We simulate gravity segregation and sweep in foam EOR in miscible and immiscible displacements of oil to investigate the effect of oil on foam sweep and the usefulness of simple models designed to predict the extent of gravity override with foam. Comparison of simulation with analytical models highlights limitations of both approaches. In 1D simulations with fine grids, both first-contact and multicontact miscible displacements fit the fractional-flow model for miscible foam floods (Ashoori et al., 2010). In simulations of either sort of displacement, the "critical oil saturation" for foam stability is irrelevant because "gas" and "oil" never co-exist at the same location. Hydrocarbon kills foam abruptly at the point where the simulator changes the name of the nonaqueous phase from "gas" to "oil." In 2D homogeneous reservoirs, the model of gravity segregation of Stone (1982) and Jenkins (1984) must adapt to an inhomogeneous mixed zone for two reasons. First, if pressure varies greatly in the mixed zone, then foam mobility varies there, because in many foam models foam mobility is a function of gas viscosity. In reality, foam mobility is not sensitive to viscosity within bubbles. Second, residual oil saturation can vary in the mixed zone. For miscible displacements Stone and Jenkins' equations work well if one takes gas properties at the middle of the mixed zone. For immiscible displacements, the variation of residual oil saturation in the mixed zone can make the distance to segregation of gas shorter or longer than predicted by Stone and Jenkins depending on whether residual oil kills foam or not. "Modified SWAG" injection (injecting water simultaneously above gas) gave from 70 to 130% greater sweep than co-injection at the same injection rates in our simulations.
展开▼
