Incorporation of Two-Phase Production Data into Reservoir Characterization

Abstract

In this study, two-phase production history data is incorporated into reservoir characterization by inverse modeling. As a first step, an analytical method has been developed to compute sensitivity coefficients for two-phase flow conditions. These equations are independent of flow simulators, and can be incorporated to any finite difference or streamline simulators. They only require one simulation run to compute the sensitivities for all time steps. In this study, we have chosen streamline simulation, which is order of magnitude faster than finite difference simulations. In the second step, a dual loop iteration technique is employed which uses conjugate gradient relaxation in order to bypass the explicit construction of the sensitivity coefficient matrix and the inversion of the product of sensitivity coefficient matrix and its transpose. Thus, a substantial reduction in the computational cost for solving the inverse problem is achieved, including the help of streamline simulation. Finally, this technique was validated using synthetic as well as field cases. The field case study was conducted on a sandstone reservoir. P. 151

Development and Application of Inverse Model Using Genetic Algorithm for Reservoir Characterization

“…Ates et al (1998) and Wen et al (1998) modified an existing optimization algorithm developed for single-phase reservoir characterization, so as it could be apply to multi-phase reservoirs, and then applied it to a field characterization. Barman et al (2000) and Ouence (2000) examined the relationship between the fracture and the parameters that affected it, using an artificial neural network for the characterization of the fracture reservoir.…”