报告时间：2024年12月13日（周五）上午10:00-12:00
报告地点：国重608会议室
报 告 人：朱信群 副教授（悉尼科技大学）
主办单位：国重实验室
---------------------------------------------------------------------------------------------------------------------
报告简介：
Bridges are deteriorated due to operational loads and environmental erosion, and they require proper inspection and monitoring to ensure their operation and structural safety. Structural damage of bridges is typically a local phenomenon. The local damage in a region trespassed by the vehicle can be accurately identified by the local responses collected. When the vehicle trespasses over the local damage region, it induces a local response of bridges. The local damage can be identified by analyzing time-varying characteristics of dynamic responses for bridges subjected to a moving vehicle. The accuracy of bridge damage identification mainly depends on vehicle-bridge interaction. The road surface roughness, and the wheel and road contact have a big effect on the identified results. Recently, the data-driven approach, especially deep learning-based approach, has attracted the interest of researchers and engineers. The traditional deep learning-based approach needs a large amount of training data. However, the data for damage scenarios is difficult to obtain in practice. In this study, a novel transfer learning-based approach is proposed for condition assessment of bridges under moving vehicles utilizing the time-frequency characteristics of vehicle-bridge interaction systems. Convolutional neural networks (CNNs) are used to extract discriminative features from the time-varying input data of vehicle-bridge interaction. The performance of the proposed method is evaluated using two cases: condition assessment of concrete bridges and prestress evaluation of prestressed concrete bridges. Effects of the noise level, vehicle speed, and sensor location on the predicted results are also studied. Numerical results show that the proposed method can precisely locate the damage on concrete bridges using only a single sensor on the bridge deck. The method has a great potential for practical application for bridge condition assessment in practice.
报告人简介：
朱信群副教授于2001年获得香港理工大学博士学位，现任职于悉尼科技大学土木与环境工程学院，在结构健康监测和状态评估、钢-混组合结构、先进传感与信号处理技术以及基于物理的机器学习等领域积累了丰富的研究经验，主持包括澳大利亚研究委员探索基金等20余个科研项目。担任美国土木工程师学会(ASCE)会员、澳大利亚结构健康监测委员会执行委员、美国土木工程师学会工程力学领域健康监测与控制委员会执行委员，《Progress in Engineering Science》执行主编、《Advances in Structural Engineering》副主编，《Advances in Bridge Engineering》等五本期刊编委。出版专著2部，发表论文260余篇，谷歌学术h-index为43，自2020年起入选全球前2%高被引学者，在学术界具有广泛影响力。
科技处
2024年12月6日