Speaker: Prof. Siu-Kui Au

Title: In-situ Dynamic Characteristics of Structures

Date: Thursday, April 25, 2019       Time: 15:30-17:00

Venue:T5305 (Teaching Building 5, Room 305)

About the Speaker：

Dr Au is professor in the School of Civil & Environmental Engineering at Nanyang Technological University, Singapore; Honorary Visiting Professor with University of Liverpool and Chutian Chair Professor (visiting) with Wuhan University. He performs fundamental and applied research in engineering risk methods (see monograph) and structural health monitoring. Recent work focuses on theory and practice of Bayesian operational modal analysis for full-scale structures (see monograph). A chartered civil engineer in Hong Kong, Dr Au is recipient of IASSAR Junior Research Prize, Nishino Prize, JSPS Fellowship and Tan Chin Tuan Fellowship. He chaired the ASCE Dynamics Committee (2014-2017); is associate editor of Structural Health Monitoring and ASCE-ASME Journal of Risk & Uncertainty Analysis in Engineering A & B; and editorial board member of Earthquake Spectra and Probabilistic Engineering Mechanics. Dr Au has published 100+ journal papers with 3800+ independent citations (Web of Science).

Abstract：

The dynamic characteristics, or 'modal properties', of a structure include primarily the natural frequencies, damping ratios and mode shapes. They are the interface between the mechanical properties of a structure (e.g., stiffness, mass) and its response under dynamic loads such as wind and earthquakes. New design concepts, tightened performance criteria and targets for cost-effectiveness have led to an increasing demand for in-situ modal identification, providing a factual basis for decision making in vibration control, retrofitting or health monitoring projects. Depending on budget and constraints, the modal properties may be determined using free (no loading), forced (known specified loading) or ambient (unknown broadband loading) vibration tests. The last category, also known as ‘operational modal analysis’ (OMA), has attracted much attention in both theory and practice as they provide an economical means for in-situ testing. Under challenging field environment and in the absence of loading information, however, issues on reliable identification and test planning become especially relevant. This talk will give an overview of Bayesian approach for operational modal analysis, which offers a fundamental means for identifying modal properties as well as quantifying their uncertainty consistent with modelling assumptions and data information. Recent developments have made implementation practical even in construction sites. As one of the latest advances, beyond the ability to calculate uncertainty for given data, ‘uncertainty laws’ have now been discovered that allow one to understand how identification uncertainty depends on test configuration (e.g., number, location and quality of sensors). They reveal the achievable precision of OMA and provide a strong scientific basis for planning ambient vibration tests. Challenges in OMA and future directions will also be discussed.