Modeling natural convection in heterogeneous porous media using encoder-decoder convolutional neural networks

Document Type : Research Article


1 student/tarbiat modares university

2 Faculty of Civil and Environmental Engineering. Tarbiat Modares University, Tehran, Iran


Introduction: Natural convection is an important phenomenon in porous media problems. It is encountered in a variety of applications, including in enhanced oil recovery systems and geothermal reservoirs. Physics-based numerical models are widely used to simulate natural convection in porous media. Although these models are usually effective, they commonly suffer from high computational costs. This is notably problematic in repetitive runs at large time and space scales, as in uncertainty analysis, data assimilation, and sensitivity analysis. In recent years, at least four different methods have been proposed to overcome this challenge, including optimizing the numerical solution algorithm, parallel computing, cloud computing, and data-driven methods. In most cases, while data-driven models are capable of handling low-dimensional problems, they have not been very successful in dealing with high-dimensional problems, both accurately and time efficient. To overcome these challenges, we propose using the encoder-decoder convolutional neural networks (ED-CNNs) for heterogeneous porous media. We apply the ED-CNN in the context of ‘image-to-image’ regression in the following two use cases in the context of natural convection simulations: (1) as a meta-model to estimate the heat map from the Rayleigh number distribution, and (2) as an optimizer to estimate the Rayleigh number distribution from the heat map.

Methodology: The proposed ED-CNN is employed to model the hypothetical example of a square porous enclosure filled with a saturated porous medium. The boundaries are impermeable, and temperatures at two opposite side walls are different, resulting in the formation of natural convection. Heterogeneity in the Rayleigh number across the problem domain is applied through zonation.
A numerical modeling tool is used to generate steady-state heat maps based on a number of randomly selected Rayleigh numbers. The numerical model input-outputs are transformed into square-shaped jpg images of 64 × 64 resolution. Two ED-CNNs are trained, one as a meta-model and the other as an optimizer. Different numbers of training input-output images (including 1000, 2000, 4000, and 5000) generated from the numerical model are employed to evaluate the performance of proposed networks. Two evaluation criteria are used to assess the performance of the developed ED-CNN models: (1) the root mean squared error (RMSE), and (2) the coefficient of determination (R^2-score). The ED-CNNs have been developed using Keras and Tensorflow python libraries.
Results and discussion: Results show that the ED-CNN accuracy, both as a meta-model and as an optimizer, is satisfactory. For the meta-model case (i.e. prediction of the temperature distribution from the Rayleigh map), the RMSE is mostly smaller than 0.15, and the R^2-score is around 0.92. In the case of ED-CNN as optimizer (i.e. estimation of the Rayleigh distribution from the heat map), RMSE is mostly in the interval [0.017-0.034], while the R^2-score is around 0.89. Acceptable results can be obtained using 2000 input-output image pairs and 150 epochs for the meta-model case, and 4000 image pairs and 200 epochs for the optimizer case. Analysis of the spatial distribution of errors shows that maximum errors occur in the middle of the problem domain where the heat map is least sensitive to the Rayleigh number. The ED-CNN model is also evaluated as an uncertainty analysis tool by comparing maps of mean and standard deviation based on the numerical model and ED-CNN predictions, showing a significant agreement with estimation error between them.
Conclusion: In this paper, we examine the performance of ED-CNNs, as a specialized architecture of deep neural networks, to solve the forward and inverse problems of natural convection in porous media. For this purpose, we frame the problem as one of image-to-image regression and show that the developed model is able to provide high accuracy approximations with limited training samples, effectively solving the curse of dimensionality problem associated with heterogeneous domains. In practice, the proposed methodology can be applied to image datasets obtained from not only numerical modeling, but also high-resolution imaging and non-destructive scanning techniques, to either estimate the temperature distribution due to natural convection, or to characterize the porous media based on the temperature distribution.


Arefin, A.M.E. (2016). Thermal analysis of modified pin fin heat sink for natural convection. In 2016 5th International Conference on Informatics, Electronics and Vision (ICIEV), pp. 1-5). IEEE.
Ataei-Dadavi, I., Chakkingal, M., Kenjeres, S., Kleijn, C.R. and Tummers, M.J. (2019). Flow and heat transfer measurements in natural convection in coarse-grained porous media. International Journal of Heat and Mass Transfer, 130, 575-584.
Balaji, C., Srinivasan, B. and Gedupudi, S. (2020). Heat transfer engineering: fundamentals and techniques. Academic Press.
Braester, C. and Vadasz, P. (1993). The effect of a weak heterogeneity of a porous medium on natural convection. Journal of Fluid Mechanics, 254, 345-362.
Fajraoui, N., Fahs, M., Younes, A. and Sudret, B. (2017). Analyzing natural convection in porous enclosure with polynomial chaos expansions: Effect of thermal dispersion, anisotropic permeability and heterogeneity. International Journal of Heat and Mass Transfer, 115, 205-224.
Gu, J., Liu, T., Wang, X., Wang, G., Cai, J. and Chen, T. (2018). Recent advances in convolutional neural networks, Pattern Recognit. 77, 354–377.
He, Q., Barajas-Solano, D., Tartakovsky, G. and Tartakovsky, A.M. (2020). Physics-informed neural networks for multiphysics data assimilation with application to subsurface transport. Advances in Water Resources, 141, 103610.
Hur, C. and Kang, S. (2019). Entropy-based pruning method for convolutional neural networks. The Journal of Supercomputing, 75(6), 2950-2963.
Ji, X., Yan, Q., Huang, D., Wu, B., Xu, X., Zhang, A., ... and Wu, M. (2021). Filtered selective search and evenly distributed convolutional neural networks for casting defects recognition. Journal of Materials Processing Technology, 292, 117064.
Jiang, Z., Tahmasebi, P. and Mao, Z. (2021). Deep residual U-net convolution neural networks with autoregressive strategy for fluid flow predictions in large-scale geosystems. Advances in Water Resources, 150, 103878.
Kamrava, S., Tahmasebi, P. and Sahimi, M. (2020). Linking morphology of porous media to their macroscopic permeability by deep learning. Transport in Porous Media, 131(2), 427-448.
Kreyenberg, P.J., Bauser, H.H. and Roth, K. (2019). Velocity Field Estimation on Density‐Driven Solute Transport With a Convolutional Neural Network. Water Resources Research, 55(8), 7275-7293.
Kumar, D., Roshni, T., Singh, A., Jha, M.K. and Samui, P. (2020). Predicting groundwater depth fluctuations using deep learning, extreme learning machine and Gaussian process: a comparative study. Earth Science Informatics, 13(4), 1237-1250.
Mo, S., Zhu, Y., Zabaras, N., Shi, X. and Wu, J. (2019). Deep convolutional encoder‐decoder networks for uncertainty quantification of dynamic multiphase flow in heterogeneous media. Water Resources Research, 55(1), 703-728.
Nield, D.A. and Bejan, A. (2017). Convection in Porous Media, Springer International Publishing, Cham.
Rajabi, M.M., Ataie-Ashtiani, B. and Simmons, C.T. (2015). Polynomial chaos expansions for uncertainty propagation and moment independent sensitivity analysis of seawater intrusion simulations. Journal of Hydrology, 520, 101-122.
Rajabi, M.M., Fahs, M., Panjehfouladgaran, A., Ataie-Ashtiani, B., Simmons, C.T. and Belfort, B. (2020). Uncertainty quantification and global sensitivity analysis of double-diffusive natural convection in a porous enclosure. International Journal of Heat and Mass Transfer, 162, 120291.
Rajabi, M.M., Javaran, M.R.H., Bah, A.O., Frey, G., Le Ber, F., Lehmann, F. and Fahs, M. (2022). Analyzing the efficiency and robustness of deep convolutional neural networks for modeling natural convection in heterogeneous porous media. International Journal of Heat and Mass Transfer, 183, 122131.
Ronneberger, O., Fischer, P. and Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. In International Conference on Medical image computing and computer-assisted intervention, Springer, Cham, 234-241.
Shen C. (2018). A transdisciplinary review of deep learning research and its relevance for water resources scientists, Water Resour. Res. 54(11), 8558–8593.
Soboleva, E.B. (2018). Density-driven convection in an inhomogeneous geothermal reservoir. International Journal of Heat and Mass Transfer, 127, 784-798.
Tahmasebi, P., Kamrava, S., Bai, T., Sahimi, M. (2020). Machine learning in geo-and environmental sciences: from small to large scale, Adv. Water Resour., 142, 103619. j.advwatres.2020.103619
Tartakovsky, A.M., Marrero, C.O., Perdikaris, P., Tartakovsky, G.D. and Barajas‐Solano, D. (2020). Physics‐informed deep neural networks for learning parameters and constitutive relationships in subsurface flow problems. Water Resources Research, 56(5), e2019WR026731.
Varol, Y., Oztop, H.F., and Avci, E. (2008). Estimation of thermal and flow fields due to natural convection using support vector machines (SVM) in a porous cavity with discrete heat sources, Int. Commun. Heat Mass Transf. 35(8), 928–936.
Vu, M.T. and Jardani, A. (2022). Mapping discrete fracture networks using inversion of hydraulic tomography data with convolutional neural network: SegNet-Fracture. Journal of Hydrology, 609, 127752, 2022.127752
Wei, H., Zhao, S., Rong, Q. and Bao, H. (2018). Predicting the effective thermal conductivities of composite materials and porous media by machine learning methods. International Journal of Heat and Mass Transfer, 127, 908-916.
Zhu, Y. and Zabaras, N. (2018). Bayesian deep convolutional encoder–decoder networks for surrogate modeling and uncertainty quantification. Journal of Computational Physics, 366, 415-447.
Volume 17, Issue 4 - Serial Number 174
December 2022
Pages 85-100
  • Receive Date: 30 January 2022
  • Revise Date: 06 August 2022
  • Accept Date: 15 August 2022
  • First Publish Date: 15 August 2022