Fitting Foam-Simulation-Model Parameters to Data: I. Coinjection of Gas and Liquid

Abstract

Summary Cheng et al. (2000) present a simple method to fit foam-simulation parameters without oil to data for pressure gradient as a function of superficial velocities of gas and liquid. The key in this process is the identification of “high-quality” (high gas fraction) and “low-quality” foam regimes. The method is essentially the same for the foam-model parameters in foam models in STARS (Cheng et al. 2000), UTCHEM (Cheng et al. 2000), or ECLIPSE (Schlumberger 2010). Often, however, available data are more limited: pressure gradient for one scan of foam quality at fixed total superficial velocity. We show how to extend this method to the more-limited data set. The transition in regimes occurs at the foam quality with the maximum-pressure gradient. We illustrate the method by fitting parameters to several published data sets. Our approach is simple and direct. The model fit would be appropriate for an enhanced-oil-recovery process involving foam injection at finite water fraction, but not a surface-alternating-gas foam process involving large slugs of gas and liquid. For the latter process, the model fit should focus on data relevant to that process (i.e., at extremely high foam quality). The approach assumes an abrupt transition between high- and low-quality-foam regimes (e.g., a large value of epdry in the STARS foam model). If the transition is less abrupt, the parameter values quickly obtained by this method could provide the initial guess for a computer-based least-squares fit of all parameters, including a smaller value of epdry, and a check on the parameters so obtained.