Physical-Based Inversion for Subsurface Flow and Transport Modeling

Wednesday, 17 December 2014
Ye Zhang, Univ of Wyoming, Laramie, WY, United States, Jianying Jiao, University of Wyoming, Laramie, WY, United States, Dongdong Wang, University of North Carolina at Chapel Hill, Geological Sciences, Chapel Hill, NC, United States and Juraj Irsa, Schlumberger, Houston, TX, United States
A new and computationally efficient fluid flow and transport inverse theory has been developed for characterizing, calibrating, and modeling aquifers. The theory is capable of simultaneous estimation of model boundary conditions (for simple transient problems, also the initial conditions) and fluid flow and transport parameters, i.e., spatially distributed permeabilities, source/sink rates, storativity, and dispersivity. The theory is robust to measurement errors and strong parameter variability. Effective parameters can be estimated to represent unresolved heterogeneity, e.g., sub-grid features and spatially variable recharge. The theory has been extended to new problems including parameter structure identification, unsaturated and variably saturated flows (e.g., directly estimating the soil retention functions), joint flow and transport inversion (e.g., containment source identification), uncertainty analysis (e.g., integrating subsurface static and dynamic data via geostatistical inversion), and high performance computing (e.g., solving large inversion systems with parallel computing). This presentation will summarize the body of the inversion research and discuss new directions for future work.