ALRDC is the Artificial Lift Research and Development Council that is focused on advancing the knowledge and practice of artificial lift technology and business. This year, the ALRDC Sucker Rod Pump Workshop was held in Oklahoma City, Oklahoma. Several oil and gas companies presented on the design, optimization, and challenges faced with Artificial Lift. Full presentations and abstracts can be viewed on ALRDC.
Our Technical Director, Jesse Filipi, presented on “Using IIoT—Enabled Machine Learning to Autonomously Optimize RodLift Wells.” The presentation focused on Ambyint’s approach to deliver a new, autonomous technique for optimizing rod lift wells using IIoT—enabled controllers and machine learning. By combining IIoT—enabled edge controllers producing higher quality data and physics-based mathematics with artificial intelligence and machine learning techniques, it is possible to dynamically adapt the operating set points of rod pumping equipment. This results in wells that produce more oil, use less power, require less maintenance, and fail less frequently.
The presentation was followed by Brian Arnst, Director of Optimization at Ambyint, who co-presented with Michael Sullivan of ConocoPhillips (COP) on “Economically Enabling Autonomous Pump by Exception on Sub-3 BOPD Permian Basin Stripper Wells Using IIoT—Enabled Edge Devices. Jointly, they discussed the results of a recent 20 well pilot deployed in Crane County, TX. COP was faced with hundreds of wells that averaged 3 BOPD and used Ambyint’s solution to prove the value of the technology as an economically viable solution to extend the economic life of low oil rate stripper wells.
Over the course of the three-month pilot, the COP team operated wells by exception through the Production Optimization Platform. The operator saw immediate results through automated alerts when mechanically induced downtime events occurred. Our Ambyint solution was able to deliver an increase in dialed in wells from 15% to 65%. Additional benefits realized included a reduction in overpumping, a 25% reduction in average runtime, and 300% reduction in cycling events. This created a 58% reduction in power consumption and 139% increase in run life. The full results and associated cost reduction are shown in the table below.