Deconvolution is a recent milestone in well test analysis which converts multirate pressure data into a single drawdown at constant rate. It yields a pressure derivative with a duration equal to the duration of the test, thus providing more pressure data to interpret than with conventional techniques; and gives access to the true radius of investigation of the test. In addition, the deconvolved derivative is free from distortions due to the pressure derivative calculation algorithm and from errors introduced by incomplete or truncated rate histories. Theoretically, deconvolution is only valid for linear systems. In practice, it is also used, with pseudo-pressures, for dry gas, and even for gas condensate and volatile oil below saturation pressure. As these systems are highly non-linear, there has been some concerns about the validity of such an approach.
This paper investigates the use of deconvolution for gas condensate and volatile oil below saturation pressure. Using compositional simulation, it applies deconvolution to different reservoir geometries using two-phase pseudo-pressures, which linearize the reservoir-fluid system by replacing the two-phase fluid below saturation pressure by a single fluid equivalent. Comparison with deconvolution using single-phase pseudo-pressure (for gas condensate) or pressure (for volatile oil) shows that the resulting deconvolved derivatives are similar, and in particular exhibit similar late time behaviours, thus justifying the simpler pseudo-pressure (resp., pressure) approach. Simulation results were verified with actual data from vertical and horizontal wells in a number of gas condensate and volatile oil reservoirs.