MODELFILTER REDUCTION OF FREQUENCY LIMITED SYSTEMS

Abstract

Model reduction of frequency limited systems techniques plays a vital role in analysing a system of very high order such as telecommunications, network systems, wave propagation, transmission lines, columns distillation, chemical reacters and economical systems etc. Model reduction is associated with systems which are related to delay equations, equations of partial differential. Model reduction reduces the complication of the higher order original system, due to which analysis of such system becomes uncomplicated in specified frequency range. There are several techniques related to model reduction. Generally, balanced truncation technique produces approximate lower model from relatively higher order model original systems for throughout entire range of frequency. However, there are many usages that demands the approximation for required interval or band. Simulations of such systems signifies that the errors (subtraction of reduced system from its of the related models of original system) should be small, and such aspect should be considered. Gawronski and Juang introduced another technique in which the approximation is considered in certain frequency band. Reduced order systems produced from this technique has stability issue. To address such limitation, several model reduction techniques where proposed to overcome the stability problem of reduced systems. Results of such proposed techniques were tedious and time taking. On the basis of this, enhanced frequency limited model reduction techniques were needed. This research mainly emphasis on model reduction technique that maintains stability, input and output properties of the original systems and produce effective results with least approximation error. In this thesis, comparison of several techniques of reduced-order modelling along with different examples are being presented to show their efficacy