Model Reference Adaptive Control Applied to the Improvement of the Operational Conditions of a Sucker Rod Pump System

Several studies have shown that the satisfactory operation of oil wells with sucker rod pumps is attributed to the techniques and methods able to control the performance of the well. The Polytechnic School of Federal University of Bahia, through the Artificial Lift Lab, has nowadays a reduced plant of a sucker rod pump with an artificial well of 32m height fully instrumented, with full access and visible. Among some studies already performed with this sucker rod pump, there is the use of a dynamic model to control the fluid level in annular well. The presence of uncertainties is related to the fluid characteristics in the well, associated with either the electrical or the mechanical assembly, may jeopardize the desired performance of the PID controller. These uncertainties usually do not contribute to good results and may put in risk the well's productivity. The aim of this paper is to show a model reference adaptive controller to deal with some uncertainties in the system model to control the level of the annular well. The results show that the adaptive controller is able to deal satisfactorily with changes, such as: in the pumped fluid composition (parts of water, oil and gas), which is a very common situation in real production fields.