Rate-transient analysis (RTA) is routinely used for unconventional production data analysis and forecasting. RTA uses rate normalization along with material balance time to estimate:(a) flow regimes, (b) the time of end of transient flow, and (c) the drainage volume and to thereby predict the estimated ultimate recovery (EUR) of wells. However, rate normalization approximates deconvolution, and material balance time is only strictly applicable for constant-pressure or constant-rate systems. This work investigates the validity of rate normalization and material balance time for synthetic and tight-oil well examples producing under variable bottomhole flowing pressure (BHP) conditions.
This work generates synthetic examples with different BHP histories and uses deconvolution to estimate a constant-pressure rate/material balance time response. First, we plot and compare the normalized rate vs. material balance time, the deconvolved rate/material balance time and the unit-pressure-drop rate vs. constant-pressure material balance time. The latter plot represents the constant-pressure solution of the 1D flow of a slightly compressible fluid (reference case). Second, we evaluate the plots of material balance time vs. time, deconvolved material balance time vs. time and constant-pressure material balance time vs. time. Third, we fit the reference type curve to a plot of the normalized rate vs. material balance time and the deconvolved rate vs. deconvolved material balance time to determine the reservoir properties to then estimate the EUR using time superposition. We conclude by illustrating the application of these steps to tight-oil wells in which we use deconvolution to estimate the unit-pressure-drop rate and the constant-pressure material balance time.
The results of this study are the following. First, BHP changes alter the slope of the log-log normalized rate-vs. material balance time plot. Second, BHP variations introduce error to the behavior of the material balance time vs. time function leading to incorrect estimates of the time of end of transient flow. Consequently, normalized-pressure rate and material balance time are not always reliable variables to properly identify the flow regime(s) and thus, to correctly estimate the time of end of transient flow and EUR. Alternatively, applying deconvolution to rigorously account for the pressure variations and generate the unit-pressure-drop rate and the constant-pressure material balance time solves these problems.
This paper investigates the validity of rate normalization and material balance time in RTA of unconventional reservoirs. Caution is needed when applying rate normalization and material balance time since these might lead to incorrect estimates of the time of end of transient flow and EUR. For this reason, deconvolution should be included as an integral part of the RTA workflow.