Production forecasting of shale wells has proven to be a challenging task. Several empirical models have been proposed for shale well production data analysis; the modified hyperbolic (MH) model is the most popular one utilized by the industry due to its simplicity. However, the parameter "minimum decline rate Dmin" required in the MH model always depends on an evaluator's experience. In addition, in the MH model we assume that the decline during late boundary dominated flow (BDF) is exponential, which is rarely appropriate as pointed out by Fetkovich and by Spivey et al.
In this paper, we first present important criteria that have been overlooked in existing empirical models but that are necessary for a robust and accurate production decline model. It turns out that the modified hyperbolic model does not violate these criteria. As a result, on the basis of a two-segment hyperbolic model, we propose an optimization algorithm that can be applied to accurately estimate five important production parameters for a well that is in the BDF regime
initial production rate qi,
initial decline rate Di,
initial Arps b parameter bi for early transient flow,
final Arps b parameter bf for late BDF, and
optimal switching time tc from early transient flow to late BDF.
In addition, we developed a three step diagnostic approach for the analysis of flow regime that a well has gone through. The proposed diagnostic approach is effective in reducing the uncertainty in well flow regime and the estimation of production parameters. The merits of this new algorithm are demonstrated with the application to analysis of real production data in Eagle Ford and Bakken reservoirs.
The methods proposed in this paper are the cornerstone to predict well estimated ultimate recovery (EUR). In addition, this work has significant impact on many other related projects such as the construction of type well production profiles, optimal completion design and probabilistic decline curve analysis since they all depend on the five production parameters.
Since 2004, the advance of technology in horizontal drilling and multistage fracturing has made unconventional resources an increasingly significant source of hydrocarbon supply in the global energy market (Holditch 2010). While these unconventional resources bring great benefits to the industry, the characterization of reservoir fluid flow in shale formations has proven to be challenging due to the significantly different fluid-driven mechanisms. Numerical simulation is a tool that is frequently used in the industry to characterize the reservoir fluid flow in shale formations. This method requires extensive manpower, time and cost, and as a result, much effort has been made to develop other efficient methods (e.g., fast marching method, semi-analytic approach, etc) that can be applied to find a both reliable and fast solution to those important engineering and economic problems (King et al; 2016; Xue et al, 2016; Wang et al., 2017). In addition to unconventional reservoir characterization, production forecasting of shale wells has also proven to be a challenging and important problem, since it has significant consequences in investment decision making and is a major component of reserves estimation required for reports to regulatory agencies.