The core task of structural health monitoring (SHM) is upgrading tunnels and bridges from "periodic checkups" to "continuous ECG monitoring": strain, temperature and vibration are collected continuously along the entire structure, so crack propagation and surrounding-rock settlement are quantified before they worsen. Distributed Fiber Strain Sensing (DTSS-BOTDR) delivers ±5με strain accuracy and ±0.5°C temperature accuracy, and has been deployed at scale in a record-cross-section highway tunnel. This article decodes the technical principle, the spec meanings and field deployment experience.
Why Transport Infrastructure Needs Structural Health Monitoring
When a highway tunnel or a major bridge fails, the response window is measured in hours and the loss in the hundreds of millions. A large batch of infrastructure built over the past few decades is entering its middle age; material aging, accumulated loading and geological activity act together, and risk is no longer hypothetical.
Three real pressures are forcing the demand for monitoring. The first is aging. Concrete carbonation and rebar corrosion are gradual processes; by the time cracks are visible to the eye, the structure has often already passed a critical threshold. The second is extreme climate. Heavy rain, freeze-thaw and temperature swings repeatedly draw surface water in, surrounding rock softens on contact with water, and the stress pattern of the tunnel lining shifts quietly. The third is overload and heavy loads. Heavy-truck axle loads chronically exceed design values, fatigue damage accumulates on the deck, and the stress curve of the girder is the thermometer of structural health.
These variables point to one conclusion: infrastructure safety can no longer rely on inspections every few years plus visual judgment; structural safety monitoring is the inevitable choice.
The Limits of Traditional Monitoring: Plenty Where You Can See, Little Where You Can't
The traditional approach relies mainly on manual inspection, supplemented by point-type instruments such as level gauges, total stations and displacement meters. In long, narrow, enclosed structures like tunnels, the limits are obvious.
The inspection cycle determines the risk blind window. Tunnels are typically inspected every ten days, monthly or quarterly; between two rounds the structural state is a complete blank. Point-type instruments cover only their mounting positions; a secondary-lining crack can sit tens of meters from the nearest sensor and go unnoticed. Bridges are easier to instrument, but the number of monitoring points is constrained by cost — a few hundred points covering a multi-kilometer deck, and the sparseness speaks for itself.
There is also a commonly ignored problem: point sensors fail on their own. Connector oxidation, power interruption, cable damage — if maintenance can't keep up, the monitoring network "shrinks" year by year, and the O&M unit often has no idea. The distributed fiber approach turns the sensor into the fiber itself, so the failure mode shifts from "tens of thousands of points silently broken" to "one fiber — broken or not, visible at once."
The Fiber Sensing Approach: Strain, Temperature and Vibration in Three Coordinated Channels
Fiber sensing does structural monitoring by laying the fiber along the structure and letting the measured object "speak for itself." For a tunnel, the fiber runs along the inner surface of the secondary lining or through the structure body, so the entire line is within measurement range — the concept of "point coverage" simply disappears.
Three physical quantities work together, none dispensable:
- Strain channel (DTSS-BOTDR). Measures the structure's deformation along the fiber axis. Crack propagation, girder bending and surrounding-rock squeeze all show up as abnormal segments on the strain curve, with ±5με accuracy and measurement time down to 1 second.
- Temperature channel (DTS). Captures water seepage, fire and ambient temperature-field changes, cross-checking the strain channel — when a seepage-cooling zone overlaps a strain concentration segment, the problem location becomes unambiguous.
- Vibration channel (DAS). Senses dynamic excitation such as vehicle loads, blasting and earthquakes, used for rapid damage screening after an impact event.
The three channels share one fiber trunk, separated by wavelength-division and time-division multiplexing. In engineering terms, a single monitoring cable performs multi-parameter acquisition, cutting both installation cost and later maintenance.
A Record-Cross-Section Tunnel Case: The Practice Template for Structural Monitoring
Nothing is more convincing than putting the scheme into a real project — in this case, a highway-tunnel structural health monitoring project. At the time of completion, this tunnel was one of the largest cross-section highway tunnels. A large cross-section means high span, complex surrounding-rock conditions and high construction risk, so the structural stress state demands extremely high monitoring sensitivity.
The project used distributed fiber sensing for full-cross-section monitoring of the tunnel structure. Strain and temperature data stream back continuously, and abnormal segments are marked directly on the map with position and magnitude — duty staff can locate problem segments without understanding the sensor principle at all. After the tunnel opened to traffic, its structural state went from "unknown" to "known"; maintenance scheduling gained a data basis instead of relying on experience and convention.
The technical difficulty of this class of project lies in the laying process. The lining curvature of a large-cross-section tunnel varies widely and fiber bend radius is limited; ensuring the fiber couples tightly to the structure without incurring extra loss from bending is where delivery capability is tested far more than equipment selection. This is why such projects usually require a supplier that has both equipment capability and engineering execution capability.
How to Read the Monitoring Metrics: The Engineering Meaning of Strain, Cracks and Settlement
For an operating unit, the monitoring system must answer not "how much data was collected" but "is the structure still sound." A few key metrics need to be read correctly.
Strain is the leading indicator of structural state. Strain is deformation per unit length; με is microstrain, one με equal to one part per million. ±5με accuracy means a length change on the order of 0.05 mm over a 10 m structure can be resolved — sensitive enough to capture the stress redistribution that precedes crack initiation. A strain curve climbing steadily over a segment is often more valuable as a warning than a photograph of a crack.
Cracks are read as a distribution, not a single point. A point-type crack meter watches only its own one crack, while the fiber strain curve shows the strain distribution over the entire segment, with the crack position appearing as a peak or step in the curve. Reading adjacent segments together tells you whether a crack is isolated or spreading into a connected band.
Settlement must be judged against geological conditions. Settlement data in a tunnel monitoring report cannot be interpreted in isolation from the surrounding-rock grade and groundwater conditions; the same settlement rate means very different danger levels in water-rich soft rock versus hard rock. The system should deliver "data + criteria," not bare data.
Fiber Monitoring vs. Manual Inspection vs. Traditional Sensors
The three approaches each suit different scenarios; they are clearest side by side.
| Dimension | Manual inspection | Traditional point sensors | Distributed fiber monitoring |
|---|---|---|---|
| Coverage | Spot checks along the route | Mounting positions | Continuous, line-wide |
| Real-time | Periodic | Near real-time | Second-level real-time |
| Fault self-check | None | Hard to detect | A broken fiber is known instantly |
| Per-point cost | Manpower rises with cycle | More points, more cost | Along-line coverage, low marginal cost |
| Suitable structures | All | Bridges, buildings | Linear structures: tunnels, bridges, galleries |
Long-Term O&M Value: Three Changes Brought by Data Accumulation
The return of a monitoring system does not come on commissioning day, but three to five years later.
First, maintenance shifts from "repair on schedule" to "repair on need." With a continuous strain and temperature baseline, normal structural fluctuation and abnormal change are distinguishable, and repair funds are spent where they matter. Second, life assessment becomes evidence-based. Years of continuous data reflect the structural degradation trend better than a single inspection, providing quantitative support for major-repair and strengthening decisions on bridges and tunnels. Third, emergency response accelerates. After a sudden geological hazard, the system delivers a line-wide damage screening result within half an hour — an order of magnitude faster than re-organizing a dedicated inspection.