A New GOR Forecasting Methodology for Unconventional Reservoirs

Abstract

The producing GOR trend of unconventional wells is a function of the complex interaction of fluid properties, geology, completion design, well spacing/stacking geometry, and operational changes. These complexities bring as an outcome a wide variation of GOR trends with time. This paper introduces a novel normalized forecasting process to build GOR type curves along with the uncertainty ranges. The long-term GOR range is comparable to theoretical PVT model expectations and offset wells with longer production history. The proposed methodology requires integrated multi-discipline efforts and is not a machine-learning modeling process. Groups of wells are selected based on similar geology and fluid properties while removing wells with poor data quality and operational issues. Next the normalized cumulative oil/EUR is introduced to compare the cumulative GOR curves. Empirical models are then generated to fit the GOR curves and forecast long-term GOR trends. A novel approach for terminal GOR range is applied to constrain the results from empirical models. The normalized process creates proven accurate gas predictions compared to the first three years of well performance. Two different approaches are used to validate the forecasted GOR. The first approach is to derive the terminal GOR range with the available PVT report. The second approach is to benchmark with offset wells with long history. Both approaches show the forecasted GOR is within the reasonable range. The normalized forecasting method has been tested for different areas in Delaware and Midland basins, covering a wide GOR range, and demonstrates its robustness in the forecast.