Frameworks in Problems of Structural Identification Systems

Nikolay Karabutov
2017 International Journal of Intelligent Systems and Applications  
The new approach to structural identification of nonlinear dynamic systems under uncertainty is proposed. It is based on the analysis of virtual frameworks (VF), reflecting a state of a nonlinear part system. Construction VF is based on obtaining special an informational set describing a steady state of a nonlinear dynamic system. Introduction VF demands an estimation of stru ctural identifiab ility of a system. Th is concept is associated with nonlinearity o f system and properties VF. The
more » ... od of an estimation of structural identifiability is proposed. The appearance of the insignificant virtual frameworks, not satisfying to the condition of structural identifiability, is considered. Algorith ms for an estimat ion of a nonlinearity class on the basis of the analysis of sector sets are proposed. Methods and procedures of the estimation of framework single-valued and mult iple -valued nonlinearit ies are proposed. The method of the structurally-frequency analysis is proposed and applied to validate the obtained solutions. VF is proposed for identification of an order and a spectrum of eigenvalues of a linear d ynamic system. The possibility of application VF for the problem solving of identification static systems is shown.
doi:10.5815/ijisa.2017.01.01 fatcat:kvjciifbpjh4vcjpso3cbvwwrq