If the duration of the flow period preceding a buildup test is not sufficiently long an apparent semi-log straight line may still be seen both on a type curve match and a Horner plot. The analysis of this semi-log straight line results in the wrong value of permeability but a reasonable p ∗. A longer shut-in does not correct the problem unless it is preceded by a longer flow period. The use of "equivalent time" to account for the short flow period introduces problems of its own.
In most oil or gas well test interpretations, it is important to determine if the data obtained are such that a meaningful semi-log analysis can be performed. In recent years, this has been done by "typecurve matching" the data and noting, from the typecurve, the approximate start of the semi-log data. Once this has been determined, the data following that point are analyzed by selecting a semi-logarithmic straight line, the slope of which can be used to determine the transmissibility of the reservoir.
The curves used for "typecurve matching" are, generally, derived from drawdown tests but they can be used for buildup tests if the data are first desuperposed. When rigorous desuperposition can not be done, an approximate method is often used, which is usually adequate for determining the approximate start of the semi-log straight line for use with Horner or similar plots.
In a large number of buildup tests recently analyzed by the authors, it was observed that, often, two tests on the same well gave significantly different results of interpretation. The method of interpretation was the same for both tests, namely: from the typecurve match, determine the approximate start of the semi-log straight line, and then use that in the Horner analysis. Both sets of desuperposed data showed the presence of an acceptable semi-log straight line. The principal difference between the two sets of tests was the duration of the flow period - one was significantly longer than the other. Why were the interpretations so different? and which one is the correct one?
In order to resolve the problem a theoretical investigation was undertaken. Pressure buildup datawere simulated for flow periods of varying durations and these synthetic pressures were then analyzed by the standard methods.
The reservoir model selected for this presentation is that of an infinite, homogeneous reservoir with a vertically fractured well - the fracture has an infinite conductivity, and wellbore storage is considered to be negligible. Such a model fits a very large number of wells which have been hydraulically fractured. The solution - in dimensionless terms - for this model was presented by Gringarten, Ramey and Raghavan in 1972 (ref 1) and is shown in "typecurve matching" format in Figure 1. This solution applies to a constant rate drawdown but it can be used, with the principle of superposition, to derive buildup information for any shut-in time Δt, following a flow period of duration t. (ref 2)
Equation (1) (Available In Full Paper)
