CONICET | Buscador de Institutos y Recursos Humanos

A novel unified approach to two-degrees-of-freedom control is devisedand applied to a classical chemical reactor model. The scheme isconstructed from the optimal control point of view and along the lines ofthe Hamiltonian formalism for nonlinear processes. The proposed schemeoptimizes both the feedforward and the feedback components of the controlvariable with respect to the same cost objective. The original Hamiltonianfunction governs the feedforward dynamics, and its derivatives are part ofthe gain for the feedback component. The optimal state trajectory isgenerated on-line, and tracked by a combination of deterministic andstochastic optimal tools. The relevant numerical data to manipulate allstages come from a unique off-line calculation, which provides designinformation for a whole family of related control problems. This ispossible because a new set of PDEs (the variations equations) allow torecover the initial value of the costate variable, and the Hamiltonequations can then be solved as an initial-value problem. Perturbationsfrom the optimal trajectory are abated through an optimal state estimatorand a deterministic regulator with a generalized Riccati gain. Both gainsare updated on-line, starting with initial values extracted from thesolution to the variations equations. The control strategy is particularlyuseful in driving nonlinear processes from an equilibrium point to anarbitrary target in a finite-horizon optimization context.