桥梁健康监测技术研究现状及展望

为进一步推进桥梁健康监测技术的发展，保障桥梁运营安全，依据近20年国内外桥梁健康监测（BHM）领域的学术研究现状，总结了BHM在系统及适用性、结构损伤监测算法、监测数据预处理、损伤结构安全预警及数字孪生技术方面取得的最新进展，确定BHM技术目前的研究热点和未来的发展方向。综合分析表明：在BHM系统及适用性方面，研究结构响应参数与健康指标的关联机制，研发长寿命非接触自动采集的智能传感装置，建立针对多源数据采集、传输、存储、分析、评价、预警于一体的自动化、网络化、智能化综合系统是重点研发方向；在结构损伤监测算法方面，设置针对异质场景的不同人工神经网络及修正方法选择建议集，针对多源信息流构建基于数据驱动与模型修正实时交互的多层级耦合智能算法是主要研究热点；在监测数据预处理方面，进一步研发基于深度学习的多源异构数据融合方法，建立复杂环境影响下的损伤结构动态信号提取算法，实现结构监测数据的精准分离是未来研究的热点；在损伤结构安全预警方面，研究重心集中于预警指标和预警体系的建立以及基于可靠度理论与监测数据的常规损伤安全评估，以结构监测数据反映总体力学行为并结合局部损伤的智能检测信息进行服役性能评价是未来的主要发展方向；数字孪生技术在BHM中尚属起步，将数字孪生技术融入多层级复合算法，建立结构多源异构大数据智能融合机制，形成数字联通、实时互动的智能化桥梁运维监测体系是重要发展方向。

Current Status and Prospects of Research on Bridge Health Monitoring Technology

To advance the development of bridge health monitoring (BHM) technology and ensure the safety of bridge operations, the current academic research status was investigated. The latest progress made by BHM in terms of systems and applicability, damage monitoring algorithms, data preprocessing, safety warning, and digital twin technology was reviewed. Then, current areas of active research and future development directions of BHM technology were identified. Comprehensive analysis shows that, in terms of the BHM system and its applicability, the correlation mechanism between structural response parameters and health indicators must be studied, and smart sensor devices with long-life noncontact automatic collection should be developed. Establishing an automated, networked, and intelligent integrated system for multisource data collection, transmission, storage, analysis, evaluation, and early warning is a key research direction. In terms of structural damage monitoring algorithms, methods of selecting different artificial neural networks and correction methods based on heterogeneous scenes are needed. Meanwhile, the construction of multilevel coupled intelligent algorithms based on data-driven and model correction real-time interaction for multisource information flow is a main focus of research. In terms of monitoring data preprocessing, further research on and development of multi-source heterogeneous data fusion methods based on deep learning are required. Dynamic signal extraction algorithms for damaged structures under the influence of complex environments must be established, and the precise separation of structure monitoring data must be realized. In terms of safety and early warning of damaged structures, early warning indicators and systems should be established to supplement conventional damage safety assessment based on reliability theory and monitoring data. The use of structural monitoring data to reflect overall mechanical behavior, combined with intelligent detection information of local damage for service performance evaluation, is an important development direction. Digital twin technology is still in its infancy in BHM. Integrating digital twin technology into multilevel composite algorithms, establishing a structured multisource heterogeneous big-data intelligent fusion mechanism, developing a digital interconnection, real-time interactive intelligent bridge operation, and maintenance monitoring system, are other important development directions.