Application of stabilizing predictive control to boilers

A multivariable state space formulation of the GPC (generalized predictive control) law that can be implemented using output feedback is described. A particular nonminimal state space model allows the current state to be known without the use of an observer. The penalty on the terminal value in the GPC cost criterion is modified to yield a predictive control law that ensures stability of the closed loop system for arbitrary costing horizons. This control law, termed stabilizing predictive control (SPC), is extended to achieve asymptotic tracking of the reference signal. The SPC law is then applied to a continuous time boiler model with good preliminary results