
Identification of a time-varying mechanical system using the Akaike information criterion
Author(s) -
Michal Růžek,
Simon Chesné,
Didier Rémond
Publication year - 2017
Publication title -
mechanics and industry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.256
H-Index - 18
eISSN - 2257-7777
pISSN - 2257-7750
DOI - 10.1051/meca/2016066
Subject(s) - akaike information criterion , bayesian information criterion , identification (biology) , system identification , polynomial , estimation theory , mathematics , torsion (gastropod) , signal (programming language) , computer science , control theory (sociology) , mathematical optimization , algorithm , statistics , mathematical analysis , data mining , artificial intelligence , medicine , botany , surgery , control (management) , biology , programming language , measure (data warehouse)
International audienceThis article deals with the use of Akaike information criterion in the case of an identification problem of a time varying mechanical system. We studied a prismatic beam with a disk excited in torsion to illustrate how the proposed approach can be implemented. The boundary conditions of the beam can changein a controlled way during the experiment. Therefore, our system can be considered as a one d.o.f. system with time-varying parameters. A method based on least-square estimates is used for the identification of the parameters. The use of the Akaike information criterion allows to choose automatically the order ofpolynomial basis used during the estimation step of the derivatives of the measured angle signal. It is shown through an experimental validation that the AIC criterion is a robust and automatic tool for numerical estimation of signal derivatives