工程可靠性与随机力学国际联合研究中心

2023年第12期（总第74期）学术报告

工程力学研究中心第21期报告

文远讲坛270期 (线上)

报告主题

TOPIC

既有混凝土结构的贝叶斯评估

Bayesian assessment of existing concrete structures

报告人

SPEAKER

   Prof. Robby Caspeele

比利时根特大学教授，高端外国专家

报告时间

TIME

2023年12月14日（周四）下午15:00-16:00

线上会场

ZOOM

Zoom ID: 89033225549

Password: 300207

主持人

CHAIR

陈建兵教授、彭勇波教授

报告内容

Abstract

   An increasing amount of the current international activities in the construction sector are oriented towards the assessment of existing structures and the associated repair and upgrading interventions. Although the last decennia significant progress has been made to analytically and numerically simulate the degradation process of concrete structures and their associated structural performance prediction in time, structural engineers are challenged to deal with the large uncertainties involved. Besides a typical lack of information, the degradation processes are associated with large uncertainties and when executing additional investigations and measurements, significant model and measurement uncertainties play a dominant role in the reliability-based performance prediction. Bayesian updating techniques provide a unique engineering tool to adequately combine available information, enabling to make inferences where classical statistical approaches fail. After revisiting some basics on Bayesian approaches, it is explained and illustrated how Bayesian updating techniques provide an undeniable advantage in several steps of the assessment process. Degradation models whose parameters generally are difficult to find consensus about, can gradually be updated when more information becomes available. Variables in structural reliability calculations can be updated on the basis of direct as well as indirect measurement data. Sources of information contained in different types of measurements or even visual inspections can be combined in order to make much more informed inferences and to reduce uncertainties in the reliability-based performance assessment, compared to when such data is analysed separately. Semi-probabilistic assessment approaches can be adjusted on the basis of updated uncertainties. And finally, the positive effect of quality control can be taken into account when assessing the structural reliability. The integration of these approaches into the practical assessment process is proven to be feasible through several case studies and an outlook is given on future engineering challenges to integrate such approaches further in the life-cycle assessment of existing concrete structures.

报告人简介

Speaker Bio

Robby Caspeele is professor in the field of Structural Reliability at Ghent University (UGent) since 2013 and one of the technical directors of the Magnel-Vandepitte Laboratory for Structural Engineering and Building Materials. His main areas of expertise relate to (1) full- and semi-probabilistic structural reliability methods and Bayesian updating techniques in relation to performance-based life-cycle design and analysis of new and existing concrete structures, (2) structural robustness and membrane action (mainly) in concrete structures in relation to design against progressive collapse as a result of foreseen and unforeseen events such as fire, extreme loading, terrorist attack, degradation and human errors and (3) structural assessment of existing structures in relation to time-dependent degradation, combining numerical structural analysis approaches with Bayesian updating techniques, decision making for structural interventions and lifetime performance optimization. He is author of more than 70 publications in peer-reviewed journals and over 190 papers in international scientific conference proceedings. He is also strongly involved in international committees and working groups related to the development of technical guidelines and standards. Among others, he is a board member of the Joint Committee on Structural Safety, co-convenor of fib Commission 3 on existing structures and chair of the Belgian mirror committee for the Eurocodes.