Association rules mining for long uptime sucker rod pumping units

Abstract

The uptime of sucker rod pumping units (SRPUs) varies significantly due to differences in geological reservoir conditions, as well as design and operation parameters of the SRPUs. Conventional approaches encounter challenges in accurately quantifying the intricate interplay of these factors with uptime. This study investigates the application of association rules mining in exploring reliable measures in long uptime SRPUs. We create an uptime dataset for 5789 sucker rod pumping wells with 16 features encompassing reservoir conditions, production characteristics, SRPUs design, operational settings, and uptime. A novel algorithm is proposed to mine association rules between uptime and other features, yielding over 1,000,000 rules. These rules reveal relationships and impacts on long uptime. By following these rules, adjustments can be made to improve the uptime of SRPUs through alterations in design and operational parameters. Two case studies were presented to illustrate that by adjusting the design and operational parameters of SRPUs based on the association rules, the probability of improved operational performance can be increased by 30 % and 8 % respectively. This study explores the association rules of long uptime SRPUs from a data-driven paradigm, providing scientific guidance for subsequent oil well design. The research methodology is also valuable for other mechanical equipment.