MODEL ORDER REDUCTION USING FREQUENCY RESPONSE MATCHING AND APPLICATION TO INSTRUMENTATION SYSTEM

Abstract

Two fresh approaches to reduce the order of linear systems in frequency domain are proposed. In first approach dominant poles are retained whereas zeros are synthesised by frequency response matching of the original and reduced order system. In this approach, there is exact matching of the steady state parts of the response at any frequency and simultaneously transient is minimized at that frequency. In the second approach point to point matching method has been used. In this method the magnitudes of original and reduced order system are forced to match at a given frequency. Selection of number of frequencies at which magnitudes are equal depends, upon the order of reduced model. Both the methods have been illustrated with examples and their salient features are also brought out. An application of reduced order modelling in an instrumentation systems has been presented.