Lawrence Berkeley National Laboratory, Report LBNL-40042

iTOUGH2 Sample Problems

Stefan Finsterle

Lawrence Berkeley National Laboratory, Earth Sciences Division
University of California, Berkeley, CA 94720


Summary. iTOUGH2 is a program for parameter estimation, sensitivity analysis, and uncertainty propagation analysis. It is based on the TOUGH2 simulator for non-isothermal multiphase flow in fractured and porous media [Pruess, 1987, 1991]. The report iTOUGH2 User's Guide describes the inverse modeling framework, and provides the theoretical background. The report iTOUGH2 Command Reference contains the syntax of all iTOUGH2 commands. The present report describes a variety of sample problems solved by iTOUGH2.

Each sample problem focuses on a few selected issues. iTOUGH2 input features and the usage of program options are described. Furthermore, interpretations of selected inverse modeling results are given. Problem No. 1 is a multipart tutorial, describing basic iTOUGH2 input files for the main iTOUGH2 application modes; no interpretation of results is given. Problem No. 2 focuses on non-uniqueness, residual analysis, and correlation structure. Problem No. 3 illustrates a variety of parameter and observation types, and describes parameter selection strategies. Problem No. 4 compares the performance of minimization algorithms and discusses model identification. Problem No. 5 explains how to set up a combined inversion of steady-state and transient data. Problem No. 6 provides a detailed residual and error analysis. Finally, Problem No. 7 illustrates the model character of the estimated parameters.

All input and data files needed to run the sample problems are part of the iTOUGH2 distribution package. The TOUGH2 and iTOUGH2 input file names are sam# and sam#i, respectively, where # stands for the problem number. Multipart problems have file names of the form sam#p%i, where # is the problem number and % is the part number. Data files are identified by a ".dat" file extension.

The key to a successful application of iTOUGH2 is (i) a good understanding of multiphase flow processes, (ii) the ability to conceptualize the given flow and transport problem and to develop a corresponding TOUGH2 model, (iii) detailed knowledge about the data used for calibration, (iv) an understanding of parameter estimation theory and the correct interpretation of inverse modeling results, (v) proficiency in using iTOUGH2 options. This report addresses issues (iv) and (v).