Uncertainty Assessment Using Experimental Design and Risk Analysis Techniques, Applied to Offshore Heavy-Oil Recovery

Abstract

Abstract The increasing complexity in the development of heavy oil fields is directly associated with the uncertainties in the fluid and reservoir characterization, particularly, in offshore scenarios where difficulties for well testing and fluid and core sampling are present. A probabilistic analysis, instead of a deterministic one is the natural way to face these expected uncertainties. The approach proposed in this work uses experimental design techniques to determine the parameters that have large contribution into the Net Present Value (NPV) of the prospect being analyzed. In the example presented NPV is estimated based on the accumulated oil production response of a flow simulator over a period of 30 years. An uncertainty analysis was done using information about the probabilities of the uncertain parameters. The decision tree technique was used to map all possible outcomes and then to estimate the Expected Monetary Value. A program was written to manage the input/output files of a commercial black-oil reservoir simulator in order to run a total of 1,728 simulations and estimate the NPV for each one of those parameter combinations. In a second analysis, the uncertainty density distribution was derived based on the histogram of the NPV results, assuming that the values of the uncertain parameters cover the entire range of variability. This procedure was applied to a synthetic case in which the uncertain parameters selected were: porosity, absolute and relative permeability, area of accumulation and exploitation scheme (horizontal or vertical wells).