Online multilevel controller structure for a class of dynamical systems

A two-level optimization control algorithm based on the minimum variance approach is developed to control the dynamic behaviour of large-scale interconnected subsystems under certain constraints. In the first level, the control action and adjoined parameters are obtained. In the second level, the parameters of both the coordinator and interactions between subsystems are obtained. Developments are introduced to the estimation adaptive scheme so as to determine the controller coefficients for each subsystem effectively. To show the effectiveness of the proposed algorithm, it is used successfully to control the behaviour of the head-box of a paper machine. Computer simulation results are presented to illustrate the potential of the approach proposed