Evaluating and optimizing well-completion procedures can be difficult because of a lack of understanding about the reservoir and completion dynamics. In many cases, a well completion is a compromise between optimum reservoir, well, completion, and operational factors. In this paper, we discuss the use of a form of artificial intelligence, the Artificial Neural Network (ANN), as a tool to enhance our ability to improve well economics. ANN models trained on reservoir, well, and completion information can predict well cumulative production with an acceptable degree of accuracy (less than 15% average absolute error).14  Sensitivity studies of these networks show that for a given reservoir quality, the completion method can significantly affect the well's production outcome. Case histories are presented in which an ANN sensitivity analysis was used to justify changes in completion/stimulation procedure. These ANN-enhanced completions resulted in improved well economics compared to the standard completion-optimization methods normally used in these fields.

This paper presents case histories in which ANN helped establish well completion priorities that improved well economics. ANN technology has been used in areas where a well-completion experience base exists. The completions analyzed with ANN techniques include the Red Fork formation in Roger Mills and Custer counties in Oklahoma; the Frontier formation in Lincoln County, Wyoming; and the Granite Wash formation in Roberts County, Texas. Specific areas of interest include any quantifiable well, reservoir, or completion characteristic that affects well production (porosity, thickness, pressure, geology parameters, well characteristics, completion/stimulation procedures, etc.). Net present value (NPV) was used for quantifying the economic effect of ANN enhancement on well-completion economics.

You can access this article if you purchase or spend a download.