Control of managed pressure drilling systems using nonlinear predictive generalized minimum variance approach based on a Volterra model

Abstract

This paper proposes a nonlinear predictive generalized minimum variance (NPGMV) control scheme for automatic control of the managed pressure drilling (MPD) systems in the presence of disturbances. Since the exact model of the system is not usually available in practice, the hydraulic flow model of MPD is described by an autoregressive second-order Volterra model. The conventional least-squares method is applied to input-output data, thereby identifying the Volterra model. Bottom-hole pressure regulation and kick handling are achieved through the control scheme. To deal with a reservoir kick, the proposed method switches to flow control mode automatically, and prevents the reservoir fluid influx into the well surface. The proposed method also has the capability to keep the bottom-hole pressure above the reservoir pressure during a formidable scenario such as pipe connection. In addition, the robustness of the controller in the face of heave disturbance and uncertainty is investigated. To show the effectiveness of the proposed method, the comparison study of several scenarios with a switching PI controller is provided and demonstrates the NPGMV control outperforms other approaches with respect to steady-state performance.

Keywords: Managed Pressure Drilling (MPD); Minimum variance control; Nonlinear control systems; Predictive control; Volterra model.