A new supervision-based adaptive control for pressure process rig to overcome bursting phenomena under low persistently excitation

A modified supervision algorithm for an indirect adaptive control of a pressure process is presented. The aim is to improve control performance under low persistently excitation that is common in the industrial process in which adaptive controllers are known fail to guarantee the boundedness of parameter estimates and the other closed-loop signals. The proposed supervision algorithm is designed to detect low sufficiently conditions quickly as possible, analyse monitoring signals using recursive computation and make a logic decision for updating the controller parameters or freezing the adaptation. The control performance is verified through experiments on a lab-scale pressure processrig and compared with the original adaptive controller.