(2024002) Leveraging Machine Learning Models for Optimization


Luke Beaudry, dv8 Energy

Incomplete fillage conditions where the downhole pump does not completely fill up with incompressible liquid has been widely accepted to have detrimental effects on pumping efficiency and moreover the equipment longevity in sucker rod pumping applications.
Methods of synchronizing the pump displacement to the wells inflow and thus reducing incomplete fillage has been of keen interest to the industry. 

A sophisticated pump off control (POC) algorithm called Advanced Fillage Mode (AFM) & Fluid Level Model with a continuous feedback mechanism has been shown to significantly reduce incomplete fillage pumping cycles using a variable frequency drive (VFD) for speed control. 

Presentation Information

Wed 2:00PM - 2:50PM, Room 104
Thu 10:20AM - 11:10AM, Room 104

Annual Conference Info