We've been building bridges ever since the wheel was invented, and a method for crossing over a rocky-bedded river ford was needed. The earliest bridges might have been timber contrivances, but these were soon replaced by sturdier stone artefacts, cemented together by early lime and mortar, some of which are still standing today.

Those structures, of course, were massively over-engineered by modern standards, and the loads they had to carry are a mere fraction of those borne today. But even they are subject to natural decay over time. An evolving infrastructure has caused many older bridges to be bypassed, which has led to them being dismantled.

Modern bridge building, spanning the last couple of hundred years, uses a wider range of technologies, each of which is subject to its own particular form of decay. They are also constructed to satisfy different criteria, such as cost, strength, location, and design elegance - the latter always seeking a balance between the lightness of a structure and its strength.

We are not unfamiliar, therefore, with the problems that beset each type of bridge, and the resources that are needed to both detect problems and to rectify them. The problem, though, is that there are many thousands of bridges that need to be maintained both here and in the United States. The scale of the problem is exacerbated by the fact that maintenance schedules can only be established by noting the scale of a particular problem and the rate of its decline through a series of repeated visits. We are not only concerned with legacy bridges, either, as every bridge that has been designed, analysed, and constructed requires a schedule of asset inspections to ensure they continue to operate at their designed performance level.

BENTLEY WORK FEATURED AT MICROSOFT IGNITE
Dan Vogen, vice president of road and rail asset management at Bentley Systems, was part of a case study panel at Microsoft's Ignite conference that explored the need to change how we deal with bridge inspection compliance and other issues that owners of bridge assets face. "Bridge inspection and management have stagnated over the years, and the typical approach is that whether we're talking about governmental oversight and legislation, or just as an asset owner, a scheduled inspection involves inspectors going into the field, using a range of tools and other mobile applications, producing photographs, videos, and audio notes to record information about the state of an asset so that we can compare notes on its condition over a period of time," Vogen said. "The production of such documentation and the submission of inventory conditions has been going on for many years."

Vogen continued, "There has been some evolution in the process, perhaps a recognition that we can do better, such as differentiating between a new bridge and one that is in an area with difficult physical conditions - salt, sea water, or heavy weather - requiring a different approach and more frequent inspections. In the United States we are faced with the Federal Highway Administration's new mandate for bridge inspections - the National Bridge inventory guidelines. These new standards will move inspectors away from rigid schedule-based to performance-based inspections."

MICROSOFT'S HOLOLENS AND DIGITAL TWINS
Vogen described the two critical processes that has allowed bridge inspectors to become smarter in the way they handle bridge inspections. The first is the use of all available technology to view and record the condition of a bridge at a particular point in time, so that a digital twin can be made of the structure and its condition. The second is that Microsoft's HoloLens technology can be used to create an immersive view of the bridge using photogrammetry processed through Bentley's ContextCapture, so that bridge inspectors can view and visually interpret data within their own office environment.

The second point is important, as it enables successive inspections to be made in the field more frequently by less experienced teams, producing accurate data from pre-programmed flights to be fed into a digital twin timeline. Updated digital twin models can then be viewed and analysed in the office by immersing the bridge inspector in the model to examine the extent of degradation over any part of the asset over a specific period.

A digital twin, creating a virtual copy of an existing structure, is composed of photos and video footage from drone and other surveys using Bentley's ContextCapture technology, in situ sensors and cameras installed to monitor stresses and strains, and other information.

I asked about a particular problem that we have locally - the rusting of internal strands of wire in a cable stay bridge, apparently invisible during physical inspections, and Vogen explained that the problem is not uncommon and that magnetic flux leakage cable scan technology is now being used to detect such invisible issues.

ContextCapture is able to process the imagery captured from a drone flight or any other method of producing photogrammetry to make a reality mesh, somewhat highly triangulated in the case of bridges, that could consist of tens of millions of polygons. These polygons need to be rendered to create the immersive model. The reality mesh, however, is just one aspect of a bridge's digital twin. "We need to add in any available extra information that we have," Vogen explained. "For a steel beam, for instance, we would want to include its material strengths so that we can perform load rating and other analyses.

"The photogrammetry and video provided to ContextCapture allows us to create a virtual model onto which we can overlay so much other information," he added.

"To demonstrate how much more we can now do, consider corrosion or cracking where you can get information from sensor feedback, which can be used to provide all kinds of overlays that would describe or simulate the levels of corrosion found. That's been easy for many years, but with immersive reality you can move around the model, zoom in, and inspect the corroded element closely and even take measurements - a full three-dimensional interaction."

Vogen continued, "You are not watching it on a monitor, either. HoloLens will create a life-size image for you, and you are there, on that bridge. Instead of moving the Creeper vehicle over a 10-foot section of the bridge, you can go into the model and walk 10, or a 100 feet, seamlessly."

Even drones have limitations. They provide a primitive limited field of vision in spite of their mobility, but full context spatial awareness is only possible when the drone virtual imagery is supplemented by the augmented reality of the full digital twin model.

Using a digital twin has other benefits too. Instead of taking two pictures of the crack in a bridge's concrete base two years apart, with different inspectors who might not quite capture the same critical detail, we can use the digital twin to locate the exact spot and provide a true geometric measurement of any further deterioration. Furthermore, artificial intelligence and machine learning algorithms can be used to automatically identify such conditions and defects during an inspection and highlight them for attention.

Vogen explained that pre-inspection set-ups allow the footage from a drone's ContextCapture survey to look for such issues, using trained learning algorithms. The software can then suggest areas for inspection - identifying what you might want to focus on from an AI/ML basis, which can be highlighted for closer visual inspection on a subsequent ContextCapture mission or by visiting the bridge.

Prediction is also a big part of performance-based inspection, Vogen said. "With improved analytics we can have a really good geometric understanding of a bridge's crack progression or corrosion levels and allow us to take preventive measures."

"We already know "how bridges deteriorate," Vogen concluded. "What we are now able to do is to inspect them remotely, more frequently, and with greater accuracy."

www.bentley.com