Improved sucker-rod system design and optimization

Abstract

A new sucker-rod simulator saves computer run time and provides sufficient accuracy for conceptual sucker-rod-system design. A new method expands the scope of the American Petroleum Institute's Recommended Practice (RP) 11L for up to seven tapers (including sinker bar) of any rod material that obeys Hooke's Law. Secondary findings apply to the creation of a consistent sucker-rod nomenclature, a procedure for history matching field data, and the potential for saving computer run and getting preliminary accuracy from an existing simulator. The new sucker-rod simulator developed by this research saves run time and meets the 20% error tolerance normally associated with conceptual design. The new simulator uses the method of characteristics and nondimensional partial differential equations. Prior to 1990, no sucker-rod design method met the error tolerances for conceptual, preliminary and final system design. Also prior to 1990, the only sucker-rod simulators that ran profitably fast did so by using a coarse grid. A coarse grid saves run time by minimizing the number of rod subdivisions (elements). The number of rod elements directly affects simulation accuracy. Minimizing the number of elements reduces accuracy. Improving the methodology behind RP 11L expands 11L's scope to include composite sucker-rod materials and sinker bar. The specific examples are for fiberglass-steel rod strings and for rod strings with up to seven tapers. RP 11L was originally intended for all steel rod strings without sinker bar. This research resolves the literature's chaotic sucker-rod notation by creating a sucker-rod nomenclature that is consistent with the Society of Petroleum Engineers' standard. Since field data validates an existing simulator, a history-matching procedure is developed. The validated simulator validates the new (conceptual-quality) simulator. While validating the conceptual-quality simulator, the existing simulator demonstrated the ability to save computer time and to meet the 10% error tolerance normally associated with preliminary design. Initially, this was considered a coincidence. However, a review of previous work revealed a consistent pattern. The pattern includes constant motor rpm, averaging the results of the second and third cycles, and 10 to 20 rod elements.