Multimode Optical Fiber Sensing
Although the vast majority of fiber optic strain sensors use single mode fiber, multimode fiber has a higher nonlinear threshold that enables higher light levels and lower noise, while the diversity of spatial modes can be used to develop sensors that are inherently immune to signal fading.
The purpose of this research was to develop detection, interrogation, and data processing techniques that leverage the unique features of multimode fibers to build next-generation fiber sensors with increased functionalities and performance.
Distributed fiber optic sensors have been developed at the Naval Research Lab and elsewhere for a variety of applications including structural health monitoring, perimeter security, seismic sensing, and underwater acoustic arrays. Many successful fiber sensors take advantage of Rayleigh scattering using a technique known as phase-sensitive optical time domain reflectometry (Φ-OTDR). These systems operate by detecting the back-scattered light in an optical fiber that results from random fluctuations in the index of refraction. Changes in the strain along the fiber are monitored by detecting changes in the backscattered light.
The vast majority of Rayleigh-based OTDR sensors use single mode fiber (SMF) and are limited by two significant disadvantages. First, Rayleigh scattering is a weak process, so single mode fiber sensors are often limited by low light levels. Higher input light levels could mitigate this issue, but nonlinear thresholds limit the optical power that can be injected into a fiber. Second, Rayleigh-based SMF sensors are susceptible to signal fading, which occurs when the backscattered light destructively interferes with itself, degrading sensor performance.
In order to make quantitative measurements of the strain along the fiber, most existing Rayleigh sensors measure the phase of the backscattered light. These phase-measuring systems require coherent detection of the backscattered light and localize detection by measuring the phase difference between two reflecting regions. The length of the reflecting regions depends on the pulse duration, so optimizing spatial resolution drives the need to use short pulses of light, which reduces backscattered light levels and increases noise. Laser phase noise can also be a significant limitation in this approach since coherent detection is required.
In contrast, systems that detect only the amplitude of the backscattered light have the potential to have reduced noise levels. These amplitude-measuring systems can use longer pulses of light since the reflecting region and localized sensing region are now one and the same, which also reduces the dependence on laser phase noise. However, in general, the amplitude has a nonlinear dependence on the strain, so traditional amplitude-measuring systems cannot make quantitative measurements.
The use of multimode fiber offers several advantages over single mode fiber in OTDR sensors. Multimode fiber has a higher power threshold for nonlinear effects and larger capture fraction of backscattered light, which can enable lower noise measurements. In addition, the rich diversity of spatial modes in the fiber provides additional useful information. First, by measuring the entire back-scattered speckle pattern, which is a summation of light from all the spatial modes, signal fading can be eliminated. Furthermore, the information contained within the amplitude of all the spatial modes is sufficient to extract a linear strain response. This insight enabled a new type of OTDR sensor that only uses the amplitude of the backscattered light to make quantitative strain measurements. Results show that multimode-fiber-based OTDR sensors can provide functionality and performance beyond what is possible with single mode fiber.
This work was done by Matthew J. Murray for the Naval Research Laboratory. For more information, download the Technical Support Package (free white paper) below. NRL-0080
This Brief includes a Technical Support Package (TSP).
Multimode Optical Fiber Sensing
(reference NRL-0080) is currently available for download from the TSP library.
Don't have an account? Sign up here.