Data-driven switched control for pressure regulation in Managed Pressure Drilling

The problem to regulate the bottomhole pressure in Managed Pressure Drilling is investigated with the approach of Unfalsified Adaptive Control. The candidate controller set is designed with the knowledge of drilling process. The falsification and unfalsification conditions are presented. In these conditions, the output data are used directly to test the current active controller. A data-driven controller switching algorithm is designed to monitors and switches controller based on only measured data. The values of the process parameters are not used at all so that the control law is robust to the parameters uncertainty. Convergence theorem of the algorithm is proved and the performance of the algorithm is illustrated by MATLAB simulations.