Selection of the objective function of reduction parameter optimization algorithm, and its impact on the properties of reduced MIMO models

This article presents a range of approximation error measures for reduced mathematical MIMO models of controlled objects. These measures can be of use in evaluations of reduction quality. The impact of selection of a particular approximation error measure as the objective function of the reduction parameter optimization algorithm on the frequency characteristics of the reduced model was analyzed. Particular emphasis was given to those cases where the individual channels of a MIMO model differ significantly from one another. The results are presented on the example of a one-phase zone evaporating tubes model of a BP-1150 steam boiler.