Real‐Time System Identification: An Algorithm for Simultaneous Model Class Selection and Parametric Identification. (15th June 2015)

• Record Type: Journal Article
• Authors: Yuen, Ka‐Veng; Mu, He‐Qing
• Is Part Of: Computer-aided civil and infrastructure engineering. Volume 30:Number 10(2015:Oct.)
• Page Start: 785
• Page End: 801
• Publication Date: 2015-06-15
• Subjects: Civil engineering -- Data processing -- Periodicals; Computer-aided engineering -- Periodicals; 624.0285
• DOI: 10.1111/mice.12146
• Languages: English
• ISSNs: 1093-9687

Abstract

<abstract abstract-type="main"> <title>Abstract</title> <p>In this article, a novel Bayesian real‐time system identification algorithm using response measurement is proposed for dynamical systems. In contrast to most existing structural identification methods which focus solely on parametric identification, the proposed algorithm emphasizes also model class selection. By embedding the novel model class selection component into the extended Kalman filter, the proposed algorithm is applicable to simultaneous model class selection and parametric identification in the real‐time manner. Furthermore, parametric identification using the proposed algorithm is based on multiple model classes. Examples are presented with application to damage detection for degrading structures using noisy dynamic response measurement.</p> </abstract>