Model Predictive Control Applied to a Plunger Lift Artificial Elevation System
Oil production; Plunger Lift; Model Predictive Control; Production optimization
During the productive life of most gas wells there is an accumulation of liquid at the bottom of the well which causes a pressure contrary to the flow of the reservoir, reducing the production of the well. In this way, one of the possible artificial lift methods to solve this problem is represented by the Plunger Lift. It consists of a low-cost installation and maintenance technique that uses a piston to, among other uses, increase the efficiency of liquid removal from the system. However, this method requires a well-adjusted controller to appropriately define the opening and closing periods of a motor valve installed in the wellhead for production control. Predictive control is a technique that can be applied to perform such task, performing predictions of the future behavior of the plant in order to obtain its optimum performance. In this work, the implementation and application of a linear predictive controller to a simulated well operated by Plunger Lift was performed. The controller was applied under different conditions of draining ability, with and without, of the gas present in the annular space. Results of MPC without drainage were compared with conventional versions of Plunger Lift control in the oil industry through graphical analysis and observation of performance parameters. This controller obtained better results regarding the control of the average speed of plunger ascent and accumulated production. In the comparison of MPC with and without drainage, the version capable of manipulating the gases from the annular space was able to cause the system to operate at speeds of piston ascent in safer ranges.