Evolution of Control Strategies to Minimize Pressure Fluctuation During Connections in MPD Deepwater Operations

Abstract

MPD techniques are used nowadays to precisely control the pressure profile throughout the wellbore. They are usually employed in drilling with narrow pressure windows, such as through highly fractured reservoirs, regions near depleted wells, regions with small fracture pressure, among others. Low drilling pressures may result in undesirable influxes from reservoir fluids, whereas high drilling pressures may cause mud loss and may damage the near-well region, reducing the future well productivity. This work details the development of a MPD simulator (hydraulics and control strategy). The hydraulic model is transient and based on the explicit and simultaneous solution of mass and momentum flow equations. This model is different from the usual control purpose hydraulic models since the well geometry is actually discretized, allowing a more precise determination of the pressure profile along the well. Furthermore, this discretization also enables the simulation of the delay between an action at the choke and the corresponding response at the bottom hole, which is fundamental for a realistic evaluation of a control strategy. A nonlinear control strategy was also created in this work. This strategy is based on the aim of a smooth exponential evolution of the choke pressure until it reaches its setpoint. The execution of the hydraulic simulation several times is not required, as in typical model predictive controllers (MPCs). The control strategy is also not so susceptible to control parameters, as in PID classical controllers. An experimental drilling plant was employed for validating the nonlinear control strategy. The nonlinear control algorithm and the classic control scheme were compared. Using real-time measurements, the automated continuous drilling unit performance is analyzed under a scenario of flow disturbance. The fast and precise hydraulic model coupled with an efficient control method makes the present simulator feasible for real-time applications.