A Numerical Simulator For Predicting Chemical Flood Performance

Abstract

Abstract INTERCOMP has developed a finite difference base numerical reservoir simulator for predicting micellar/ polymer flood performance. The model solves for n polymer flood performance. The model solves for n components in three fluid phases. All components may partition amongst the phases satisfying either partition amongst the phases satisfying either pseudo-ternary phase behavior or a general, multi-component, pseudo-ternary phase behavior or a general, multi-component, three-phase flash. Polymer characteristics such as inaccessible pore volume, resistance and residual resistance factors and retention hysteresis are included. Recovery mechanisms of swelling, solubilization and interfacial effects are represented. Second order spacial accuracy for the solution allows simulation of physical dispersion with a reasonable number of grid blocks. Aside from the usual 3-D Cartesian geometry, the model allows for a general orthogonal coordinate system as well as a special conformal map used for 5-spot symmetry elements. Two model formulations have been investigated. Capitalizing on an assumption of no access volume of mixing, the second model formulation runs 5 times faster than the first. The simulators to be described in this paper are the product of the "Chemical Flooding Model Development" project, undertaken by INTERCOMP for 13 major oil companies and research organizations. The paper will present the model formulations and solution methods together with examples demonstrating simulator capability. Introduction In April, 1974 INTERCOMP initiated a research project to study the application of variational project to study the application of variational methods to reservoir simulation. The project was jointly sponsored by INTERCOMP and ten major oil companies and industry research groups. The results of this project were of sufficient interest to encourage the proposal of an applications project. This led to the Chemical Flooding Model Development Project, begun in the Fall of 1975 and sponsored jointly by INTERCOMP and thirteen oil companies and research groups. There were several reasons why chemical flooding was chosen for the model development project. First, the work was timely. Interest in chemical flooding was increasing rapidly and a tool was needed both to aid in the interpretation of laboratory experiments and to scale up laboratory experiments for field applications. Second, the research project had indicated that there were certain features of the variational techniques which made them attractive for simulating the chemical flood process. In particular, the high order of accuracy readily obtained with the variational techniques could allow the tracking of small slugs of material from injector to producer without the non-natural dissipation of the slug due to the numerical dispersion usually exhibited by standard, first order correct finite-difference based simulators. Also, indications were that grid orientation effects would be essentially eliminated by use of the proper variational formulation. Paralleling the development of the variational based model was the development of a finite-difference based simulator. Although this simulator was developed principally to test the efficiency of the variational principally to test the efficiency of the variational model, it has itself proven to be an effective applications tool. The programs developed during this project were limited to two dimensions and two phases. Consequently, following the project completion in the Fall of 1977, a project extension was initiated with aim at enhancing the model physics and expanding the model's field simulation capabilities. The result was the development of two new finite-difference based simulators which, between themselves, differ fundamentally only in the treatment of the phase behavior. This paper discusses the development of the finite-difference based chemical flood simulators.