Effect of fluid properties models in the prediction of bottom-hole flowing pressures for multiphase systems
Academic Article
Overview
Research
Additional Document Info
View All
Overview
abstract
Simultaneous flow of oil, water and gas in pipes occurs throughout the production systems involved in flowing fluids from the reservoir to the surface. Accurate prediction of the pressure drop in a multiphase flow system is essential for proper design of well completions, artificial-lift systems, surface flowlines and gathering lines. The interdependence of flow regimes, flow rates of the different fluid phases, and of fluid properties, makes the prediction of pressure drop in a multiphase system a complex and iterative procedure. The thermodynamic behavior of the flowing hydrocarbon mixture is one of the most dominant factors affecting multiphase flow. All pressure drop models for multicomponent multiphase flow require the knowledge of mass transfer, or species partitioning, between the vapor and liquid phases and the physical properties of the different phases. Regardless of the procedure used to predict the bottomhole flowing pressure (BHFP), it is vital to have reliable predictions of the fractions of liquid and vapor flowing at a certain point in the well, and of the densities and viscosities of those phases. We have developed a numerical simulator to predict the BHFP in multiphase systems using the Beggs and Brill (1973) model which is one of the most widely used and accepted in the petroleum industry. This simulator has two alternative procedures to predict the phase separations along the tube and the fluid phase properties. One is the conventional approach by using fluid property correlations, and the second is by a thermodynamic Equation of State (EOS). The BHFPs predicted from both procedures were compared with measured values of BHFP data from 56 wells with a wide variation in producing rates, gas/oil ratios, depths, tubing sizes, compositions, and water cuts. The average absolute mean predicted error for our procedure was 9.73% with a standard deviation of 7.55%, while the conventional approach gave a 28.44% error with a standard deviation of 21.79%. We believe that whenever compositional data are provided the evaluation of phase separation and fluid properties using a thermodynamic EOS should always provide better estimates of the BHFP than generalized fluid property correlations. For volatile and condensate fluid systems fluid property correlations would introduce even larger deviations.
