Full-field finite element model updating using Zernike moment descriptors for structures exhibiting localized mode shapes. (15th April 2019)

• Record Type: Journal Article
• Authors: Ngan, Jun Wei; Caprani, Colin C.; Bai, Yu
• Is Part Of: Mechanical systems and signal processing. Volume 121(2019)
• Page Start: 373
• Page End: 388
• Publication Date: 2019-04-15
• Subjects: Model updating -- Finite element -- GFRP -- Zernike moment descriptors -- Mode shapes; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2018.11.027
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: We use ZMDs and spatial interpolation to classify mode shapes for model updating. A laboratory footbridge exhibiting a localized mode shape is used. The traditional MAC does not provide good model updating in this case. Characteristics of the approach are identified for practice. Abstract: The Modal Assurance Criterion (MAC) is a simple and powerful indicator to correlate experimental and analytical mode shapes. However, the MAC is not optimal for use in model updating of structures with localized mode shape features, since it is a global single index. Consequently, an alternative to the MAC is needed to optimally update numerical models of such structures for improved representation of structural dynamic behaviour. Developed from techniques in image processing and pattern recognition, Zernike moment descriptors (ZMDs) have been proposed as alternative correlation indicator in this study. They are highly sensitive to image features and so offer good promise to correlate mode shapes of relevant structures. This paper presents a framework for the use of ZMDs in model updating of structures exhibiting localized mode shapes. A particular example of such structures increasingly being used in civil engineering are glass fibre-reinforced polymer (GFRP) structures, such as buildings and footbridges. The proposed approach is applied to a full-scale pultruded GFRP footbridge with localized features to produce a more faithful numerical model. The effectiveness of both MAC and ZMD are considered. It is found that the use of ZMDs as the target response metrics improves mode shape correlation, giving improved FE representation of the GFRP footbridge. This paper informs on the applicability of ZMDs in FE model updating of structures with localized mode shapes and has recommendations for their use in practice.