
45. An analytical solution for one-dimensional nonlinear consolidation of soils with continuous drainage boundary