Anthony D. Puckett

Dynamics of a noncontacting, white light Fabry–Perot interferometric displacement sensor

A white light extrinsic Fabry-Perot interferometer is implemented as a noncontacting displacement sensor, providing robust, absolute displacement measurements with micrometer accuracy at a sampling rate of 10 Hz. This paper presents a dynamic model of the sensing cavity between the sensor probe and the nearby target surface using a Fabry-Perot etalon approach obtained from straightforward electromagnetic field formulations. Such a model is important for system characterization, as the dynamically changing cavity length imparts a Doppler shift on any signals circulating within the sensing cavity. Contrary to previously published results, Doppler-induced shifting within the low-finesse sensing cavity is shown to significantly distort the measurement signal as recorded by the sensor. Experimental and simulation results are compared, and the direct effects of cavity dynamics on the measurement signal are analyzed along with their indirect impact on sensor performance. This document has been approved by Los Alamos National Laboratory for unlimited public release (LA-UR 12-00301).

Experimental Validation and Uncertainty Quantification of a Single-Mode Optical Fiber Transmission Model

The radial intensity distribution of a transmission from a single-mode optical fiber is often approximated using Gaussian-shaped spatial field distributions. While such approximations are useful for some applications, they do not accurately describe optical transmission intensity off of the axis of propagation. A recent paper was presented that more accurately describes the intensity distribution, and this paper presents a simple experimental setup that verifies the models accuracy through formal uncertainty quantification procedures. Agreement between the model and the experiment is established both on and off of the axis of propagation. These results are then discussed in the context of displacement sensor designs based on the optical lever architecture. Transmission behavior off of the axis of propagation controls the sensor performance when large lateral offsets (25-1500 μ m) exist between transmitting and receiving fibers. The practical implications of modeling accuracy over this lateral offset region are discussed as they relate to the development of high-performance, intensity-modulated optical displacement sensors. Specifically, the sensitivity, linearity, resolution, and displacement range of a sensor are functions of the relative positioning of the sensors transmitting and receiving fibers. It is concluded that the predictive capability of the model presented in this paper could enable an improved methodology for high-performance sensor design.

A performance comparison of transducer designs for interferometric fiber optic accelerometers

Interferometric fiber optic accelerometers constitute a high-responsivity, high-resolution sensing architecture, with achievable sensitivities of several rad/g and resolutions in the micro-g range, depending on the specific configuration. Fiber Bragg grating (FBG) optical accelerometers offer ease of multiplexing but are inherently less sensitive than their interferometric counterparts. Fiber-based accelerometers have the usual optical advantages of being lightweight, electromagnetically immune, and non-spark emitting over traditional (piezo-electric) accelerometer architectures. Among fiber optic sensing methodologies, both interferometric and FBG accelerometers can be interrogated using phase-based demodulation, which offers advantages over intensity-based sensing schemes such as increased linearity, repeatability, and insensitivity to extraneous measurands. The performance of an accelerometer is often characterized in terms of its bandwidth, sensitivity, and resolution, all of which depend on the specific transducer design (the mechanical architecture) as well as the optical interrogation architecture. For a given optical interrogation architecture, a fundamental tradeoff exists in accelerometer transducer design between bandwidth and sensitivity; attempts to increase bandwidth will generally result in a decrease in sensitivity. This paper investigates the frequency and displacement characteristics that govern this tradeoff for several transducer configurations, in order to determine a pair of configurations that offer the greatest sensitivity for a given optical interrogation methodology (interferometric or FBG), at a prescribed bandwidth. The feasibility of several mechanical architectures is assessed based on the physical dimensions required for a given configuration to achieve a primary resonance of at least 15 kHz. The deflection of those configurations under their own self-weight is then considered a measure of accelerometer sensitivity in the measurement band below primary resonance. This paper has been reviewed by Los Alamos National Laboratory and received the following release number: LA-UR 10-00671.

Performance characterization of an intensity modulated fiber optic displacement sensor

A testbed simulating an intensity-modulated fiber optic displacement sensor is experimentally characterized, and the implications regarding sensor design are discussed. Of interest are the intensity distribution of the transmitted optical signal and the relationships between sensor architecture and performance. Particularly, an intensity-modulated sensors sensitivity, linearity, displacement range, and resolution are functions of the relative positioning of its transmitting and receiving fibers. In this paper, sensor architectures with various combinations of these performance metrics are discussed. A sensor capable of micrometer resolution is reported, and it is concluded that this work could lead to an improved methodology for sensor design. This paper has been approved by Los Alamos National Laboratory for unlimited public distribution (LA-UR 10-06637).

Understanding the effects of Doppler phenomena in white light Fabry–Perot interferometers for simultaneous position and velocity measurement

In static tests, low-power (<5 mW) white light extrinsic Fabry-Perot interferometric position sensors offer high-accuracy (μm) absolute measurements of a targets position over large (cm) axial-position ranges, and since position is demodulated directly from phase in the interferogram, these sensors are robust to fluctuations in measured power levels. However, target surface dynamics distort the interferogram via Doppler shifting, introducing a bias in the demodulation process. With typical commercial off-the-shelf hardware, a broadband source centered near 1550 nm, and an otherwise typical setup, the bias may be as large as 50-100 μm for target surface velocities as low as 0.1 mm/s. In this paper, the authors derive a model for this Doppler-induced position bias, relating its magnitude to three swept-filter tuning parameters. Target velocity (magnitude and direction) is calculated using this relationship in conjunction with a phase-diversity approach, and knowledge of the targets velocity is then used to compensate exactly for the position bias. The phase-diversity approach exploits side-by-side measurement signals, transmitted through separate swept filters with distinct tuning parameters, and permits simultaneous measurement of target velocity and target position, thereby mitigating the most fundamental performance limitation that exists on dynamic white light interferometric position sensors.

The use of time reversal methods with Lamb waves to identify structural damage in a pipeline system

Harsh environmental and operating conditions often leave pipeline systems prone to cracks, corrosion, and other aging defects. If left undetected, these forms of damage can lead to the failure of the pipeline system, which may have catastrophic consequences. Most current forms of health monitoring for pipeline systems involve nondestructive evaluation (NDE) techniques. These techniques often require a pipeline system to be taken out of operation at regularly scheduled intervals so that a technician can perform a prescribed NDE measurement. Such a measurement also requires direct access to the pipes exterior or interior surface. This access may require excavation if the pipe is underground and the removal of insulating layers when present. This research proposes the use of Macro-fiber composite (MFC) actuators for damage detection in pipeline structures. Because MFC actuators are durable and relatively inexpensive, they can be permanently bonded to the surface of a pipe during installation. Therefore, measurements for damage detection can be performed at any time, even while the system is still in operation. The time reversal methods use the propagation of Lamb waves to evaluate the structural health of a pipeline system. A burst waveform is used to excite Lamb waves in a pipe at an initial location using an array of MFC patches. The measured response at the actuation location is reversed in time and used as the excitation signal at the second location. The initial excitation signal is then compared to the final response signal. The performance of the time reversal methods was compared to the traditional methods of Lamb wave propagations using standard tone burst waveforms.

Simultaneous measurement of displacement and velocity using white light extrinsic Fabry-Perot interferometry

Low-power (<;5 mW) white light extrinsic Fabry-Perot interferometric sensors offer micron-level absolute position measurement capability for devices under test (DUTs) over large (cm) displacements. Because the position is determined directly by demodulating the interferogram, the sensors are robust to most common sources of power fluctuations but sensitive to DUT velocity, which induces Doppler shifting. This shifting results in interferogram bias errors as large as 50-100 μm for DUT velocities as low as 0.1 mm/s. This paper derives a model for predicting this Doppler-induced bias in the interferogram and exploits the derived relationship with a dual-filter optical architecture to obtain simultaneously the DUT velocity and bias-corrected absolute displacement. This approach is successfully demonstrated with a prototype in lab bench experiment. (Approved for release; LA-UR #12-23426).

Performance limitations of a white light extrinsic Fabry-Perot interferometric displacement sensor

Non-contacting interferometric fiber optic sensors offer a minimally invasive, high-accuracy means of measuring a structures kinematic response to loading. The performance of interferometric sensors is often dictated by the technique employed for demodulating the kinematic measurand of interest from phase in the observed optical signal. In this paper a white-light extrinsic Fabry-Perot interferometer is implemented, offering robust displacement sensing performance. Displacement data is extracted from an estimate of the power spectral density, calculated from the interferometers received optical power measured as a function of optical transmission frequency, and the sensors performance is dictated by the details surrounding the implementation of this power spectral density estimation. One advantage of this particular type of interferometric sensor is that many of its control parameters (e.g., frequency range, frequency sampling density, sampling rate, etc.) may be chosen to so that the sensor satisfies application-specific performance needs in metrics such as bandwidth, axial displacement range, displacement resolution, and accuracy. A suite of user-controlled input values is investigated for estimating the spectrum of power versus wavelength data, and the relationships between performance metrics and input parameters are described in an effort to characterize the sensors operational performance limitations. This work has been approved by Los Alamos National Laboratory for unlimited public release (LA-UR 12-01512).

Experimental verification of a model describing the intensity distribution from a single mode optical fiber

The intensity distribution of a transmission from a single mode optical fiber is often approximated using a Gaussian-shaped curve. While this approximation is useful for some applications such as fiber alignment, it does not accurately describe transmission behavior off the axis of propagation. In this paper, another model is presented, which describes the intensity distribution of the transmission from a single mode optical fiber. A simple experimental setup is used to verify the models accuracy, and agreement between model and experiment is established both on and off the axis of propagation. Displacement sensor designs based on the extrinsic optical lever architecture are presented. The behavior of the transmission off the axis of propagation dictates the performance of sensor architectures where large lateral offsets (25-1500 μm) exist between transmitting and receiving fibers. The practical implications of modeling accuracy over this lateral offset region are discussed as they relate to the development of high-performance intensity modulated optical displacement sensors. In particular, the sensitivity, linearity, resolution, and displacement range of a sensor are functions of the relative positioning of the sensors transmitting and receiving fibers. Sensor architectures with high combinations of sensitivity and displacement range are discussed. It is concluded that the utility of the accurate model is in its predicative capability and that this research could lead to an improved methodology for high-performance sensor design.

Performance Comparison of Fiber Optic Tips in Interferometric Displacement Measurements

Fiber optic displacement sensors have many potential advantages over traditional displacement measurement techniques, including small size, immunity to electromagnetic interference, electrical isolation, and high resolution. In this report, we focus on an interferometric fiber optic sensor, where the gap between the fiber tip and the device under test forms a Fabry-Perot resonant cavity. An optical interrogator measures the reflected intensity at wavelengths ranging from 1510 to 1590 nm. The spacing between resonant frequencies allows us to determine the distance from the tip to the device under test. We consider ferrule connector angled physical contact (FC/APC), ferrule connector ultra physical contact (FC/UPC) and unpolished cleaved tips and compare their influence on sensor performance. A plane wave propagation model is proposed for predicting tip effects. Comparisons are made on the basis of sensor measurement range, resolution, and sensitivity to changes in test conditions. In this paper, we discuss the experimental setup, detail our analysis, and present test results with recommendations for the applications of each tip.