A new approximation methodology for linear multivariable large-scale systems

Mohamed Belhocine (1, 2) , Bogdan Marinescu (1, 2)

1 ECN - École Centrale de Nantes
2 LS2N - Laboratoire des Sciences du Numérique de Nantes

A new approximation methodology is proposed in this paper to reduce the dynamic models of linear multivariable dynamic systems with slow decay of Hankel spectrum. The main problem is that for such systems, the dynamics are difficult to evaluate using directly the classic indexes like the energy of the state variables or the residues related to the modes and the existing techniques may not give always satisfactory results. For this, we propose here a mixed approach based on the energy of the impulse response and the link between the modes and the states to better evaluate the dynamics and finally reduce the model. This preserves the dynamic structure of the system and allows one to obtain acceptable results with reduced models of much lower order than the ones obtained by the classic approaches. Theoretical results are first presented and a power system practical example is next treated to validate our methodology.

Domaines: Automatique / Robotique

Mohamed Belhocine, Bogdan Marinescu. A new approximation methodology for linear multivariable large-scale systems. Advanced Control and Diagnostic, Nov 2017, Bucarest, Romania. ⟨hal-02510549⟩