New data-driven approaches may improve bridge safety monitoring

Bridges provide vital links in urban and remote Canadian communities. Canada has more than 50,000 publicly owned bridges, as reported in a 2020 Public Infrastructure Survey. Many are reaching the end of their lifespan: 11% (or 5,865 bridges) were reported to be in poor or very poor condition. While the design life of a bridge is generally 50 to 75 years, they can safely last longer with good monitoring and timely maintenance, which requires a significant public investment.

The National Research Council of Canada's AI for Logistics program is addressing these challenges by bringing our researchers together with experts from Housing, Infrastructure and Communities Canada, Esri Canada and the University of Manitoba. They are developing improved methods for life-long monitoring of infrastructure conditions and decision support for maintenance investment. To conduct their research, they are using a highway bridge in Manitoba.

This AI for Logistics project is entitled, "A Data-Driven Approach to Enhance Transportation Infrastructure Resiliency through Bridge Monitoring." The research is sponsored by Defence Research and Development Canada's Canadian Safety and Security Program (CSSP), which is managed in partnership with Public Safety Canada. The project to develop this important technology is valued at more than $2 million, including $1.4 million from the CSSP.

A key challenge area of the CSSP is research and development to reduce the risks to Canadians from low-frequency, high-impact events. These rare events with significant consequences span a wide range including pandemics, tsunamis, terrorist attacks, or catastrophic bridge failures from structural deficiencies and aging infrastructure. For example, the catastrophic bridge failure on the Concorde overpass in Laval, Quebec on September 20, 2006 resulted in 6 injuries and 5 fatalities. A subsequent review found improper inspection, deterioration of concrete and cracks all contributed to the bridge collapsing.

The impact of climate change and extreme weather on bridge performance and health

While catastrophic bridge events are not new, climate change and extreme events can have an increased impact on the structural health and performance of bridges. Risks and vulnerabilities to the structures include temperature changes and natural disasters, such as floods and other extreme weather-related events.

Weather extremes and access challenges in Canada's remote and northern regions can limit opportunities for bridge inspection and can involve greater risks for inspectors attempting to access exposed bridge sections. However, the elevated risk of climate impacts on bridge health in these areas make regular inspection increasingly important.

Two distinct monitoring technologies are being used and compared in the project: satellite-borne radar sensors and in-situ sensors attached to the surface of structural elements of the bridge to assess strain, acceleration and temperature.

Dr. Chaouki Regoui, from our Digital Technologies Research Centre, leads the AI components part of the project. Dr. Daniel Cusson, from our Construction Research Centre, provides leadership in the areas of satellite-based bridge monitoring using Synthetic Aperture Radar (SAR) imagery. SAR imagery is a type of active data collection where a sensor produces its own energy and then records the amount of that energy reflected back after interacting with the Earth or the built environment.

Professor Aftab Mufti of the University of Manitoba's Structural Innovation and Monitoring Technologies Resource Centre, in collaboration with Professor Douglas Thomson and Professor Jonathan Regehr, leads the analysis of traffic and environmental influences on bridge deterioration. Esri Canada is providing Geographic Information System expertise and tools. The Data Science Division at Housing, Infrastructure and Communities Canada is developing models to explore the effects of bridge disruptions on socio-economic indicators.

The lessons learned from studying the highway bridge in Manitoba are expected to be applicable to the monitoring of similar bridges elsewhere in Canada. This will enhance the potential for remote bridge monitoring and reduce physical and economic risks related to bridge failure. One element of this is the team's plan to develop an AI-based bridge health predictor that governments could use to support decision-makers in their operation and investment in bridge maintenance, renewal or replacement.

Building a data vault to help keep infrastructure secure

This project also contributes to developing our AI for Logistics program's integrated data repository (data vault). The data vault contains a wide range of logistics-related information, including infrastructure and contextual data, which is important for studying the dynamics of traffic and environmental patterns on infrastructure such as bridges. This project will contribute to the data vault by adding on-board analytical and visualization capabilities, enhancing its usefulness for research involving data analytics and machine learning.

Margaret McKay, the program leader for AI for Logistics, is hopeful that this project and others like it will make a big difference for Canadians. "The knowledge and tools that this project provides will help governments ensure bridge safety in all regions, no matter how challenging the geography or climate. That is a big win for all of Canada," she says.

Certainly, this project has the potential to help keep Canadian infrastructure more secure and to reduce risk and disruptions for those who use bridges in their daily lives.