Artykuły naukowe (WEl)

Stały URI dla kolekcji

https://oa.zut.edu.pl/handle/20.500.12539/61

Przeglądaj

Przeglądaj Artykuły naukowe (WEl) wg Temat "frequency analysis"

Teraz wyświetlane 1 - 2 z 2
  • Miniatura
    PozycjaOpen Access
    Comparison of frequency domain methods for discrete-time, linear time-varying system with invariant eigenvalues
    (International Society for Advanced Research, 2007-10) Orlowski, Przemyslaw; Institute of Control Engineering Szczecin University of Technology
    The main aim of this paper is to compare and evaluate frequency methods applicable for discrete-time (DT) linear time-varying (LTV) systems, in particular: two-dimensional (time, frequency) transfer function (2D-TF), timeaveraged 2D-TF and approximated Bode diagrams calculated using SVD-DFT approach and power spectral density (PSD) properties. The main evaluation criteria is possible applicability to feedback system stability analysis. The paper begins from short theoretical background of frequency methods applicable for LTV systems. Further properties of these methods are compared and discussed on the basis of particular case of parameter controlled switching DT LTV system.
  • Miniatura
    PozycjaOpen Access
    An extension of Nyquist feedback stability for linear time-varying, discrete-time systems
    (International Journal of Factory Automation, Robotics and Soft Computing, 2007) Orlowski, Przemyslaw; Institute of Control Engineering, Szczecin University of Technology
    The paper concerns on extending the classical Nyquist theorem to stability analysis of linear time-varying (LTV) discrete-time (DT) feedback control systems. Frequency methods, are well- known tool for analysis and synthesis for linear time-invariant systems. Unfortunately, the methods cannot be applied for LTV systems. The main objective is to show that Bode plots approximated using SVD-DFT are adequate methods for evaluating stability margins as well as external stability for LTV systems. We assume discrete-time state space models with time dependent system matrices defined on a finite time horizon. To solve the problem we employ discrete Fourier transform and singular value decomposition of a system matrix operator as well as properties of power spectral density.