In August 2007, the I-35W bridge in Minneapolis collapsed
during evening rush hour. Thirteen people died.
Hundreds of vehicles were on the bridge when it fell.
The investigation found that a gusset plate —
a critical connecting piece in the steel truss structure —
had been undersized from the original design
and had been accumulating stress for decades.
The bridge had been inspected. It received a rating
that flagged it as structurally deficient.
But structurally deficient in bridge inspection terminology
does not mean imminent collapse risk.
It means something needs attention at some point.
The problem is that periodic visual inspection,
which is still the primary method for most bridge assessments,
cannot tell you how fast a problem is developing
or how close a structure is to a critical threshold.
It tells you what is visible on the day someone looked.
The state of bridge inspection today
There are approximately 620,000 bridges in the United States.
Federal law requires they be inspected at least every two years.
Those inspections are primarily visual. An inspector walks the structure,
looks for visible cracking, corrosion, spalling concrete, section loss in steel.
They rate what they see on a numerical scale.
The rating informs maintenance priority and funding allocation.
Visual inspection is valuable. It catches a lot.
But it has a fundamental limitation that no amount of inspector skill overcomes.
It cannot see inside materials.
It cannot detect a fatigue crack that has not yet reached the surface.
It cannot measure how much steel remains in a rebar
that has been corroding from the inside for twenty years.
It cannot tell you how the acoustic properties of a concrete deck
have changed as alkali-silica reaction has been slowly
degrading the aggregate bond.
For that information you need acoustic and ultrasonic methods.
What structural health monitoring actually involves
Structural health monitoring is the practice of instrumenting a structure
with sensors that provide continuous or frequent data
on its physical condition and response to loading.
For bridges the sensor suite typically includes strain gauges
that measure how the structure deflects under traffic loads,
accelerometers that capture vibration response,
and increasingly acoustic emission sensors
that detect the stress waves produced by active crack growth
and other damage mechanisms.
The acoustic emission component is particularly valuable
because it is the only approach that detects damage
as it is happening rather than after it has already occurred.
When a fatigue crack in a steel girder grows by a fraction of a millimeter
under a passing truck load, it releases acoustic energy.
A sensor bonded to that girder can detect that release.
Over time, the rate of acoustic emission events from that location
tells you whether the crack is stable or accelerating.
That rate-of-change information is what periodic inspection cannot provide
and what makes the difference between knowing a crack exists
and knowing whether it requires urgent intervention.
Acoustic Testing Pro builds the acoustic emission sensors
and monitoring infrastructure suited to this kind of
continuous structural surveillance —
https://acoustictestingpro.com/sensor-technologies/acoustic-emission-sensors/
— the foundation of a monitoring program that gives engineers
real data on structural condition between inspection visits.
The data infrastructure challenge for bridge monitoring
A single instrumented bridge generates continuous data streams
from potentially dozens of sensors.
A state department of transportation managing hundreds of bridges
faces a data management problem that requires
centralized platforms capable of aggregating multi-site data,
automated analysis that flags anomalies without requiring
an engineer to manually review every sensor reading,
and reporting that connects to asset management workflows.
The edge-to-cloud architecture that works for industrial facility monitoring
applies here with some modifications for the outdoor, distributed environment.
Sensors need to operate across wide temperature ranges.
Power is often not available at sensor locations,
requiring battery or energy harvesting solutions.
Connectivity in rural bridge locations may be limited,
requiring local data buffering and opportunistic transmission.
None of these are unsolved problems but they require
more careful system design than a controlled industrial setting.
What has changed in the last five years
Sensor costs have dropped to the point where instrumenting
a bridge with a meaningful acoustic emission network
is no longer reserved for research projects and flagship structures.
Wireless sensor nodes eliminate most of the installation cost
that made wired monitoring systems prohibitive for routine deployment.
Cloud platforms designed for structural health monitoring
have matured to where they handle the data management problem
without requiring custom infrastructure for every deployment.
And the political will to invest in infrastructure monitoring
has increased following several high-profile failures —
not just Minneapolis but the Fern Hollow Bridge collapse in Pittsburgh in 2022
and the Francis Scott Key Bridge in Baltimore in 2024,
which while caused by a vessel strike rather than structural failure
raised public awareness of bridge vulnerability significantly.
The technology to do better bridge monitoring exists.
The funding frameworks to deploy it at scale
are slowly catching up to what the technology makes possible.
A question worth sitting with
Every bridge you have crossed in the last month was last fully inspected
at some point in the past two years.
Some of them have sensors providing continuous data.
Most of them do not.
The ones that do not are being managed on the assumption
that nothing significant has changed since the last inspection.
That assumption is usually correct.
The times it is not correct are the ones that make the news.
What would it take, in your view, to make continuous structural monitoring
standard practice for critical infrastructure rather than the exception?
Top comments (0)