Background: Why Traditional Point Sensing Is Hitting a Ceiling
In the field of bridge and tunnel monitoring, we have discussed FBG sensors many times — high sensitivity, immunity to electromagnetic interference, long service life, with a single sensing point achieving ±1µε accuracy and millisecond-level response. But FBG is, at its core, a point sensor: one channel corresponds to one measurement point.
Once a project scales up, the problems begin. For a large cable-stayed bridge, dead-load monitoring plus live-load testing can easily require more than 200 FBG sensing points, each needing fusion splicing and individual calibration. With so many fibers on site during construction, attenuation accumulates within channels and management difficulty doubles — by the time the system is commissioned, the schedule has already slipped by two weeks. More critically: point sensing can only tell you "this point has failed," but structural deterioration in bridges rarely starts from a single well-defined point — crack propagation and strain redistribution caused by rebar corrosion develop continuously along the member.
This is where Distributed Fiber Optic Sensing (DFOS) comes in.
Three Scattering Mechanisms, Three Ranges: Choosing the Right Technology Route Is Key
DFOS uses the optical fiber itself as the sensing medium, detecting strain, temperature or vibration along the line by analyzing light scattered back within the fiber. Depending on the scattering mechanism, mainstream technologies fall into three categories:
Rayleigh scattering — high-frequency vibration, high precision over short distances
Based on Optical Time Domain Reflectometry (OTDR) or Optical Frequency Domain Reflectometry (OFDR). OFDR offers the highest spatial resolution of any distributed sensing technology today — down to the 1mm level with ±1µε/M accuracy — which is why we used OFDR for the 3D reconstruction of the strain field on a cable-stayed bridge model: during loading, the color gradients in the software matched the mechanical analysis results almost point for point. However, OFDR's sensing distance is limited, generally within 100 meters, making it suitable for small and medium bridges or fine-grained monitoring of critical local regions.
Brillouin scattering — long-range strain, dual temperature + strain parameters
BOTDR (Brillouin Optical Time Domain Reflectometry) and BOTDA (Brillouin Optical Time Domain Analysis) are the most common long-range solutions in bridge and tunnel monitoring. Strain accuracy is about ±20µε, spatial resolution 1–2 meters, with a range of tens of kilometers. Single-ended BOTDR is particularly suitable for retrofit projects on existing structures, since no double-ended fiber routing on the bridge is required; BOTDA offers higher accuracy (±15µε) but requires a double-ended loop, making installation somewhat more complex. In practice, a common combination is BOTDA for strain and DTS (Distributed Temperature Sensing) for temperature separately.
Raman scattering — pure temperature sensing, long range
Distributed Temperature Sensing (DTS) infers temperature from the ratio of Stokes to anti-Stokes light intensity in Raman scattering, independent of strain. The range can exceed 30 kilometers, with ±0.1°C accuracy and 1–2 meter spatial resolution. It is particularly well suited to tunnel fire monitoring and ground temperature field analysis.
In a 2025 study published by Germany's BAM Federal Institute for Materials Research and Testing, DAS (Distributed Acoustic Sensing, based on coherent detection of Rayleigh scattering) was used in destructive testing on a decommissioned bridge — cracks were deliberately induced, and the fiber "listened" to the acoustic emission signals generated by crack propagation, achieving meter-level localization entirely wirelessly and contactlessly. This ability to turn a fiber into a string of microphones is something traditional point sensors simply cannot do.
From Data to Decisions: The Digital Twin as the Last Piece of the DfOT Puzzle
Distributed sensing solves the "can we measure it" problem, but "can we understand it" still requires a digital twin to translate the data into engineering language.
A 2026 study published in the Journal of Infrastructure Intelligence and Resilience presented a complete closed-loop framework: the physical layer's wireless sensor network (acceleration, strain and displacement sensors) collects real-time data; the data layer pushes it via MQTT to edge gateways; the model layer uses a finite element model (FE Model) calibrated with measured data as the digital twin baseline; the computation layer performs modal parameter identification and anomaly detection; and the decision layer outputs alerts and maintenance recommendations.
Within this framework, distributed fiber sensing provides continuous strain field data and is the most important data source for updating the digital twin model — every load test and every abrupt temperature change triggers another closed-loop calibration of the model with fresh measured data. Essentially, the core logic of this system is: a digital twin is not built once; it is "fed" into existence with real data.
The Spyckstraße Bridge case (Germany, built in 1976, a three-span steel-concrete composite girder bridge) illustrates this best. The research team used short-term health monitoring data (vibration, displacement, modal frequencies) combined with a genetic algorithm to iteratively calibrate the digital twin model, keeping the difference between numerical and measured modal frequencies within 3%. After calibration, they ran a parametric analysis of prestressing strand corrosion — reducing the effective prestress by just 10% pushed the tensile stress at the bottom of the midspan section beyond the design limit, and the modal frequencies also showed a measurable downward shift. This is not something you can simply calculate — it is a conclusion you can only trust after the model has been calibrated with measured data.
Three Real-World Cases: Where DFOS + Digital Twins Land
Case 1: Semmering Base Tunnel, Austria — DFOS guiding shotcrete curing
Graz University of Technology embedded a 60-meter OFDR sensing cable in the tunnel's initial lining, positioned at segment joints, corresponding to roughly 6,000 monitoring points with 10mm spatial resolution and 1µε/M strain resolution. The system continuously monitored strain evolution during shotcrete curing throughout construction and identified early microcrack initiation locations that traditional point sensors could not capture at all — the initial crack width at that location was only 0.05mm, invisible to the naked eye, and it was not discovered by visual inspection until it had grown to 0.3mm three months later. DFOS provided a warning at least three months in advance, and the contractor adjusted the curing plan for that area accordingly.
Case 2: An urban interchange in Beijing — hybrid FBG + distributed strain dual-system
We mentioned this case in our earlier fiber sensing series: the system deployed both FBG point strain sensors (at midspan and support locations) and a distributed BOTDR strain cable (running the full length of the main girder bottom slab). Comparing data from two static load tests revealed that the strain anomaly at the girder ends appeared earlier than at midspan — the distributed cable captured this gradient evolution completely, while the FBG sensors could only be retrofitted at that location after the fact. The data pointed to the following deterioration mechanism: leakage from the deck expansion joint caused rebar corrosion and expansion in the end crossbeam, triggering stress redistribution at the girder ends. This conclusion changed the maintenance priority — the plan had been to reinforce the midspan deterioration first, but the expansion joint waterproofing should have been addressed first.
Case 3: Grist Mill composite girder bridge, Maine — BOTDA + "sensing textile" rapid deployment
A team at the University of Massachusetts Lowell developed a scheme that embeds distributed optical fibers into a textile (a "sensing textile"), replacing the traditional method of directly taping fibers in place. On the Grist Mill Bridge, they ran loading tests with 4 heavy vehicles using BOTDA, and the sensing textile captured the complete strain distribution curve along the span of the main girder, matching theoretical calculations. The advantages of the sensing textile are: factory prefabrication, controlled quality, and on-site work limited to fixing the textile rather than point-by-point fusion splicing — for a bridge of the same scale, the traditional approach required 5 workers working 3 continuous days, while this approach was completed by 2 workers in half a day.
Practical Recommendations: Selection Is About Scenarios, Not Spec Sheets
Selecting a distributed fiber sensing system may look like choosing an interrogator (spatial resolution, range, accuracy), but it is really a judgment about the application scenario:
Main cables and hangers of long-span suspension/cable-stayed bridges involve long strain monitoring lengths and high accuracy requirements. We recommend BOTDA with distributed temperature compensation (simultaneously removing the temperature effect on the Brillouin frequency shift), a range starting from 10km, and 2-meter spatial resolution is sufficient.
Initial and secondary linings of long tunnels require high spatial resolution (to capture microcracks) and fast response during construction monitoring. We recommend OFDR — limited range but the highest accuracy. Note that OFDR cables must not have movable connectors, or the coherent signal will be disrupted and the data cannot be recovered.
Retrofit of existing structures (where double-ended routing is not possible): prefer a single-ended BOTDR solution, accepting slightly lower accuracy and longer sampling times as the trade-off. Before deployment, always measure the full-link attenuation with an OTDR first; any section with total channel loss exceeding 20dB needs an optical amplifier repeater.
Distributed Acoustic Sensing (DAS) is still at the research and small-scale pilot stage in the bridge and tunnel domain, with high interrogator prices (roughly 150,000–250,000 RMB per unit for domestic units, 500,000+ RMB for imported brands). But for scenarios that require "listening" to crack propagation and evaluating damage evolution, it is currently the most cost-effective option available.
Conclusion: The Essence of Monitoring Is Reducing Uncertainty
Whether it is FBG or DFOS, digital twins or AI diagnostics, the fundamental purpose of bridge and tunnel monitoring has always been the same: reducing uncertainty in O&M decision-making. Traditional methods rely on experience — high uncertainty; periodic inspection relies on sampled data — moderate uncertainty; only continuous full-field monitoring combined with digital twin calibration can push uncertainty down to a level acceptable for decision-making.
The role of distributed fiber sensing in this chain is clear: it turns "sampling" into "full field" and "periodic" into "continuous." But the technology itself does not automatically create value — sensor selection, cable installation workmanship, data calibration workflows, and continuous updates of the digital twin model must all be done solidly for the digital twin to mean anything. If you just lay a bunch of fiber, hook up an interrogator and look at curve plots, that is no different from installing GPS but never checking the navigation.
Technology is the tool; decision-making is the goal. Don't mistake the means for the end.