Deep Conditional Generative Adversarial Network Combined With Data‐Space Inversion for Estimation of High‐Dimensional Uncertain Geological Parameters. Issue 3 (16th March 2023)

• Record Type: Journal Article
• Authors: Fu, Wenhao; Zhang, Kai; Ma, Xiaopeng; Liu, Piyang; Zhang, Liming; Yan, Xia; Yang, Yongfei; Sun, Hai; Yao, Jun
• Is Part Of: Water resources research. Volume 59:Issue 3(2023)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2023-03-16
• Subjects: inverse modeling -- generative adversarial network -- data‐space iInversion -- deep learning; Hydrology -- Periodicals; 333.91
• DOI: 10.1029/2022WR032553
• Languages: English
• ISSNs: 0043-1397

Abstract

Abstract: Inverse modeling can provide a reliable geological model for subsurface flow numerical simulation, which is a challenging issue that requires calibration of the uncertain parameters of the geological model to establish an acceptable match between simulation data and observation data. The general inverse modeling method needs to iteratively adjust the uncertain parameters, which is a difficult and time‐consuming high‐dimensional sampling problem. To address this problem, we propose a deep‐learning‐based inverse modeling method called pix2pixGAN‐DSI. In this method, the deep‐learning‐based image‐to‐image generative adversarial network (pix2pixGAN) is constructed to directly predict the posterior parameter fields from the posterior dynamic responses obtained by the data‐space inversion (DSI) method. This inverse modeling method does not need to iteratively adjust the uncertain parameters, which improves computational efficiency. The effectiveness of the proposed method is verified through a Gaussian model case and two non‐Gaussian channelized model cases. Through the analysis of posterior realizations, matching and forecast of production data, and uncertainty quantification, the results show that the proposed method can obtain reasonable estimates without iteration and parameterization. Plain Language Summary: Numerical simulation is an effective means to characterize the flow of subsurface fluid and achieve efficient development of subsurface resources. Inverse modeling can calibrate the uncertain parameters of the numerical model and thus ensure simulation accuracy. However, most inverse modeling methods are based on iteratively adjusting the uncertain parameters, which requires performing a great deal of time‐consuming numerical simulations. In the last decades, significant progress has been made in the field of machine learning, especially in the field of deep learning. In this study, we use the generative adversarial network (a key achievement of deep learning) in combination with the data‐space inversion method to establish a novel inverse modeling workflow. The effectiveness of the proposed method is validated by three case analyses. Key Points: A deep conditional generative adversarial network is developed to generate posterior parameter fields from the posterior dynamic responses The data‐space inversion method is implemented to sample the posterior dynamic responses from prior model simulations and observed data The proposed method is end‐to‐end, avoids parameterization, and can deal with complex geological parameter inversion problems