A robust estimation of the Expected Ultimate Recovery (EUR) for fracture stimulated multi-layered tight gas wells, early in the production life of the wells, is the cornerstone for a meaningful project economic analysis, reserves classification and future field development plan, however, it actually presents significant challenges to reservoir engineers.
The complex flow dynamic of a tight gas reservoir causes that the direct application of the most commonly used reservoir engineering techniques to characterize Original Gas In Place (OGIP) such as Static Material Balance (P/Z vs Gp plot) and Flowing Material Balance (FMB), and EUR estimation such as Decline Curve Analysis (DCA) can only provide reliable results within an acceptable level of uncertainty when significant production decline data is available (minimum 6 to 9 years of production history in the case of Big Lake wells). This makes early predictions of future production decline behavior hence EUR highly uncertain, sometimes even unrealistic and very subjective to the analyst, with significant potential impacts.
The objective of this work is to present a practical workflow that provides suitable constraints to be applied to the otherwise poorly constrained DCA technique for EUR estimation by making appropriate adjustments to the results obtained from the standard reservoir engineering techniques when applied to the early production decline data of multi-layered tight gas wells, based on relevant correlations from a detailed production and reservoir analysis of numerous field cases from the Big Lake Field in the Cooper Basin.
Specifically, the study illustrates: (1) that the OGIP from P/Z determined early in the production life of a well, although using imperfect Build-Up pressures (static pressure values taken at the well which are still building pressure due to reservoir characteristics and sampling period), can be reasonably corrected to better represent the long term connected OGIP to the well, through a correlation between P/Z and FMB analysis on existing wells. (2) The integration of Production Logging Tests (PLT) data and petrophysical properties provides important inputs in order to better characterize the vertical heterogeneity of the reservoirs and simulate the aggregate Recovery Factor (RF) in a multi-layer production scenario. (3) A more constrained multi-segment DCA, constructed honoring the results from point 1 and 2 better represents (within 6% deviation) the production forecast of over 40 wells in the Big Lake Field after a year and half of additional production and it is expected to be a close approximation of the EUR of the wells (assuming current operating conditions).