Simplification of the mathematical models of objects in simulator complexes

Mathematical models of control systems are important for problems that are solved in real time in the software for simulator complexes of complicated machinery (pilot simulators, transport simulators, etc.). Simplification of the models by excluding weakly affecting components allows one to reduce loading on the computational medium of simulators. The Schur real decomposition approximates a simplified model to the model of the initial system. A new method to increase the quality of approximation with allowance for transient and established values of object reaction, that is, approximation of transfer functions, is suggested. It is made by decreasing deviations over the matrices of a transfer function between the initial and simplified models. The applicability and effectiveness of this method of simplification are proved using a dynamic model of a pilot system as an example. © 2009 Springer Science+Business Media, Inc.