Abstract
Constant volume depletion (CVD) data are a primary tool in reservoir management of gas condensate and volatile oil reservoirs. Much time is spent tuning equations of state (EoS) to match lab data. If the lab data are invalid, that time has been wasted. Worse, predictions made with an EoS tuned to invalid lab data will lead to poor reservoir management and inaccurate forecasting. It is always important to avoid unnecessary expense, but especially so in the current economic environment.
In contrast to many laboratory tests, the CVD test includes a variety of different measurements, all of which must be internally consistent if the results are to be valid. In particular, the gas and liquid phases at each depletion step must be in equilibrium, and the data must preserve both overall mass balance and component-by-component molar balance.
Almost all published methods for quality-checking CVD data rely on some combination of the following: 1) comparing calculated and measured composition of the liquid remaining after the last depletion step to verify mass and/or mole balance; 2) calculating equilibrium ratios and preparing the Hoffman, Crump, Hocott plot to verify equilibrium; and 3) inspecting graphs of gas composition, liquid composition, and/or equilibrium ratios vs. pressure to ensure smooth behavior. Although not rigorous, as are the equilibrium and mass/mole balance criteria, there are a number of accurate correlations for various fluid properties that may be used in validating CVD test data.
As a result of this study:
A workflow is proposed to provide an intensive quality check for CVD data. The workflow uses a variety of tools to verify that: a) mass and mole balances are satisfied, b) gas and liquid are in equilibrium at each depletion step, and c) physical properties of the relevant phases are consistent with accurate correlations for liquid density, Z-factor, and bubble point pressure.
A new method is proposed for verifying the overall mass balance and component molar balance even when the composition of the liquid remaining after the final depletion step is not available.
Each of the correlations used in the workflow have been evaluated using a small database of CVD test data that have been screened for validity, to determine which correlations work best for each physical property.
New correlations are proposed for cumulative moles produced and for molecular weight of the C7+ fraction of the equilibrium gas.
A new method for setting a more stringent constraint on the density of the liquid phase is proposed.
Surprisingly, the presence of a straight line on the Hoffman, Crump, Hocott plot does not guarantee that the gas and liquid phases are in equilibrium as is commonly believed, nor does a straight line on a plot of log (ki) vs Tbi guarantee equilibrium.
