Conventional Plunger Lift Optimization

Anvar Akhiiartdinov – University of Tulsa
Dr. Eduardo Pereyra – University of Tulsa
Dr. Cem Sarica – University of Tulsa

2020 ALRDC/SWPSC Artificial Lift Workshop
Oklahoma City, OK, USA
Feb 17-20

This presentation attempts virtual flow metering for conventional plunger lift wells by using pressure signals and arrival sensor data. An artificial neural network was applied to data obtained from a multiphase flow loop. It was found that this information is sufficient to model gas flow rate. The model does require calibration. Predicted flow rates could be used for allocation purposes and the after-flow time can be optimized based on the predicted flow rate.

Casing pressure and gas flow rate were also modeled using controller set points and initial conditions. Integration of “on” and “off” models predicted cumulative production with 7.7% error. The current understanding of plunger lift physics dictated the limitations of the optimization results. Optimization was able to increase gas production in the studied example by 8.5%. The described optimization model could complement existing alarm systems.