Question 1 of 8: Reservoir-Engineering Short Answers

Problem 1: Reservoir-Engineering Short Answers (25 marks — 5 marks each)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

1. Retrograde behavior. A retrograde gas-condensate fluid is one whose two-phase envelope is tilted far enough that the reservoir temperature falls between the critical temperature and the cricondentherm. Along an isothermal depletion path at such a temperature, crossing the dew-point line from the single-phase gas region does the opposite of what intuition from a simple pure substance suggests: retrograde condensation is liquid dropping out of solution as pressure falls below the dew point (normally you expect liquid to condense as pressure rises), because on this part of the phase envelope the dew-point locus bends back on itself — isothermal cooling of the pressure (not the temperature) pushes the system further into the two-phase region until a maximum liquid volume is reached, after which further depletion revaporizes the condensate. That re-vaporization leg is retrograde vaporization: liquid disappearing (turning back to gas) as pressure continues to fall, again the reverse of the normal vaporization-on-heating/depressurizing intuition. The key distinction is the direction of the liquid-volume curve versus pressure on the same isotherm: condensation (liquid volume rising) on the upper, high-pressure leg between the dew point and the maximum, and vaporization (liquid volume falling) on the lower leg between the maximum and abandonment pressure — both retrograde because both occur on a pressure decline, the opposite sense to non-retrograde single-component behavior.

2. Waterflooding an undersaturated reservoir. An undersaturated reservoir sits above its bubble point, so the only energy initially available for volumetric depletion is fluid and rock expansion — a small compressibility effect that supports very little primary recovery before pressure falls to the bubble point and a much more efficient solution-gas-drive (or worse, an inefficient one) mechanism takes over. Injecting water while the reservoir is still undersaturated keeps average pressure above the bubble point for as long as possible, which (a) avoids ever letting free gas evolve in the pore space, preserving the oil's own solution-gas drive energy for later and avoiding the very low sweep/high-mobility-ratio behavior of a free gas phase; (b) uses water, an incompressible and cheap injectant, as direct voidage replacement, which is far more volumetrically efficient than depending on rock/fluid expansion alone; and (c) with an early start, the flood front can sweep oil at a favorable water/oil mobility ratio while the reservoir oil viscosity is still at its lowest (undersaturated oil viscosity rises as pressure drops toward the bubble point, so keeping pressure high also keeps the oil easiest to displace). This is the classic case for "pressure maintenance" waterflooding, started before bubble point is ever reached.

3. Four independent straight-line methods for reserves determination. (i) Volumetric gas material balance — $p/z$ vs. cumulative gas produced $G_p$ plots as a straight line for a closed, volumetric dry-gas reservoir; extrapolating to $p/z=0$ gives the original gas in place $G$, and to the abandonment $p/z$ gives remaining reserves. (ii) Havlena–Odeh material balance for oil — casting the general oil MBE as $F=N(E_o+mE_g+E_{w,f})+W_eB_w$, a plot of $F$ vs. $E_o$ (no gas cap/water drive) or $F/E_o$ vs. $E_g/E_o$ (with a gas cap) is a straight line whose slope/intercept give $N$ (and $m$). (iii) Decline-curve analysis — an exponential decline plots as $\log q$ vs. $t$ (straight line, slope $-D/2.303$) and a harmonic decline plots as $\log q$ vs. $N_p$ (straight line); extrapolating either to the economic rate limit gives remaining reserves directly. (iv) Constant water-drive-index material balance — plotting $F/E_o$ vs. $W_e/E_o$ (or, equivalently, testing a candidate aquifer model until the water-drive index $F/(N E_o+W_e B_w)$ is constant with time) is a straight line of unit slope through the origin once the correct aquifer model and $N$ are found, and is itself used as the diagnostic for picking $N$.

4. Most generally applicable aquifer-influx method. The van Everdingen–Hurst unsteady-state (unsteady) model is the most generally applicable. Schilthuis's steady-state model assumes the aquifer instantaneously reaches a new equilibrium after every pressure change at the reservoir/aquifer boundary — only valid for a very active, high-permeability, effectively infinite aquifer, and it systematically over-predicts influx early in life. The Hurst "modified steady-state" approach patches this with an empirically-fitted time-lag but has no rigorous basis for extrapolation. Van Everdingen–Hurst instead solves the diffusivity equation for radial (or, with the Carter–Tracy or linear variants, other) aquifer geometries directly, produces the dimensionless influx function $W_{eD}(t_D)$ tabulated for any aquifer size and boundary condition (infinite, finite, no-flow outer boundary), and superposes the actual, arbitrarily-varying pressure history at the reservoir/aquifer boundary via convolution ($W_e=\sum U\,\Delta p\,W_{eD}$). Because it is derived from first principles (not an equilibrium assumption) and handles both early transient and late-time pseudosteady/steady flow as limiting cases of the same solution, it applies correctly across small, medium and large aquifers and across the whole life of the field — which is why it, or its Carter–Tracy computational variant, is the industry-standard method rather than Schilthuis or Hurst.

5. Reservoir simulation for reserves. Simulation is used, rather than decline-curve or material-balance methods, when the reservoir is too complex or too early in its life for those simpler tools to be reliable: (a) early life, before enough production/pressure history exists to define a decline trend or a material-balance straight line — a geologically-constrained static model plus a dynamic simulation is the only way to forecast recovery and reserves; (b) geological complexity — layered reservoirs with cross-flow, faulted or compartmentalized structures, strong permeability heterogeneity, or a gas cap/aquifer whose geometry material balance's single-tank assumption cannot represent; (c) multiple wells and changing operating strategy — infill drilling, workovers, artificial lift changes or completion changes through time, which a single-tank decline model cannot honor well-by-well; and (d) improved/enhanced recovery evaluation — waterflood pattern design, gas or chemical injection, thermal recovery — where the incremental reserves depend on a displacement process (sweep efficiency, fingering, gravity override) that only a full multi-phase, multi-dimensional flow model can predict. Once enough history exists, the model is history-matched against observed rates and pressures, then run forward under one or more development scenarios to book both proved and probable/possible reserves.