Fiber Optic Sensor

What is a Fiber Optic Sensor

A fiber optic sensor is a measurement device in which an optical fiber, or a transducer optically coupled to a fiber, converts a physical or chemical quantity into a change in a light signal. The measurand modulates one of four optical properties: intensity (how much light returns), wavelength (the color that returns, as in a fiber Bragg grating), phase (the optical path length, exploited by interferometers), or polarization (the orientation of the light field). A detector and an instrument called an interrogator read that modulation and convert it back to engineering units such as microstrain, degrees Celsius, bar, or g.

What separates a fiber optic sensor from a conventional electrical sensor is the medium. The sensing element is silica or polymer glass, a dielectric, and no electrical current flows at the measurement point. This single fact produces the family's defining advantages: complete immunity to electromagnetic and radio-frequency interference, no spark source in explosive atmospheres, freedom from lightning and ground-loop problems, the ability to run the sensing point tens of kilometers from the electronics, and tolerance of corrosive and high-temperature environments that destroy electrical gauges. A single fiber thinner than a human hair can also carry many sensors at once.

The field grew out of telecommunications. Low-loss silica fiber, commercialized in the 1970s, gave researchers a transmission medium so sensitive to strain, temperature, and pressure that the same effects that degrade communication links became the basis for measurement. The fiber Bragg grating, first written photo-inductively by Hill and colleagues in 1978 and made practical by the UV side-writing method in 1989, turned the fiber core itself into a wavelength-encoded gauge. In parallel, optical time-domain reflectometry (OTDR), developed for locating faults in telecom cable, evolved into distributed temperature, strain, and acoustic sensing that reads thousands of points along an ordinary fiber.

The application scale is broad. At the high end, distributed acoustic sensing (DAS) instruments listen along buried pipelines and subsea cables over ranges that can exceed 50 km, and distributed temperature sensing (DTS) maps the temperature profile of a power cable or an oil well over tens of kilometers. At the point-sensor end, FBG arrays instrument aircraft wings, wind-turbine blades, bridges, dams, and the windings of power transformers. At the simplest end, an intensity-based plastic-fiber unit on a factory amplifier detects a small part at a conveyor in tens of microseconds. No single device covers this whole span: selection is the act of mapping a measurement need to the right optical principle, fiber, and packaging.

Four engineering metrics dominate the value of a fiber optic sensing system: the measurand and its sensitivity (for example, pm per microstrain for an FBG), the spatial coverage (a single point, a multiplexed array, or a fully distributed profile with a stated spatial resolution), the environmental envelope (temperature, strain, and chemical limits set by the coating and packaging), and the interrogator (its wavelength band, accuracy, acquisition rate, and channel count). These four together decide both the capability and the total installed cost, because the interrogator is usually the most expensive single component.

Intrinsic and Extrinsic Classification

The first and most fundamental classification of fiber optic sensors is whether the fiber is itself the sensor or merely the light pipe. In an intrinsic sensor the measurand acts on light while it is still inside the fiber, so the fiber core is the transducer. In an extrinsic sensor the fiber carries light to and from a separate external sensing region, and the modulation happens outside the fiber. This distinction decides whether you can build a distributed or multiplexed network and how the sensing region is engineered.

Intrinsic sensors are the workhorse of structural and infrastructure monitoring. Because the fiber core carries the measurement, a single fiber can host many fiber Bragg gratings at different wavelengths (quasi-distributed) or be read continuously along its full length (fully distributed). This is what lets one buried fiber become thousands of virtual sensors along a pipeline, a power cable, or a dam. The trade is that the sensitivity is set by the physics of silica, so cross-sensitivity between strain and temperature must be separated by design, for example by pairing a strained grating with a strain-isolated reference grating.

Extrinsic sensors decouple the sensing region from the fiber. An extrinsic Fabry-Perot interferometric (EFPI) sensor forms a tiny optical cavity at the fiber tip whose length changes with pressure or displacement, giving very high resolution in a small package and serving as an EMI-immune alternative to an electrical pressure sensor in harsh fields. A gallium arsenide (GaAs) crystal or a fluorescent phosphor at the fiber tip gives an absolute temperature reading that is independent of light intensity, which is the standard method for monitoring the hot spots inside power transformer windings and for measurement in strong microwave and MRI fields. Sapphire-fiber extrinsic probes extend point temperature sensing into the 1000 degrees Celsius and above range for furnaces and gas turbines.

A second axis cuts across this one: whether a sensor is a single point, quasi-distributed, or fully distributed. A single-point sensor reports one location. A quasi-distributed sensor multiplexes many discrete points (typically FBGs) on one fiber, each at a known position. A fully distributed sensor returns a continuous profile, where every meter of fiber is effectively a sensor, with the resolution set by the interrogation method. The right axis combination, intrinsic versus extrinsic and point versus distributed, is the first decision in any fiber sensing project, because it determines the entire instrument architecture and budget.

Mainstream Sensing Principles

Within the intrinsic and extrinsic classes, four sensing principles cover almost all commercial fiber optic sensors: intensity modulation, fiber Bragg grating (wavelength), interferometric (phase), and distributed scattering. Each has a characteristic measurand, accuracy, reach, and cost. The table below compares the four, after which each is explained.

Intensity-based sensing is the simplest and oldest method: the measurand changes how much light reaches the detector. In a factory photoelectric fiber unit, a plastic optical fiber carries light from an amplifier to a sensing head and back, and an object breaking the beam or reflecting it changes the received power, the same detection principle used by a self-contained photoelectric sensor. These units are cheap, rugged, and fast (a digital fiber amplifier such as the Keyence FS-N40 series reaches response times of about 23 microseconds), but because the reading is raw intensity it drifts with source aging, connector loss, and fiber bending, so it is used for detection (presence or absence) rather than precise metrology.

Fiber Bragg grating (FBG) sensing is the dominant precision point-sensing method. A periodic refractive-index modulation written into the core reflects one narrow Bragg wavelength, typically near 1550 nm, and passes the rest. Strain and temperature change both the grating period and the index, shifting that wavelength by about 1.2 pm per microstrain and roughly 10 to 13 pm per degree Celsius. Because the information is in wavelength, not amplitude, FBGs are immune to intensity loss, and many gratings at different wavelengths can share one fiber by wavelength-division multiplexing. Suppliers include HBM FiberSensing, FBGS, FiSens, and Proximion, read by interrogators such as the Luna HYPERION and HBM FS22.

Interferometric sensing encodes the measurand in optical phase and offers the highest sensitivity of any fiber technique. Mach-Zehnder, Michelson, Sagnac, and Fabry-Perot configurations measure minute changes in optical path length caused by strain, pressure, or acoustic waves. The Sagnac interferometer is the basis of the fiber optic gyroscope used in inertial navigation. An extrinsic Fabry-Perot cavity at a fiber tip gives sub-nanometer displacement resolution for pressure and microphone sensors. The cost of this sensitivity is complexity, careful path matching, and demanding signal processing.

Distributed scattering sensing reads the light that scatters back from every point along an ordinary fiber, turning the whole fiber into a continuous sensor. Three scattering mechanisms are used. Rayleigh scattering, read by phase-sensitive OTDR, is the basis of distributed acoustic sensing (DAS) for vibration and intrusion, with spatial resolution of about 1 to 10 m, covering along a single fiber what a point vibration sensor would otherwise measure at one location. Raman scattering, whose anti-Stokes to Stokes intensity ratio depends on temperature, is the basis of distributed temperature sensing (DTS), resolving about plus-or-minus 1 degree Celsius at roughly 1 m resolution, in effect placing thousands of temperature sensor points along one cable. Brillouin scattering, whose frequency shift depends on strain and temperature, drives distributed strain sensing (BOTDA and BOTDR). Suppliers include AP Sensing, Silixa, Bandweaver, and OptaSense.

Fiber Types, Materials, and Standards

The fiber and its coating are as decisive as the sensing principle, because they set the optical performance, the temperature ceiling, and the mechanical durability. The first choice is single-mode versus multimode glass, and beyond glass, large-core plastic optical fiber for industrial intensity sensing. The designations 9/125, 50/125, and 62.5/125 give the core and cladding diameters in micrometers.

Single-mode fiber (9/125 micrometer, OS2 per ITU-T G.652) carries one spatial mode near 1550 nm and preserves the optical phase and a clean reflected spectrum. It is mandatory for FBG, interferometric, and most Rayleigh and Brillouin distributed sensing, where a single mode is required to read wavelength or phase cleanly. Its small core demands precise connectors and splices but gives the lowest loss and the longest reach.

Multimode fiber (50/125 micrometer OM2 or OM3, or legacy 62.5/125 micrometer OM1) carries many modes and a larger core that collects more backscattered light. Raman DTS commonly uses multimode fiber because the Raman intensity ratio does not depend on phase, so modal effects are tolerable and the extra collected power improves the temperature measurement. The larger core also eases launch alignment in rugged field installations.

Plastic optical fiber (typically around a 1 mm core) is the medium of industrial intensity sensing. It is light, flexible, and bendable to very small radii (down to roughly 1 mm in fine fibers), making it ideal for routing a photoelectric beam into a tight machine space. It is cheaper and easier to terminate than glass, but its high attenuation limits it to short distances and its working temperature is modest, so it serves detection rather than precision measurement.

The coating sets the temperature ceiling and chemical durability of the sensing fiber. Standard acrylate recoats serve about -40 to +85 degrees Celsius, polyimide recoats reach roughly +200 degrees Celsius, and metal coatings or regenerated and sapphire fibers extend point sensing to +1000 degrees Celsius and beyond for furnace and turbine work. Relevant standards include the IEC 61757 series specific to fiber optic sensors, the ITU-T G.652 and G.651.1 transmission-fiber standards, ISO 18674 for geotechnical fiber monitoring, and the IEC 60079 hazardous-area series that governs use in explosive atmospheres. The table below maps environments to fiber and coating choices for initial selection.

Key Specification Parameters

Fiber optic sensing datasheets mix sensor specifications with interrogator specifications, and the two must be read together because the measurement performance is the combination of both. The parameters below are the ones that actually drive selection, organized by sensor side and instrument side.

Sensitivity is the change in the optical signal per unit measurand. For an FBG near 1550 nm it is about 1.2 pm of wavelength shift per microstrain and roughly 10 to 13 pm per degree Celsius. Sensitivity multiplied by interrogator wavelength resolution gives the smallest measurand change you can resolve, so a 1 pm interrogator paired with a 1.2 pm per microstrain grating resolves below 1 microstrain. Measurement range for distributed systems is the fiber length the instrument can read, from a few kilometers to tens of kilometers, and it trades against spatial resolution and acquisition time.

Spatial resolution applies to distributed systems and is the shortest length along the fiber that can be resolved as an independent point. Raman DTS commonly achieves about 1 m, with advanced systems reaching the 0.1 m class. DAS phase-OTDR typically resolves 1 to 10 m, set mainly by the optical pulse width, where a 100 ns pulse gives about 10 m. Finer spatial resolution generally costs measurement range and signal-to-noise, a fundamental distributed-sensing trade.

Wavelength accuracy and stability are interrogator metrics for wavelength-encoded systems. A reference instrument such as the Luna HYPERION si155 specifies about 1.2 pm wavelength accuracy and about 1.3 pm wavelength stability across roughly the 1500 to 1620 nm band, with a dynamic range near 25 dB. These figures, divided by sensor sensitivity, set the floor on strain or temperature resolution. Acquisition rate is how fast the instrument reads all sensors; the si155 covers about 10 Hz to 5 kHz, enough for vibration and dynamic strain, while a DTS system may take seconds per full temperature trace.

Channel count and multiplexing capacity decide how many sensors one instrument serves. Swept-laser FBG interrogators offer up to 4 parallel optical channels, each a wide wavelength band, and because each grating occupies a different wavelength slot, one channel can carry tens of gratings in series and a full instrument can read hundreds of FBGs. Distributed interrogators instead resolve thousands of virtual points along a single fiber, set by range divided by spatial resolution.

Environmental and mechanical limits come from the fiber, coating, and packaging. Key figures are the operating temperature range (acrylate to about +85 degrees Celsius, polyimide to about +200, sapphire and regenerated gratings to +1000 and beyond), the maximum working and ultimate strain (ultra-high-strength FBGs carry ultimate strain limits of about 5 to 6 percent, with practical working strain far lower for fatigue life), the minimum bend radius (large in single-mode, very small in plastic fiber), and the chemical compatibility of the coating. The output of the system is an electrical or digital signal from the interrogator, commonly Ethernet, so the field network stays optical while the control interface is conventional.

Selection Decision Factors

To turn the preceding chapters into a specific system, follow the decision sequence below. In fiber sensing the most expensive mistake is choosing the architecture (point versus distributed, intrinsic versus extrinsic) after committing to a fiber route, so resolve the high-level questions first. These steps can serve as a fixed RFQ template.

1. Measurand and architecture: First decide what you measure (strain, temperature, vibration, pressure, presence) and whether you need a single point, a multiplexed array, or a fully distributed profile. This selects FBG, interferometric, distributed scattering, or an intensity unit before anything else.
2. Reach and spatial resolution: Specify the fiber length and, for distributed systems, the required spatial resolution. Reach and resolution trade against each other and against acquisition time, so state the operating point, for example 1 m resolution over 10 km, rather than both extremes at once.
3. Sensitivity and resolution budget: Multiply sensor sensitivity (pm per microstrain or per degree Celsius) by interrogator wavelength resolution to confirm the smallest measurand change you can detect meets the application, with margin.
4. Fiber and coating: Choose single-mode for FBG, interferometric, Rayleigh, and Brillouin; multimode for Raman DTS; plastic fiber for industrial intensity units. Match the coating (acrylate, polyimide, metal, sapphire) to the temperature and chemical environment from Chapter 4.
5. Interrogator: Confirm wavelength band, accuracy, stability, acquisition rate, and channel count against the sensor network. The interrogator is usually the dominant cost, so size channel count to the final sensor population plus spare.
6. Cross-sensitivity and packaging: Separate strain from temperature by design, for example a strained grating plus a strain-isolated reference grating, and specify the mechanical package (weldable, bondable, embeddable, surface-mount) for the host structure.
7. Certifications and environment: For hazardous areas confirm the passive optical network meets the relevant IEC 60079, ATEX, or IECEx requirements, and for geotechnical and structural work reference ISO 18674 and IEC 61757. Confirm ingress protection and vibration ratings for the interrogator enclosure.
8. Total cost of ownership: Account for the interrogator, the sensing fiber and connectors, splicing and installation labor, calibration, and the value of the data over the asset life. A distributed system has a high instrument cost but replaces hundreds of discrete sensors and their wiring, often winning on a per-point basis over a long route.

One commonly overlooked dimension is serviceability and the supplier ecosystem. Fiber sensing is a young field with consolidating vendors, so check that interrogator software is maintained, that spare gratings and patch cords are available, and that field splicing and re-termination can be done locally. Luna Innovations (which absorbed Micron Optics, OptaSense, and LIOS), HBM FiberSensing, AP Sensing, Silixa, and Bandweaver maintain established product lines and support for FBG and distributed systems, while Keyence, Omron, SICK, Banner, and Baumer dominate industrial intensity-based fiber units with broad distribution and short lead times.