Work

A Multilevel Inverse Analysis Framework for Damage Identification Leveraging Reduced-Order Modeling of Sensor-Structure Interaction

Year: 2025

Type: article

Abstract: Identifying small-sized damage in its early stage allows one to take appropriate action in a timely manner to minimize the maintenance concerns and thus becomes imperative in structural health monitoring (SHM). While the low-frequency vibration measurements are widely used in damage identification due to convenient data acquisition, they are insensitive to the small-sized damage. High-frequency response on the other hand is capable of reflecting the small-sized damage effect. However, a high-dimensional finite element model oftentimes is required to enable high-frequency response evaluation. As a result, the inverse analysis based upon such model to conduct damage identification will become computationally expensive or even intractable. In this research, we establish a multilevel damage identification scheme to decompose a single complicated task into a set of consecutive tasks that are computationally amenable. It encompasses a coarse-level damage localization between substructures, followed by a fine-level damage identification through inverse optimization in the substructure localized with damage occurrence. High-frequency impedance/admittance information acquired from embedded piezo-electric transducer is employed. This framework starts from the reduced-order modeling of sensor-structure interaction with component mode synthesis (CMS) technique which can yield the initial damage location estimation amongst the substructures by means of sampling-based similarity analysis that jointly considers multiple objectives. Subsequently, inverse analysis leveraging multi-objective optimization is conducted on the substructure flagged with damage occurrence to pinpoint damage location and severity accurately. Systematic case studies are carried out for validation and discussion.

Source: Journal of Computational and Nonlinear Dynamics

DOI: 10.1115/1.4070201

Authors: Kai Zhou, Yang Zhang, Jiong Tang

Institutions: Hong Kong Polytechnic University, University of Connecticut

Language: English

Topic: Structural Health Monitoring Techniques

Subfield: Civil and Structural Engineering

Field: Engineering

Domain: Physical Sciences

SDG: Industry, innovation and infrastructure

Open Access status: hybrid

Funders: National Aeronautics and Space Administration, Nuclear Safety and Security Commission, Hong Kong Polytechnic University