Abstract :
Here, it is studied the control of integrating systems in the presence of model uncertainty. For this kind of system, a method is
proposed to overcome one of the major barriers to the practical implementation of the existing robust MPC approaches: the
assumption that the steady state of the true plant is known. To deal with unknown steady states, the controller incorporates a statespace
model in the incremental form, which is a model framework frequently adopted by MPC packages. In this case, it is shown
that for integrating systems, minimizing the integrating states at steady state is not sufficient to guarantee the stability of the
uncertain plant. It is proposed a modified cost function that allows the controller to stabilize a family of plants, even when the
steady state is not at the origin. To compute the control law, a Min-Max problem is solved with model uncertainty assumed to be of
polytopic type. The application of the proposed controller is illustrated with the simulation of an industrial multivariable system.
For this example, the effect of the new tuning parameters is discussed.
