PRODUCTION FORECASTING FOR SHALE OIL: WORKFLOW

With the demand for oil rising, unconventional oil reservoirs have taken a prominent role in the United States as a source of crude oil. Different methodologies to estimate reserves for shale gas and coal bed methane have, thus far, proved to be reliable, but no simple yet accurate workflow has been generally accepted to forecast production and estimate reserves for shale oil. To fill this gap in technology, we proposed and validated a workflow that integrates analytical methods with empirical methods. The final methodology is both easily applied and accurate. In developing the final workflow, we evaluated several alternatives, most of which proved to be unsuitable. We also investigated the use of a filter to eliminate outliers in a systematic way, as proposed by Rastogi (2014). The workflow was successfully applied to three of four volatile oil wells in the Eagle Ford shale, with similar results. The analytical model that best matched the wells is called the Stimulated Reservoir Volume (SRV) Bounded Model by the software marketer Kappa. We tested this and other models using a Beta test version of new Kappa software. While accurate, this modeling approach is too time consuming for routine use. We found that a simple empirical approach that led to the same results as the analytical model was a 3-segment Arps decline model. The early flow regime was transient linear for all the wells; thus an Arps ―b‖ parameter of two was appropriate. When boundary-influenced flow (BIF) appeared later, b-values of 0.2 were found appropriate. The initial decline rate (Di) value during BIF was modified in mid-segment leading to a distinct third segment. Our workflow also led to reliable forecasts of production (to date) of the gas-oil ratio for the three wells.