Hybrid Techniques for Reduction of Linear Time-Invariant Systems

Shilpi Lavania (Bundelkhand Institute of Engineering & Technology, India)

Model order reduction (MOR) has an unprecedented impact on modeling of complex control system problems. This paper strives to propose three novel combinational diminution methods for order reduction of complex high order linear time- invariant (LTI) single input-single output (SISO) systems. In each method described, the diminutive approximants for denominator are derived using modified Padé which includes matching of both markov parameters and time moments, whereas numerator in each method is derived using different MOR techniques. These three methods for the evaluation of numerator approximants are Factor division algorithm, Cauer II form and Cauer III form. Further these hybrid methods are compared with the existing model order reduction techniques to analyze their performance. Performance is assessed in respect of a performance index (PI) known as Integral square error (ISE). All suggested methods guarantee stability of the obtained lower order system provided that the original high order system is stable.

Journal: International journal of simulation: systems, science & technology V18

Published: Dec 30, 2017

DOI: 10.5013/IJSSST.a.18.04.03