Model predictive control and estimation of managed pressure drilling using a real-time high fidelity flow model

Area of Science:

• Petroleum Engineering
• Control Systems Engineering
• Computational Modeling

Background:

• Managed pressure drilling (MPD) is crucial for oil and gas well control.
• Current MPD often uses simplified models, limiting optimal performance against system perturbations.
• Model predictive control schemes are common in MPD but face practical limitations.

Purpose of the Study:

• To implement a highly accurate system model for real-time MPD controller application.
• To develop a high-fidelity model that approximates offline results without expert operation.
• To evaluate the controller's performance using the new real-time model under diverse drilling scenarios.

Main Methods:

• Development and adaptation of a high-fidelity system model for real-time controller integration.
• Simulation studies to test controller behavior under various drilling conditions.
• Analysis of the impact of sparse downhole feedback measurements on model effectiveness.

Main Results:

• Successful implementation of a real-time high-fidelity model for MPD controllers.
• Demonstrated approximation of offline high-fidelity model results in real-time.
• Controller performance evaluated under simulated drilling conditions, including sparse data.

Conclusions:

• The developed real-time high-fidelity model enhances MPD controller capabilities.
• This model offers accurate performance without necessitating expert operation.
• The study validates the model's effectiveness in managing well pressures under varied drilling conditions.