Biot's equations-based reservoir parameter inversion using deep neural networks

Xiong, Fuqin; Yong, Heng; Chen, Hua; Wang, Han; Shen, Weidong

doi:10.1093/jge/gxab057

Cited by 7 publications

(2 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hou et al [17], combined with D-InSAR monitoring technology, used inversion method to inverse IPIM-G dynamic prediction model parameters for mining subsidence prediction. Xiong et al [18], combined with DNN (deep neural network), analyzed the sensitivity of seismic attributes to rock mechanics parameters to realize parameter inversion. Liu et al [19], according to the fnite element theory, proposed a displacement inverse model for solving the elastic modulus of each rock layer of the underground roadway based on the measured displacement after excavation of the underground roadway in coal mines.…”

Section: Introductionmentioning

confidence: 99%

Inversion Study of Mechanical Parameters of Tunnel Surrounding Rock Based on Similar Test

Wang

Cui

et al. 2023

Advances in Civil Engineering

View full text Add to dashboard Cite

Based on similar model tests and finite element simulations, inversion study of the mechanical parameters of the tunnel surrounding rock by combining artificial fish swarm algorithm (AFSA) and digital scattering correlation method is carried out. The inversion of the mechanical parameters of the tunnel surrounding rock is carried out by iterating the optimization algorithm to minimize the objective function value through similar model test with the displacement measurements obtained by the digital scattering correlation method as the known quantity and the nodal displacement simulation values obtained by numerical simulation as the unknown quantity. The results show the following: (1) The new research idea proposed in this paper is effective and feasible and can perform the inversion calculation of the parameters of the soft and weak tunnel surrounding rock. (2) Through the inversion calculation, the parameters related to the surrounding rock show a wave-like change: the parameters increase in the initial and final compaction stages due to the small deformation rate of the tunnel and decrease in the continuous loading stage due to the increasing deformation rate of the tunnel. (3) The displacement field calculated by DSCM in the test is basically the same as the displacement distribution characteristics of the numerical simulation displacement cloud map, and the error of nodal displacement value is within 3%.

show abstract

Section: Introductionmentioning

confidence: 99%

Inversion Study of Mechanical Parameters of Tunnel Surrounding Rock Based on Similar Test

Wang

Cui

et al. 2023

Advances in Civil Engineering

View full text Add to dashboard Cite

show abstract

“…With the advent of the powerful computer and big data, deep neural networks have been extensively used in diverse tasks such as image classification, computer vision, language translation, game intelligence, reservoir parameter inversion [40]. In the context of scientific computing, the idea to use neural network for solving PDEs has been raised since 1990s [19,17].…”

Section: Introductionmentioning

confidence: 99%

A weighted first-order formulation for solving anisotropic diffusion equations with deep neural networks

Xie¹,

Zhai²,

Liu³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

In this paper, a new weighted first-order formulation is proposed for solving the anisotropic diffusion equations with deep neural networks. For many numerical schemes, the accurate approximation of anisotropic heat flux is crucial for the overall accuracy. In this work, the heat flux is firstly decomposed into two components along the two eigenvectors of the diffusion tensor, thus the anisotropic heat flux approximation is converted into the approximation of two isotropic components. Moreover, to handle the possible jump of the diffusion tensor across the interface, the weighted first-order formulation is obtained by multiplying this first-order formulation by a weighted function. By the decaying property of the weighted function, the weighted first-order formulation is always well-defined in the pointwise way. Finally, the weighted first-order formulation is solved with deep neural network approximation. Compared to the neural network approximation with the original second-order elliptic formulation, the proposed method can significantly improve the accuracy, especially for the discontinuous anisotropic diffusion problems.

show abstract

Deep-learning-based surrogate model for forward and inverse problems of wave propagation in porous media saturated with two fluids

Liu,

Xiong,

Guo

et al. 2023

Acta Geophys.

View full text Add to dashboard Cite

Biot's equations-based reservoir parameter inversion using deep neural networks

Cited by 7 publications

References 35 publications

Inversion Study of Mechanical Parameters of Tunnel Surrounding Rock Based on Similar Test

Inversion Study of Mechanical Parameters of Tunnel Surrounding Rock Based on Similar Test

A weighted first-order formulation for solving anisotropic diffusion equations with deep neural networks

Deep-learning-based surrogate model for forward and inverse problems of wave propagation in porous media saturated with two fluids

Contact Info

Product

Resources

About