no code implementations • 30 Apr 2022 • Tianhao He, Haibin Chang, Dongxiao Zhang
Furthermore, based on the constructed TgU-net surrogate, a data assimilation method is employed to identify the physical process and parameters simultaneously.
no code implementations • 15 Apr 2022 • Nanzhe Wang, Haibin Chang, Xiangzhao Kong, Martin O. Saar, Dongxiao Zhang
In this work, we propose a closed-loop optimization framework, based on deep learning surrogates, for the well control optimization of geothermal reservoirs.
no code implementations • 31 Dec 2021 • Nanzhe Wang, Qinzhuo Liao, Haibin Chang, Dongxiao Zhang
The results show that the deep learning method can provide equivalent upscaling accuracy to the numerical method, and efficiency can be improved significantly compared to numerical upscaling.
no code implementations • 12 Oct 2021 • Nanzhe Wang, Haibin Chang, Dongxiao Zhang
Pressure and saturation are coupled with each other in the governing equations, and thus the two networks are also mutually conditioned in the training process by the discretized governing equations, which also increases the difficulty of model training.
no code implementations • 17 Nov 2020 • Nanzhe Wang, Haibin Chang, Dongxiao Zhang
In order to achieve the theory-guided training, the governing equations of the studied problems can be discretized and the finite difference scheme of the equations can be embedded into the training of CNN.
no code implementations • 25 Apr 2020 • Nanzhe Wang, Haibin Chang, Dongxiao Zhang
The trained neural network can predict solutions of subsurface flow problems with new stochastic parameters.
no code implementations • 21 Jan 2020 • Hao Xu, Haibin Chang, Dongxiao Zhang
In the proposed framework, a deep neural network that is trained with available data of a physical problem is utilized to generate meta-data and calculate derivatives, and the genetic algorithm is then employed to discover the underlying PDE.
no code implementations • 24 Oct 2019 • Nanzhe Wang, Dongxiao Zhang, Haibin Chang, Heng Li
The TgNN can achieve higher accuracy than the ordinary Artificial Neural Network (ANN) because the former provides physically feasible predictions and can be more readily generalized beyond the regimes covered with the training data.
no code implementations • 13 Aug 2019 • Hao Xu, Haibin Chang, Dongxiao Zhang
However, prior to achieving this goal, major challenges remain to be resolved, including learning PDE under noisy data and limited discrete data.
no code implementations • 29 Oct 2018 • Haibin Chang, Dongxiao Zhang
Using the training data set, a data-driven method is developed to learn the governing equation of the considered physical problem by identifying the occurred (or dominated) processes and selecting the proper empirical model.