Application of Inverse Modeling to Partitioning Interwell Tracer Tests for Estimating Oil Saturation

The objectives of this thesis are learning and combining inverse model with method of moments (MOM) for obtaining accurate and realistic oil saturation estimates. The streamline-based inverse model developed in Texas A&M University coupled with the finite-difference ECLIPSE simulator was used to calculate the oil saturation distribution from partitioning interwell tracer test (PITT) data. Inverse model was tested for both single-phase at residual oil saturation and then for two-phase flow simulations during a waterflood. First sensitivity tests were performed. Then, the MOM and inverse modeling techniques were applied to PITTs at various tracer detection limits and their accuracy compared. Furthermore, these methods are used in a complementary way by using the saturation from the MOM as the initial guess for the inverse model calculation. In two-phase flow simulations, the sensitivity of inverse modeling to relative permeability, injection rate, cross-flow and to different partition coefficients was investigated.