Stage-Wise Stochastic Deep Learning Inversion Framework for Subsurface Structure Identification

The codes for the paper "Stage-Wise Stochastic Deep Learning Inversion Framework for Subsurface Structure Identification"，the code of OCAAE and DOCRN is suitable for the 3D case, whereas it can be readily used for 2D subsurface structure by replacing the 3D operations with their 2D counterparts in network model files. such as OCAAE_models. As for the TP model, it can be used for both 2D and 3D cases, one should modify the "inputfile" and conditional file according to different problems, then run the "GEOST.exe" program to complete the simulation. The flow and transport responses data is too large to upload, one can generate this data set by TOUGHREACT or other transport simulators according to subsurface structures.