scientific article; zbMATH DE number 1391261

From MaRDI portal
Publication:4935989





This habilitation thesis is devoted to systems of ordinary differential equations with stationary random excitation terms. For stable linear systems as well as systems with polynomial nonlinearities the existence of stationary solutions and the computation of their distribution characteristics as second-order moment functions are considered. Such questions arise in the investigation of the long-time behaviour of the system response to permanently acting excitations. The results can be directly applied to discrete vibration systems with damping and an external random excitation. The presented results are essentially based on the work of Jürgen vom Scheidt and his research group on weakly correlated random functions and their application to random equations of mathematical physics [see \textit{J. vom Scheidt}, ``Stochastic equations of mathematical physics (1990; Zbl 0715.60061)]. They can be considered as extensions and generalizations of the work on random vibration systems [see \textit{J. vom Scheidt}, \textit{B. Fellenberg} and \textit{U. Wöhrl}, ``Analyse und Simulation stochastischer Schwingungssysteme (1994; Zbl 0816.60052)]. In particular, such topics are chosen where the author himself contributed to the development: modeling of random excitations, higher-order expansions of correlation functions, systems with polynomial nonlinearities and model reduction. While Chapter I introduces some preliminary results, Chapter II is devoted to the modeling of stationary random excitation processes (and homogeneous fields) with prescribed correlation functions and spectral densities as well as given smoothness of the trajectories. The basic idea consists in the description of these processes by integral functionals \(g(t,\omega)=\int^\infty_0Q(s)^\varepsilon f(t-s,\omega)ds\) of \(\varepsilon\)-correlated functions \(^\varepsilon f\). To find suitable kernel functions \(Q\) stationary solutions of so-called form filter equations with \(\varepsilon\)-correlated inputs are investigated. Chapter III starts with stable linear systems and derives higher-order expansions for the correlation function of the stationary solution with respect to the correlation length \(\varepsilon\). For systems of ODEs containing polynomial nonlinearities perturbation method is applied to find pathwise solutions in form of power series with respect to a parameter \(\eta\) controlling the nonlinearities. Under the assumption that for \(\eta=0\) the system is stable and that the excitation terms are bounded the radius of convergence of the perturbation series is estimated. The proof uses properties of the integer sequence of Catalan numbers. Further, it is proved that the perturbation series form stationary solutions. In Chapter IV for large-scale systems of linear ODEs approximations of the correlation and spectral density functions of the stationary solutions are found by means of model reduction techniques. To assess the approximation accuracy suitable error criterions are introduced and error estimates are given. Numerical results concerning random axial vibrations of a beam are presented.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4935989)