Naslov
Interpolation-based parametric model reduction forefficient damping optimization

Autori
Beattie, Christopher ; Gugercin, Beattie ; Tomljanović, Zoran

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, neobjavljeni rad, ostalo

Skup
Numerical methods for optimal control problems: algorithms, analysis and applications

Mjesto i datum
Rim, Italija, 19.06.2017. - 23.06.2017

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Podatak o recenziji nije dostupan

Ključne riječi
Model reduction ; Interpolation ; second order system

Sažetak
We consider the optimization problem of semi-active damping of vibrating systems. The mainproblem is to determine the best damping matrix which will minimize the influence of the input tothe output of the system. We use a minimization criteria based on theH2system norm. Since thatthe objective function is a non-convex function, this damping optimization problem usually requiresa large number evaluations of objective function. Thus, we propose an optimization approachthat calculates the ‘interpolatory’ reduced order model which allows significant acceleration ofoptimization process.In our approach we use parametric model reduction based on Iterative Rational Krylov Algorithm, which ensures a good approximation ofH2system norm. For the sampling parameters within theparametric model reduction we propose fixed sampling or adaptive sampling approach. Moreover, in order to preserve important system properties, we use second order structure, which in modalcoordinates allows very efficient implementation of our approach.Proposed approach provides efficient approximation of optimal parameters with significant accel-eration of the optimization process, which is also illustrated in numerical experiments.

Izvorni jezik
Engleski

Znanstvena područja
Matematika

POVEZANOST RADA