Ravi Sharma's thesis
by
Ravi Sharma, MSE
University of Texas at Austin, 2003
Supervisors: Gary A. Pope
Kamy Sepehrnoori
The objective of this study was to simulate gas condensate reservoirs for understanding the effect of liquid
blocking near the well and the effect of methanol in improving the productivity of a damaged well. A new
hybrid well model was also developed and implemented in the University of Texas Composition Simulator
(UTCOMP) to calculate the well productivity.
In a gas condensate well, as the bottom hole pressure falls below the dew point pressure, a liquid phase drops out and accumulates in the pore spaces until residual oil saturation is reached. To model this condensate dropout effect accurately, phase behavior modeling of the hydrocarbon mixture was done to match the pressure volume relationships obtained from the experiments. The presence of water in the gas condensate reservoirs made it necessary to study the effect of water on the phase behavior of the mixture. Phase behavior studies of hydrocarbon-methanol-water mixtures were also done to aid in the interpretation of increase in gas relative permeability after methanol treatment as shown by experiments. The results showed that for hydrocarbon mixtures the PREOS calculates the phase behavior well with zero binary interaction coefficients but for mixtures with water and methanol, binary interaction coefficients must be adjusted to get a better match. For mixtures with polar components, the Soave-Redlich-Kwong equation-of-state with Huron and Vidal mixing rules performs better than the PREOS.
A simulation study of a gas condensate well under conditions similar to a damaged well at Hatter’s Pond Field was done to understand the impact of water and condensate blocking. An effort was also made to understand the response of the damaged well to methanol treatment. A considerable decline in the productivity due to liquid blockage was observed. The sharp decline showed that the condensate accumulates within a few days after starting production below the dew point pressure. It was also observed that the trapping number has significant effect on well deliverability. Production rates with trapping number effect were 20-30 percent higher than those without trapping number effect. Methanol treatment showed an increase in production for a few days after treatment.
For accurate prediction of the well productivity, it is necessary to have fine grids near the well while conducting simulation studies. A new hybrid well model was developed and implemented in the UTCOMP providing a faster and easier alternative to conventional fine grid simulation without compromising the accuracy. In this well model the area near the well is modeled separately using variable width grid option available in UTCOMP and is coupled with the uniform coarse grid model of the reservoir. Two-dimensional single layer, single well simulations were done for the verification of the new model against conventional fine grid simulations. Three different fluids were used for the study that varied from lean to rich gas condensate mixtures. The results obtained were found to be in good agreement with the conventional fine grid simulation results. Also, the new well model was about three times faster than fine grid simulation.
A simplified semi-analytical well model was implemented in the UTCOMP to calculate the well productivity. Three gas condensate mixtures were used to test this model and results were found to be in close agreement with the fine grid simulation results.
Back to theses index
