Next-generation Indirect Bridge Health Monitoring Technology

北美的桥梁基础设施正在老化，并承受日益恶劣的天气条件。最近的基础设施报告卡显示，美国三分之一的基础设施面临快速恶化的风险，其中约24万座桥梁至少有50年的历史。平均每天有1.88亿人次通过这些结构缺陷的桥梁，需要价值约1230亿美元的修复。传统的桥梁健康监测（BHM）涉及昂贵且耗时的接触式传感器安装。一种有吸引力的替代方案是间接BHM（iBHM），其中传感器安装在经过的车辆中，从而实现移动的桥梁检查。然而，现有的iBHM有几个挑战，如传感器-车辆-桥梁相互作用，缺乏车辆配置的指导方针，需要一个大型的图像数据库，以及桥梁健康的可视化不足。拟议研究的长期目标是开发一种强大的iBHM技术，该技术将无需桥梁上的任何物理仪器和大型图像数据库，以实现具有成本效益的桥梁检查和维护。长期目标将通过以下五个短期目标来实现，这些短期目标分为三个紧迫主题。主题1将开发使用移动的传感器的全局损伤检测技术，方法是：i）开发鲁棒的诊断算法，以解决车辆测量中车辆-传感器-桥梁相互作用的挑战; ii）开发测试车辆的最佳配置。主题2将通过iii）一套新颖的生成对抗网络和iv）通过增强图像和视频的数据融合技术，使用较少的训练数据库生成强大的局部损伤评估技术。主题3将开发一种新的可视化技术，以跟踪损坏的进展，并使用v）一种新的桥梁信息模型优化检查频率。该提案将首次为测试车辆及其操作制定指导方针，这是准确和自主的iBHM所需的。所提出的数据融合技术和基于视觉的多损伤分类器将不需要大量的图像，从而为美国数千座受损桥梁提供具有成本效益的iBHM系统。在与美国土木工程师协会（ASCE）、加拿大土木工程协会（CSCE）、安大略交通部（MTO）和安大略专业工程师（PEO）的技术委员会协商后，拟议研究的结果将产生BHM规范，为北美桥梁检查提供最佳实践和指南。一个由3名博士，2名MESc和5名BESC学生组成的多元化团队将接受培训，为这一新兴的BHM技术做出贡献，并在工程界推广EDI。预计这项研究将为我们的公民带来安全的交通基础设施，并为政府节省数十亿美元的成本。

Bridge infrastructure in North America is ageing and enduring increasingly adverse weather conditions. The recent infrastructure report card indicated that one-third of America's infrastructure is at risk of rapid deterioration, with ~240,000 bridges at least 50 years old. On average, 188 million daily trips occur across these structurally deficient bridges, which requires rehabilitation worth ~$123 billion. Traditional bridge health monitoring (BHM) involves the costly and time-consuming installation of contact sensors. One attractive alternative is indirect BHM (iBHM), where sensors are installed in a passing vehicle, enabling mobile bridge inspection. However, existing iBHM has several challenges, such as sensor-vehicle-bridge interaction, lack of guidelines for vehicle configuration, the need for a large image database, and inadequate visualization of bridge health. The long-term objective of the proposed research is to develop a robust iBHM technology that will be free of any physical instrumentation on the bridge and a large image database to achieve cost-effective bridge inspection and maintenance. The long-term objective will be achieved through the following five short-term objectives organized in three pressing themes. Theme 1 will develop global damage detection technology using mobile sensors by i) developing robust diagnostic algorithms to solve the challenges of vehicle-sensor-bridge interaction in vehicle measurements and ii) developing the optimal configuration of the test vehicle. Theme 2 will produce robust local damage assessment technology using a fewer training database through iii) a suite of novel generative adversarial networks and iv) data fusion technology by augmenting images and videos. Theme 3 will develop a novel visualization technology to track the progression of damage and optimize inspection frequency using v) a novel bridge information modelling. This proposal will, for the first time, develop guidelines for the test vehicle and its operation, which are required for accurate and autonomous iBHM. The proposed data fusion technology and vision-based multi-damage classifier will be free of the requirement for a massive number of images, resulting in cost-effective iBHM systems for thousands of impaired bridges in America. In consultation with technical committees of the American Society of Civil Engineers (ASCE), the Canadian Society for Civil Engineering (CSCE), the Ministry of Transportation of Ontario (MTO) and the Professional Engineers of Ontario (PEO), the outcome of the proposed research will result in BHM codes to provide best practices and guidelines for bridge inspection in North America. A diverse team of 3 PhD, 2 MESc, and 5 BESc students will be trained to contribute in this emerging BHM technology and promote EDI in the engineering community. It is anticipated that this research will lead to safe transport infrastructure for our citizens and billions of dollars of cost savings for governments.