Plunger lift optimization: A comparison of fuzzy and predictive controllers with manipulation of a secondary gas valve

Abstract

One of the most common problems in the oil and gas industry is the excessive accumulation of liquids at the bottom of the well. This phenomenon usually results in reduced production rates due to high values of bottomhole flowing pressure. Plunger Lift represents a low-cost solution to this issue, using a controller to compensate parameter variations in the well and system disturbances. The use of a model predictive controller demands a precise mathematical description of the system behavior, which can be difficult to be obtained for processes with high nonlinear and hybrid aspects, such as Plunger Lift. The application of a fuzzy controller for this type of system, on the other hand, can be a more feasible task, since it dismisses the need of a mathematical model, being based solely on the experience of a human operator. This work presents the application of a fuzzy and a linear model predictive controllers to a Plunger Lift process. The controllers were designed to manipulate a secondary valve installed at the top of the annular in order to improve the range of control. The implemented control techniques aimed at production optimization and regulation of the average rise speed of the plunger. The fuzzy and predictive controllers were applied to a plunger lift well in two different configurations, with and without a secondary valve present in the surface, and compared with a fixed time controller. Disturbances in the production line and static pressures were present in the tests. Graphical and performance indexes analysis were conducted, being both implemented controllers successful in maintaining optimal production rates. The fuzzy and MPC controllers generated around 55% more gas and oil than the fixed time controller. In the configuration that utilized the secondary valve, the controllers achieved safe operational values for average plunger rise speed.