Eigen spectrum based moment matching technique for model order reduction

Abstract

In this paper a mixed approach for model order reduction is proposed. The proposed approach combines eigen spectrum analysis and Padé based moment matching technique for model order reduction. The denominator of the higher order system(HOS) is derived using eigen spectrum analysis whereas numerator dynamics are determined using Padé based moment matching technique which utilizes both time moment and Markov parameters. The efficiency and powerfulness of the method is evaluated in terms of a performance index which is known as integral square error(ISE). This method guarantees stability if the original system is stable. I. INTRODUCTIONIn the analysis of complex and intricate systems, it is often required to simplify the system for the ease of analysis, simulation and implementation. When mathematical description of any higher order system (HOS) is presented, several considerations and assumptions are made. Moreover, consideration of all aspects would make the mathematical description of the system more complex to analyze. This has preceded model order reduction (MOR) a necessary tool. Concept behind the model order reduction is to replace the higher order system with an approximated system, which provides ease of simulation, analysis and better control. Further, such a system has reduced dimensions, which significantly reduces the simulation and design time. Additionally it must be ascertained that the characteristics of the reduced order system remain pristine. The area of model order reduction is very vast and has always attracted researchers to extend the previous literature or to propose new methods. Model order reduction has been broadly classified in two domains i.e. Time domain and Frequency domain. A wide variety of techniques have been proposed in both the domains of model order reduction [1][2][3][4][5]. In time domain Singular method,Modal method,Optimal Solution etc. have been proposed [5][6][7][8][9][10]. Whereas, in frequency domain Padé approximation, Routh Approximation, Continued Fraction etc. have been proposed [11][12][13][14].

Numerical and Experimental Analysis of DVA on the Flexible-Rigid Rail Vehicle Carbody Resonant Vibration

“…However, rail vehicle coaches contain a variety of equipment under the chassis, which varies depending on the functional requirements of the coach [20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36]. As a result, the suspension characteristics of the equipment should be considered.…”