Stable and accurate reconstruction of pollutant transport is a crucial and challenging problem, including the inverse problem of identifying pollution sources and physical coefficients and the forward problem of inferring pollutant transport. Governed by advection, diffusion, and reaction processes, this transport phenomenon can be represented by the advection–diffusion–reaction (ADR) equation. In this paper, the physics-informed neural networks (PINNs) are applied to solve the forward and inverse ADR problems. To further enhance the stability and accuracy of the original PINN, two improvements are developed. The first adjusts the orthogonal grid (OG) point selection method and the other suggests adding an additional regulation function, namely, first derivative constraint (FDC). The new method is referred to as OG-PINN with FDC. To verify the effectiveness of the proposed method, five forward and inverse ADR problems are solved, and the results are compared with the analytical and reference solutions. For forward problems, the improved method can solve various ADR problems accurately and stably. For inverse problems, the ability of the OG-PINN for model parameter learning and initial distribution prediction is demonstrated and analyzed. The former gives the missed physical information in the ADR equation from the data, and the latter is used to trace the source of pollutants. The proposed method is quantitatively reliable for investigating various advection–diffusion–reaction processes.
References
1.
E.
Essouayed
, E.
Verardo
, A.
Pryet
, R. L.
Chassagne
, and O.
Atteia
, “An iterative strategy for contaminant source localisation using GLMA optimization and Data Worth on two synthetic 2D Aquifers
,” J. Contam. Hydrol.
228
, 103554
(2020
).2.
M.
Barati Moghaddam
, M.
Mazaheri
, and J.
Mohammad Vali Samani
, “Inverse modeling of contaminant transport for pollution source identification in surface and groundwaters: A review
,” Groundwater Sustainable Dev.
15
, 100651
(2021
).3.
L. A.
Khan
and P. L.
Liu
, “An operator splitting algorithm for the three-dimensional advection-diffusion equation
,” Int. J. Numer. Methods Fluids
28
, 461
–476
(1998
).4.
Y.
Wang
and K.
Hutter
, “Comparisons of numerical methods with respect to convectively dominated problems
,” Int. J. Numer. Methods Fluids
37
, 721
–745
(2001
).5.
K.
Alhumaizi
, “Comparison of finite difference methods for the numerical simulation of reacting flow
,” Comput. Chem. Eng.
28
, 1759
–1769
(2004
).6.
D. L.
Stefanovic
and H. G.
Stefan
, “Accurate two-dimensional simulation of advective-diffusive-reactive transport
,” J. Hydraul. Eng.
127
, 728
–737
(2001
).7.
B. H.
Devkota
and J.
Imberger
, “Lagrangian modeling of advection-diffusion transport in open channel flow
,” Water Resour. Res.
45
, 1
–14
, (2009
).8.
N.
Karedla
, J. C.
Thiele
, I.
Gregor
, and J.
Enderlein
, “Efficient solver for a special class of convection-diffusion problems
,” Phys. Fluids
31
, 023606
(2019
).9.
E.
Heifetz
and S.
Zucker
, “Fluid-like representation of Fickian diffusion
,” Phys. Fluids
34
, 011701
(2022
).10.
E. M.
LaBolle
, J.
Quastel
, and G. E.
Fogg
, “Diffusion theory for transport in porous media: Transition-probability densities of diffusion processes corresponding to advection-dispersion equations
,” Water Resour. Res.
34
, 1685
–1693
, (1998
).11.
Y.
Sun
, A. S.
Jayaraman
, and G. S.
Chirikjian
, “Approximate solutions of the advection–diffusion equation for spatially variable flows
,” Phys. Fluids
34
, 033318
(2022
).12.
S.
Leichombam
and R. K.
Bhattacharjya
, “Identification of unknown groundwater pollution sources and determination of optimal well locations using ANN-GA based simulation-optimization model
,” J. Water Resour. Prot.
08
, 411
–424
(2016
).13.
S.
Zhang
and X.
Xin
, “Pollutant source identification model for water pollution incidents in small straight rivers based on genetic algorithm
,” Appl. Water Sci.
7
, 1955
–1963
(2017
).14.
G.
Sole-Mari
, D.
Fernàndez-Garcia
, P.
Rodríguez-Escales
, and X.
Sanchez-Vila
, “A KDE-based random walk method for modeling reactive transport with complex kinetics in porous media
,” Water Resour. Res.
53
, 9019
–9039
, (2017
).15.
H. T.
Hwang
, S. W.
Jeen
, D.
Kaown
, S. S.
Lee
, E. A.
Sudicky
, D. T.
Steinmoeller
, and K. K.
Lee
, “Backward probability model for identifying multiple contaminant source zones under transient variably saturated flow conditions
,” Water Resour. Res.
56
, 1
–20
, (2020
).16.
F.
Cupola
, M. G.
Tanda
, and A.
Zanini
, “Contaminant release history identification in 2-D heterogeneous aquifers through a minimum relative entropy approach
,” SpringerPlus
4
, 1
–19
(2015
).17.
K.
Hornik
, M.
Stinchcombe
, and H.
White
, “Multilayer feedforward networks are universal approximators
,” Neural Networks
2
, 359
–366
(1989
).18.
L.
Lyu
, K.
Wu
, R.
Du
, and J.
Chen
, “Enforcing exact boundary and initial conditions in the deep mixed residual method
,” arXiv:2008.01491 (2020
).19.
B.
Liu
, J.
Tang
, H.
Huang
, and X.
Lu
, “Deep learning methods for super-resolution reconstruction of turbulent flows
,” Phys. Fluids
32
, 025105
(2020
).20.
Z.
Long
, Y.
Lu
, and B.
Dong
, “PDE-Net 2.0: Learning PDEs from data with a numeric-symbolic hybrid deep network
,” J. Comput. Phys.
399
, 108925
(2019
).21.
G. E.
Karniadakis
, I. G.
Kevrekidis
, L.
Lu
, P.
Perdikaris
, S.
Wang
, and L.
Yang
, “Physics-informed machine learning
,” Nat. Rev. Phys.
3
, 422
–440
(2021
).22.
M.
Raissi
, P.
Perdikaris
, and G. E.
Karniadakis
, “Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
,” J. Comput. Phys.
378
, 686
–707
(2019
).23.
L.
Lu
, X.
Meng
, Z.
Mao
, and G. E.
Karniadakis
, “DeepXDE: A deep learning library for solving differential equations
,” SIAM Rev.
63
, 208
–228
(2021
).24.
M.
Raissi
, A.
Yazdani
, and G. E.
Karniadakis
, “Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
,” Science
367
, 1026
–1030
(2020
).25.
Y.
Li
and F.
Mei
, “Deep learning-based method coupled with small sample learning for solving partial differential equations
,” Multimedia Tools Appl.
80
, 17391
–17413
(2021
).26.
Z.
Zhou
and Z.
Yan
, “Solving forward and inverse problems of the logarithmic nonlinear Schrödinger equation with PT-symmetric harmonic potential via deep learning
,” Phys. Lett. A
387
, 127010
(2021
).27.
S.
Cai
, Z.
Wang
, S.
Wang
, P.
Perdikaris
, and G. E.
Karniadakis
, “Physics-informed neural networks for heat transfer problems
,” J. Heat Transfer
143
, 1
–15
(2021
).28.
S.
Wang
, X.
Yu
, and P.
Perdikaris
, “When and why PINNs fail to train: A neural tangent kernel perspective
,” J. Comput. Phys.
449
, 110768
(2022
).29.
A. D.
Jagtap
and G. E.
Karniadakis
, “Extended physics-informed neural networks (XPINNs): A generalized space-time domain decomposition based deep learning framework for nonlinear partial differential equations
,” Commun. Comput. Phys.
28
, 2002
–2041
(2020
).30.
A. D.
Jagtap
, K.
Kawaguchi
, and G. E.
Karniadakis
, “Adaptive activation functions accelerate convergence in deep and physics-informed neural networks
,” J. Comput. Phys.
404
, 109136
(2020
).31.
Z.
Li
, X. Z.
Mao
, T. S.
Li
, and S.
Zhang
, “Estimation of river pollution source using the space-time radial basis collocation method
,” Adv. Water Resour.
88
, 68
–79
(2016
).32.
W. P.
Cheng
and Y.
Jia
, “Identification of contaminant point source in surface waters based on backward location probability density function method
,” Adv. Water Resour.
33
, 397
–410
(2010
).33.
M.
Stein
, “Large sample properties of simulations using Latin hypercube sampling
,” Technometrics
29
, 143
–151
(1987
).34.
D. P.
Kingma
and J. L.
Ba
, “Adam: A method for stochastic optimization
,” in 3rd International Conference on Learning Representations, ICLR 2015-Conference Track Proceedings
(IEEE
, 2015
), pp. 1
–15
.35.
J. H.
Ten Thije Boonkkamp
and M. J.
Anthonissen
, “The finite volume-complete flux scheme for advection-diffusion-reaction equations
,” J. Sci. Comput.
46
, 47
–70
(2011
).© 2022 Author(s). Published under an exclusive license by AIP Publishing.
2022
Author(s)
You do not currently have access to this content.