Reliability Improvement in Beam Pumps by use of a Unique Methodology Combining Real-Time Automation and Failure Analysis

Abstract

Abstract Santos has produced oil from the Cooper Basin, central Australia since 1983. A range of artificial lift systems are used, dependent on fluid rate and reservoir depth with the beam pump the dominant artificial lift system. Historically the fleet experienced high subsurface failure frequency and increasing operating cost. Information regarding pump operating conditions and the root cause of failure was inconsistent in reporting standard and not readily available. As a consequence the details and mode of failure were not well understood. This paper will focus on analysis of beam pump failure root causes and trends, and how the operator has used this data combined with automated equipment to reduce failure frequency and operating cost. Santos started a program of beam pump automation in 2006. Before 2009, Santos relied on review of individual well files to determine pump operating history and failure mode. No central collation of well history or pump failure information was available. A database was built, which compiled failure details for more than 2,000 pump installations. This database has enabled run life, failure mode and frequency to be tracked and analysed. A program of automation and control upgrades to 85% of the beam pump population has provided greater visibility of operating conditions and parameters by means of telemetry to the office. This information in conjunction with the database is now used to better understand reasons for failure, and to design future downhole configurations as a result. In 2006, all Santos beam pumps were non-automated and had a failure frequency of 0.4 failures per well per year. Since the implementation of the automated fleet and the database, data shows frequency has reduced by 70% for automated wells. In comparison, the remaining non-automated beam pumps continue to have a failure frequency of 0.3 to 0.4. Since 2006 the beam pump well count has increased by 60% while the number of annual pump repairs has reduced by 30%. The data indicates failure modes are different on wells with automation. As an example, automated wells have a lower frequency of parted rods, resulting from better rod string design and control. This paper will assist operators to benchmark beam pump run life in similar operating conditions such as the Cooper Basin. It also presents data and findings to support the benefit a beam pump automation and control program may provide. Detailed understandings of failure modes have helped Santos and its joint venture partners gain significant economic benefit, and develop "Best Practice" guidelines for beam pump system design as a consequence.