Heuristic algorithm based Controller design and stability Analysis using model order Reduction of interval system

Abstract

In this work a controller is designed for a reduced order interval newlinesystem model using heuristic algorithm newlineIndustrial processes with large number of state variables are newlinemodeled as higher order model In order to reduce the complexity in analysis newlinethe higher order model is reduced it into lower order model Various newlineconventional model order reduction techniques such as aggregation method newlineRouth Pade approximation Krylov subspace proper orthogonal newlinedecomposition and hankel norm methods are available for conventional newlinemodel representation In the proposed model order reduction technique for newlineinterval system the key characteristics of higher order system are matched newlinewith the reduced order model by retaining the dominant poles of higher order newlinemodel This is mainly due to less variation of initial time moments and newlineMarkov parameter Validation of model order reduction is performed using newlineKharitonov theorem step response nyquist response and bode response test newlineThe variations in process parameter due to perturbations bias and newlineenvironmental changes may lead to model mismatch with degraded newlineperformances Hence an attempt has been made to represent a bioreactor with newlineall its dynamics and process perturbations as an interval system The width of newlinethe interval coefficient depends on sensitivity of process parameter and its newlineparametric uncertainty In this thesis the estimation of uncertainty and level newlineof confidence are carried out with multiple experiment method newline