Restricted structure control of multiple model systems with Series 2 DOF tracking and feedforward action

Abstract

SUMMARYThe solution of a scalar optimal control problem is discussed where the feedback, series tracking and feedforward controllers are chosen to have a very simple. Each controller term may be chosen to be of reduced order, lead/lag, or PID forms, and the controller is required to minimize an LQG cost-index. The optimization is based upon a cost-function which also allows separate costing of the terms due to the feedback, tracking and feedforward controllers. The system model can be uncertain and can be represented by a set of models over which the optimization is performed. This provides a form of robust optimal control that might even be applied to non-linear systems that can be approximated by a set of linearized models.The theoretical problem considered is to obtain the causal, stabilizing, feedback, series-tracking and feedforward controllers, of a prespeci"ed form, that minimize an LQG criterion over the set of possible linear plant models. The underlying practical problem of importance is to obtain a simple method of tuning low-order controllers, given only an approximate model of the process. The results are illustrated in a power generation control problem for a system represented by 12 di!erent linearized plant models. The single feedback controller that is obtained has a simple form and stabilizes the full set of models.

Restricted structure predictive optimal control

“…Say, for example, that the plant is believed to be represented by one of a set of linear models, then a simple controller can be computed that will stabilize the set of models and minimize associated cost terms. This approach of Multiple Model Robust Control is easy to apply to this predictive control problem [40]. An alternative approach is to represent the uncertainty using additive or multiplicative uncertainties and apply H 1 control design.…”