Modelagem e controle fuzzy da densidade de fluidos de perfuração

Abstract

The drilling fluids play an important role in drilling of oil wells and the control of the physical and chemical properties must be carefully performed during its preparation. Currently, this process is accomplished with manual dosing of additives for specialized operators, and the control of the properties is carried out in laboratory analysis. The aim of this study was to propose an automated process of fluid preparation with online monitoring and control of the density by using a fuzzy logic-based control system. This process was conducted in a pilot plant for fluid preparation consisting of a helical rotor pump, two mixing tanks, sensors to monitor the density of the fluid, vibrating hopper, metering pump, and level sensor. To control the density of the fluid, a supervisory system for data acquisition and monitoring using the software Labview® was developed. This supervisory system receives real-time process data through communication protocols with acquisition boards by National Instruments and, using the control algorithm, triggers the process equipment (pumps, agitator and vibrating hopper) to keep the density of the fluid in the reference value. The control strategy proposed in this work was based on fuzzy logic and used the Mamdani-type inference rules as the basis for the development of a Fuzzy-PI controller. The Fuzzy-PI controller was compared to a classic type PI controller in disturbance-rejection and setpoint-tracking control problems. The performances of the controllers were evaluated by comparing their overshoots, rise times, settling times, and the integral error criteria IAE and ITAE. The results showed that, although the answers obtained from the evaluation criteria were not conclusive, economic and operational aspects indicated the superiority of Fuzzy-PI controller