Optical fiber sensors are used to monitor strain at elevated temperatures on modern high- temperature alloys during cyclic loading. Presented are the application and operation of metal coated silica-based fibers and extrinsic Fabry-Perot strain sensors monitoring fatigue tests at high-temperatures. The resultant strains from varying fatigue cycles and temperatures, from ambient to 2070 degrees F (1132 degrees Celsius), were monitored with surface-attached, short gage length, low finesse Fabry-Perot interferometric optical fiber sensor elements. The results demonstrate that the fiber optic strain sensors are able to withstand extreme temperatures, while maintaining a high level of performance. The capabilities of the fiber optic strain sensors make it possible to monitor material property changes during high- temperature fatigue loading.

Aircraft and rotorcraft health monitoring systems (HMS) must be capable of identifying structural damage (cracks) when the damage is very small. Typically, small is small enough that classical finite element modeling methods are inadequate. Monitoring the structure at very high frequencies and using a combination of relativistic, and absolute measures of the monitored signals can be used to provide a measure of the structures health. Passively monitoring the structure for waveforms, generated by crack growth is a promising method for incorporation into a health monitoring system. This paper discusses research into the implementation of such a system. The issue of signal de-noising using adaptive filters prior to waveform identification are addressed, and results presented.

The measurement and prediction of the propagation of stress waves in impacted bars is the focus of this paper. Ultrasonic stress pulse propagation in bars is commonly monitored by means of stainless steel gauges. Because the outputs of these conventional gauges are very small and the frequency content of the stress pulses extend into the ultrasonic range, it is always necessary to employ amplifiers of a very high amplification and a very wide frequency bandwidth. On the other hand, PZT tiles of the same physical dimensions as the conventional gauge give voltage outputs which are factors of more than 10,000 greater than the conventional gauges. Consequently, the PZT tiles do not require any amplification system but produce signals which can be directly samples. In this paper PZT tiles of dimensions 5 by 3 mm, which are cut from standard PZT patches of dimensions 30 by 30 mm, are bonded to the surface of cylindrical bars and are used for monitoring the propagation of stress pulses induced in the bars by the collinear impact of spherical balls. The velocity of the spherical balls at impact is monitored by means of a pair of infra-red sensors. This velocity is used in conjunction with the Hertzian law of contact and the associated non-linear ordinary differential equation governing the impact of spherical balls against a rod to determine the force-time history of the impact. This force spectrum is used in a finite element model of the rod to predict the propagation of stress pulses in the rod. It is shown that the correlation between the measured and predicted stress pulses at different locations on the bars is very good.

Sensors based on quartz material can be used for measurement of mechanical strain. There exist two different structures due to propagation mode of the mechanical waves: bulk acoustic wave (BAW) and surface acoustic wave devices (SAW). Another way to classify the sensors is to divide them into resonators or delay-lines (latter only SAW). The sensors are passive elements so that no additional energy supply is needed. The interrogation signal is both information and energy-carrier. Furthermore the measurement system works in the MHz range. With these properties it is possible to operate the system using wireless transmission. In this way a sensor is obtained that is well suited for applications which are not easy or impossible to realize using wired systems: e.g. torque on rotating shafts can be measured at high rates (using DSP up to 10,000 measurements/second). The principle of signal processing works as follows: several periods of a rf sine-wave are received by the sensor antenna every 10 - 50 microseconds for a duration of some 100 ns. They are coupled into the sensor via the piezoelectric effect. Depending on the sensor type a damped oscillation (resonator) or a number of reflected sine-waves (delay line) can be received. In case of the resonator the resonant frequency changes with mechanical strain applied to the sensor and can be measured via radio echo transmission. The principle of signal processing for the SAW delay line is the same as that used for coherent pulse radar: a stressed sensor causes a phase shift of the received signal which is proportional to the elongation. With a maximum elongation of about 1000 ppm of the device used and a resolution of about 5 ppm provided by the signal processing, an accuracy of 0.5% is achieved. Fluctuations of temperature can be eliminated by summing the received signals of two sensors, one of which is stressed and the other compressed. By subtracting the two received and processed signals from each other the temperature can also be measured.

This paper presents the results of experiments to measure the internal strains and temperatures that are generated in carbon fiber/epoxy composite specimens during processing using embedded fiber optic strain sensors and thermocouples. Measurements of strain and temperature, combined with a computational model, offer the potential for non-destructive, real-time determination of residual stress in composites, and may be useful for process monitoring and control. Extrinsic Fabry-Perot interferometers, Bragg grating strain sensors, and thermocouples were embedded in graphite/epoxy composite laminates prior to cure. The specimens were cured in a press, and the internal strains and temperatures developed during processing were monitored and recorded. The results are compared with expected values, and limitations of the experimental technique are discussed.

A simple and very accurate method for measuring the viscosity of liquids is presented. A vibrating lead zirconate titanate (PZT) ceramic polycrystalline resonator is totally immersed in a liquid and the oscillation behavior is studied both analytically and experimentally. The vibrating ceramic generates a shear wave traveling in the surrounding fluid normal to the surface of the resonator with heavy damping. The PZT is modeled as a vibrating thin plate. The resulting one-dimensional governing equation of motion, which includes the effect of damping, is solved with the appropriate damped boundary conditions using impedance methods. The resonance frequency of a PZT immersed in a fluid is shown to be a function of the PZT parameters and the product of the fluid density and viscosity. Fluid loading is shown to lower the resonance frequency of the ceramic and both diminish and broaden the impedance plot near resonance. Unlike previous methods utilizing crystal shear-mode resonators, this method is valid for both low and high viscosity values. The proposed method provides viscosity values within 10% accuracy compared to tabulated values. The current study also provides an equivalent circuit model for the PZT in fluid loading. Immersing the PZT in a liquid increases both the inductance and the resistance of the unperturbed (dry) ceramic. The calculated elements of the equivalent circuit compare well with the values obtained using a best-fit statistical model.

Two types of optical fiber sensors (OFS) were investigated for use in monitoring the cure of an epoxy-amine resin system: (1) an evanescent wave sensor and (2) a refractive index sensor. The evanescent wave sensor was used to detect changes in concentration of the active chemical species involved in the cure reaction via evanescent wave near-infrared spectroscopy. By using the optical fiber as an attenuated total reflection waveguide, spectra were collected over the range 1490 - 1570 nm at regular time intervals during the cure. This technique enabled the depletion of amine to be monitored. Results obtained via this method were fitted to kinetic models which allowed prediction of the reaction rate at different cure temperatures and conversions. The optical fiber evanescent wave sensor results were compared with data obtained using an established cure monitoring technique (FT-IR spectroscopy). A theoretical model of the evanescent sensor has been used which describes the relationship between evanescent absorption as a function of absorber concentration and refractive index. Predictions of sensor response were undertaken using absorption data from FT-IR spectroscopy and refractive index results as a function of cure time. The predicted sensor response was then compared with experimentally obtained sensor data. An optical fiber sensor which monitored the cure process via refractive index change was also investigated. Sensors were set up to allow simultaneous collection of data during cure from the OFS, together with data from transmission near-infrared spectroscopy and Abbe refractometry. In this way the response of the sensor to changes in the cure state of the resin, refractive index and temperature was compared.

A compact intensity-based fiber-optic vibration sensor, suitable for embedment or surface- mounting has been designed and evaluated. It employs simple and inexpensive instrumentation, and is shown to be responsive to frequencies in the range of 100 - 10,000 Hz. It was mounted onto the surface of specimens of carbon fiber reinforced composite, and proved to be capable of differentiating between the vibrational responses of damaged and undamaged panels. An attempt was made to characterize the sensor's output signal by FFT processing. The sensor also survived embedment in a cementitious composite panel. It is proposed that the device can be used as the sensing element of a real-time mechanical-health- monitoring system.

For many applications it would be highly desirable to be able to measure all three axes of strain and temperature internal to composite materials. Conventional electrical strain gauges are undesirable to embed into composite materials because of their size, conductive nature, susceptibility to electromagnetic interference, incompatibility with the host material and temperature limitations. All of the tests done to date with single element fiber sensors have been limited to the measurement of strain in the in plane dimension. This paper describes an innovative fiber sensor based on dual overlaid fiber gratings on short lengths of birefringent polarization preserving fiber that allows three axes of strain and temperature to be measured at a single point.

The U.S. Navy is developing advanced fiber optic sensors for machinery monitoring and diagnosis. Robust, inexpensive fiber sensors which can be embedded into metal structures are desired in order to make strain and temperature measurements close to the expected machinery fault. The use of several types of fiber sensors was considered for embedding applications. This paper compares the properties and performance of optical fiber-based long-period and Bragg grating temperature, strain and refractive index sensors. The comparison is based on magnitude of spectral shift of the resonance bands, cross-sensitivity to undesired perturbations, bend sensitivity, and the ease and cost of demodulating the grating signal. The sensitivities of the long-period grating sensors were found to be strong functions of the fiber parameters. Although the long-period grating sensors displayed undesired cross-sensitivities and dependence on fiber bends, they exhibited larger spectral shifts than the short-period fiber Bragg grating sensors and can be employed with relatively inexpensive sensor demodulation methods. These features and the fact that the long-period grating sensors are more economical to manufacture make them a strong candidate for naval machinery monitoring applications.

The fiber optic extrinsic Fabry-Perot sensor was embedded in composite beam to sense the strain and failure of composite structures. A tensile test was performed to confirm the strain sensitivity of the fiber optic sensor embedded in composite specimens. The strain sensitivity of the extrinsic Fabry-Perot sensor showed very good agreement with the theoretical value. The bending deformation and matrix cracking were investigated through four-point bending tests of cross-ply composite beams with embedded fiber optic extrinsic Fabry-Perot sensor. The failure due to matrix cracks in the composite beam was confirmed by an edge replica method. The strain and failure signals were separated by digital filtering from the signal of fiber optic sensor. The failure instants were obviously noticeable from the failure signal obtained from the fiber optic signal by high pass filtering. The dominant failure strain of the composite beam was determined by strain signal obtained by low pass filtering.

The measurement of quasi-static strain field using optical fibers presents a considerable challenge due to the inherent sensitivity of optical fibers to temperature. This paper summarizes recent work we have carried out on two approaches to this problem. Dual mode polarimetric measurements were investigated as a means of implementing distributed temperature measurements and radio frequency subcarrier sensors have been used to perform the same measurement on an integrated basis. These techniques are contrasted and assessed against other technologies such as Bragg gratings and dispersive Fourier transform spectroscopy on the basis of measurement capability, ease of implementation and technological maturity.

Life extension programs for military metallic aircraft are becoming increasingly important as defense budgets shrink and world economies realign themselves to an uncertain future. For existing military weapon systems, metallic corrosion damage costs an estimated $8 billion per year. One approach to reducing this cost is to develop a reliable method to detect and monitor corrosion in hidden metallic structure with the use of corrosion sensors which would give an early indication of corrosion without significant disassembly. This paper describes the current status of the development, analysis, and testing of a fiber optic corrosion sensor developed jointly by Boeing and Virginia Tech Fiber & Electro-Optics Research Center and sponsored by USAF Wright Laboratory, Materials Directorate, contract #F33615-93-C-5368. In the sensor which is being developed under this contract, the normal cladding is removed in the sensor region, and replaced with aluminum alloy and allowed to corrode on coupons representative of C/KC-135 body structure in an ASTM B117 salt spray chamber. In this approach, the optical signal out of the sensor is designed to increase as corrosion takes place. These test results to determine the correlation between sensor output and structural degradation due to corrosion are discussed.

Optical fiber grating-based sensors are proposed and demonstrated for the detection of corrosion. Two techniques are employed to indirectly monitor corrosion: (1) measuring the corrosion-induced decrease in the residual strain of a metal-coated optical fiber short period grating sensor and (2) monitoring corrosion-induced changes in the dimension of a metal- coated, long-period grating sensor.

The smart aircraft fastener evaluation (SAFE) system is an advanced structural health monitoring effort to detect and characterize corrosion in hidden and inaccessible locations of aircraft structures. Hidden corrosion is the number one logistics problem for the U.S. Air Force, with an estimated maintenance cost of $700M per year in 1990 dollars. The SAFE system incorporates a solid-state electrochemical microsensor and smart sensor electronics in the body of a Hi-Lok aircraft fastener to process and autonomously report corrosion status to aircraft maintenance personnel. The long-term payoff for using SAFE technology will be in predictive maintenance for aging aircraft and rotorcraft systems, fugitive emissions applications such as control valves, chemical pipeline vessels, and industrial boilers. Predictive maintenance capability, service, and repair will replace the current practice of scheduled maintenance to substantially reduce operational costs. A summary of the SAFE concept, laboratory test results, and future field test plans is presented.

A sensor at a fixed location in a complex structure records a complicated but unique wave pattern containing information about impact location, imparted energy and any damage created by an impact event. An intelligently designed hybrid neural network system is capable of extracting this information from the sensory signal. Such a system based on a generalized regression neural network (GRNN) is described for the purpose of impact location, energy, and damage detection. A Northrop Grumman test article is utilized to demonstrate capabilities of the system. The system performance evaluation based on the preliminary experiments is very encouraging. Further experimental evaluations of the system are planned and are described in this manuscript.

A fiber optic optical time domain reflectometer-based (OTDR) sensing system is described for detecting the extent and/or location of damage incurred by a composite material following an impact event. Two multi-mode fibers are embedded within a graphite composite to form a 'checkerboard' fiber mesh with 0.5 inch spacing. The measurement system consists of a PC- based virtual instrument, a high resolution optical time domain reflectometer, and two fiber optic sensors. Using the OTDR and processing algorithms, the location of impact can be determined within the composite panel by scanning the length of each fiber from both ends. The OTDR scanning process yields four reference points which can be used to determine the location of the impact referenced to a coordinate axes designated within the composite panel.

The feasibility of embedding piezoelectric transducers in composite panels for the measurement of phase velocities is presented. The theoretical dispersion curves for a composite panel are generated and the experimental measurements of phase velocities compared. In this paper the method for testing and for evaluating the experimental phase velocities is developed. The experimental methodology is validated by excellent agreement to theoretical values. This measurement methodology is used to detect holes drilled in composite plates to show the usefulness of the procedure. Signal losses of between 10% and 48% can be detected with this embedded transducer method.

Intrinsic Fabry-Perot optical fiber sensors were embedded to the tensile side of the 20 cm by 20 cm by 150 cm cement concrete structures. The sensors were attached to the reinforcing steels and then, the cement concretes were applied. It took 30 days for curing the specimens. After that, the specimens were tested with 4-point bending method by a universal testing machine. Strains were measured and recorded by the strain gauges embedded near optical fiber sensors. Output data of fiber sensor showed good linearity to the strain data from the strain gauges up to 2000 microstrain. The optical fiber sensors showed good response after yielding of the structure while embedded metal film strain gauges did not show any response. We also investigated the behavior of the optical fiber sensor when the specimens were broken down. In conclusion, the optical fiber sensors can be used as elements of health monitoring systems for cement concrete infra-structures.

A non-contact optical method has been developed for the remote and high-speed interrogation of optical fiber sensors embedded in a structure. The technique allows the use of passive fiber sensor requiring no local on-board electrical power or local electronics in the structure, thus simplifying the design and manufacturing of the structure, and allowing potential applications in low-cost structures where the addition of on-board power and electronics may be cost- prohibitive. For demonstration, multiple absolute extrinsic Fabry-Perot interferometric (AEFPI) strain sensor elements were embedded in a polymer matrix cross-ply laminate coupon. Coupling of broadband optical power into the embedded sensor elements over a distance of several tens of centimeters was achieved using a compact and lightweight broadband light source. Optical radiation received back from the sensors within the test specimen was optically detected and electronically processed to obtain the AEFPI strain sensor output signals using a computer software-based signal processing unit designed for this application. The ability of the opto-electronic receiver unit to both detect changes in strain dynamically was determined by quasi-satirically increasing the load on the specimen using a small loading fixture. The further ability of the system to monitor strain dynamically during rapid motion was demonstrated by moving the specimen with respect to the input and output optics. The limitations of the system due to the operation of the detection system are detailed.

This paper describes a method of using matched-pattern gage sensors that are embedded into highway pavements to classify vehicles, i.e. cars vs. trucks. The classification of vehicle type is an important technology for a variety of highway operations, e.g. traffic control, maintenance planning, weigh-in-motion, and the assignment of tolls. Vehicle classification schemes that are based on strip-crossing methods are not very robust due to the large variability of strip-crossing sequences. Visual methods still rely primarily on human identification. The method described here involves placing long gage length sensors in highway pavements. The spatial pattern of the sensor is configured so that it will match the wheel pattern of the type of vehicle that is being identified. Theoretical modeling shows that the signal received from the sensor is a cross-correlation function relating the wheel and sensor patterns in space and time. The sensor can be any one of a variety that transduce by integrating pressure along a length. The technique is demonstrated in the laboratory with PVDF and fiber optic sensors. Experimental results and computer simulations are presented as well as a discussion of the realistic possibility of using such a vehicle identification scheme under field conditions.

Civil smart structures often require displacement sensors with measurement bases between a few centimeters and a few meters with a precision of the order of 1/100 mm. Low-coherence interferometry offers these performances even for long-term measurements. Being a non- incremental setup, it does not require an uninterrupted monitoring. The main drawback of this technique resides in the fact that a separate sensor is required for each section to be measured. A typical civil structure such as a bridge requires up to 50 sensors for each span, so the complexity for this type of instrumentation and the number of connections often limit its large- scale application. It would be interesting to subdivide the fiber sensor in domains that can be measured separately but have a single lead-out connection. This contribution presents an in- line multiplexing scheme for displacement sensors based on low-coherence interferometry and partial reflectors installed in pairs along the sensing fibers. The multiplexing of up to ten displacement sensors along the same fiber line is demonstrated theoretically and experimentally. Different types of partial reflectors are also compared. The special case of structures that are constructed in sections is also analyzed.

We measure both the temperature and strain responsivities for a fiber Bragg grating and Brillouin scattering system. We then show how it is possible to use both Brillouin scattering based and FBG sensing techniques to discriminate strain and temperature effects and determine these two parameters along an optical fiber.

Smart sensor is a recent concept presenting numerous advantages such as versatility, strong electromagnetic immunity, reduction of the connectivity, high computation power, etc. In civil engineering smart sensor based systems are well suited due to the large amount of spatially distant transducers and the need of large computational power. However, such systems require long development time, especially in their software part, and beside the multitude of instrumentation problems encountered, the need of a generic model is strong. The aim of the model is the design of a software generator for distributed data acquisition system. The key of our system is in the description of an instrumentation plane under the form of a data dependence graph (DDG). The goal of the generator is to map and 'execute' that DDG on the physical architecture according to the number of transducers, their affectation to the smart sensors and a PC based system controller. In this paper, after an outline of the smart sensor concept, we describe the DDG based representation of the instrumentation plan. An example of bridge monitoring is then described. Finally, the smart sensor, the system controller and the network modelization are outlined and their ability to allow the DDG mapping with the help of local or remote variable is shown.

The expansion of smart structures and the development of intelligent materials science as a crucial research endeavor has challenged our conventional approach to new material development. The most concerned challenges are perhaps the processing and eventual use of material systems that are inherently dynamic. Therefore, new instrumentations and smart sensing and processing should be introduced. In this paper we introduce new instrumentation to characterize and develop new polymeric structures for smart sensing. Our system is now capable of on-line sensing in real operational environments. To give intelligence to the system, an artificial intelligence method has been adopted for computerized chemical detection. Using conductive polymers, this system is able to recognize different solutions based on the reading of their pH vs R (resistance).

A method of determining the contact force on laminated composite plates subjected to low velocity impact is developed using the finite element method and a neural network. The back propagation neural network is used to estimate the contact force on the composite plates using the strain signals. The neural network is trained using the contact force and strain histories obtained from finite element simulation results. The finite element model is based on a higher order shear deformation theory and accounts for von-Karman nonlinear strain-displacement relations. The nonlinear time dependent equations are solved using a direct iteration scheme in conjunction with the Newmark time integration scheme. The training process consists of training the network with strain signals at three different locations. The effectiveness of different neural network configurations for estimating contact force is investigated. The neural network approach to the estimation of contact force proved to be a promising alternative to more traditional techniques, particularly for an on-line health monitoring system.

We describe a fiber optic velocity sensor which performs non contact measurements of structural vibrations. The sensor responds directly to the velocity of the sample structure rather than the associated displacement or acceleration through measurement of the Doppler- induced frequency shift of light reflected from the vibrating structure. Light from a single- frequency laser is launched into a fiber optic system which serves the dual role of delivering light to the structure and recollecting the frequency shifted reflected light. The frequency shift, which is proportional to the sample velocity, is processed with a fiber optic Mach-Zehnder interferometer with a relative optical path imbalance. The maximum detectable velocity of the sensor is greater than 1 km/s and is scalable over several orders of magnitude with changes in the fiber optic interferometer. Resolution of the velocity sensor is 80 micrometer/second/(root)Hz.

A novel fiber optic differential pressure transducer and multiplexing system is described for real-time pressure measurements of airflow on an actuator- and SMA-controlled airfoil. The design of the pressure transducer is based upon extrinsic Fabry-Perot interferometric technology and incorporates a micromachined silicon diaphragm as the pressure sensitive element. The pressure transducer has a full scale operating range of -10 to 10 psi_g and a resolution of greater than 0.01 psi. Data is presented demonstrating the reproducible performance of the fiber optic sensor after repeated cycling and also at various temperatures. Finally, various multiplexing techniques and results are described.

This paper describes a non-destructive inspection technique based on the interaction of ultrasonic S_O Lamb waves with holes, regions of impact damage, and delaminations, in carbon fiber composite plates. The Lamb waves were detected using a surface bonded, single mode optical fiber sensor operating at a wavelength of 633 nm forming one arm of an optical fiber Mach Zehnder interferometer. Lamb generation was accomplished by use of perspex coupled piezo-electric transducers.

A new class of distributed sensors is presented which can measure both the linear, angular deflections and twist of flat composite plates. The sensor relies in its operation on a set of wires which are embedded off the neutral planes of the composite plates. The wires are arranged in a special manner which allows continuous monitoring of the deflection curve of the plate. The output signals of the wires are processed to determine the linear, angular displacements and twist at critical discrete points in the plate. The equations governing the operation of the sensor are developed using the theory of finite elements. The resulting equations provide the sensor with unique interpolation capabilities which make it possible to map the deflection and strain fields over the entire surface of the plate. The theoretical and experimental performance of the sensor are presented in both the time and frequency domains. Comparisons are given between the experimental performance of the distributed sensor and that of conventional laser sensors in order to demonstrate the accuracy and merits of the distributed sensors. The results obtained suggest the potential of this new class of sensors as a viable means for monitoring the static and dynamic deflections of flexible composite SMART plate.

This paper discusses the development of a novel composite system where some of the reinforcing fibers act as the light guide. High purity silica reinforcing fibers with a diameter of 9 micrometer were used along with an appropriate cladding material to produce a light guide, which was termed a 'self-sensing' fiber. Self-sensing fibers were embedded within a 16-ply carbon fiber reinforced composite and the resultant panels were impact tested to examine the possibility of using the self-sensing fibers as an impact damage sensor (crack detector). Similarly, three types of conventional optical fibers, with outer diameters of 30 micrometer, 50 micrometer, and 125 micrometer were also embedded within composite panels. These were also impact tested to ascertain their effectiveness as crack detectors. Results indicate that the self-sensing fibers are capable of detecting impact damage as low as 2 J and proved to be more sensitive to impact damage than the other types of fiber investigated in this study.

Optical fiber sensors for monitoring of structures (OSMOS) is a European collaborative research project which has, over the past three years, embraced a number of technological issues related to the problem of structural monitoring in the civil engineering and aerospace industries. A key technical objective of the program was the measurement of temperature and strain using a single sensor length. A laboratory prototype using the differential sensitivities of polarimeters based on the fundamental, LP₀₁ mode and the first higher order LP₁₁ mode of polarization maintaining fiber demonstrated parameter recovery to within 2 C and 5 (mu) (epsilon) . A receiver enabling quasi-distributed measurements to be made with a linear spatial resolution of 70 cm using white light polarimetry was assembled. White light polarimetry was also used in conjunction with pressure sensitive fiber to detect impact damage on a composite radome structure. Impacts of 5 Joules in magnitude were detected with a spatial resolution of around 1 cm. Microwave radio frequency subcarrier measurement techniques were used to develop the engineering processes necessary to integrate optical sensors into civil engineering structures for simulated applications trials. This enabled issues such as stress transfer, mechanical bonding and sensor protection to be addressed. For the aerospace industry, embedding of optical fiber sensors remains an important issue. Here we developed techniques for embedding connectorized fibers such that the component could be machine finished after curing, an important feature of the manufacturing process.

Regular assessment is generally necessary for the maintenance of civil engineering structures. This requires periodic or permanent inspection needing appropriate heavy and expensive equipment. The optical fiber has certain potentials for a more efficient automatic or semi automatic maintenance solution. In this paper, we study a sensor consisting of a low birefringent (LB) single-mode optical fiber, inserted between two thin steel ribbons. From a mechanical view point, the ribbon transfers its applied external loading to the fiber in the unidirectional form. This transmission causes birefringence in the optical fiber by modifying the opto-geometrical parameters related to the difference between principal stresses in the fiber core. The polarization phase shift enables then to evaluate the applied load. This system was first theoretically studied. The mechanical model assimilates the optical fiber to an elastic spring of variable stiffness. Laboratory tests were carried out to validate the proposed model. The sensors response, represented by graphs, can thus be determined. Mechanical loading can hence be measured in real time with, eventually, a several sensors multiplexing possibility. The prospects of the proposed system are worth exploring in the assessment of civil engineering materials and structures especially in 'smart ones.'

An embedded, intensity-based fiber optic sensor was previously designed and evaluated for strain monitoring in advanced fiber reinforced composites under dynamic loading conditions. The original sensor design involved the use of two multimode fibers, each with a cleaved end. These fibers were fitted into a glass capillary and were secured in position via a fusion splice at each end of the capillary. However, the effective operational strain range of this sensor design was limited primarily to tensile loading. In order to use this sensor under compressive loading regimes, it was necessary to develop a technique to construct the sensor with a known separation of the fiber end-faces. In effect, the sensor is an extrinsic Fabry-Perot interferometric sensor. The signal processing was based on a scanning monochromator. The feasibility of using the optical fiber sensor for tensile and compressive strain measurements was demonstrated. The sensor was also used to obtain in-situ stiffness reduction data during the fatigue testing of a cross-ply carbon fiber reinforced composite. An analysis of the relationship between detection sensitivity and sensor geometry is also presented.

This paper shows the way to turn a defect inherent to single-mode fiber, namely birefringence, into a prime quality for a powerful and reliable sensor. The latter is entirely devoted to weigh- in-motion (WIM) applications extended to complete active traffic management systems. After a brief description of the sensor and its principle of operation, the theoretical model is developed. Then, a full characterization made in both static and dynamic conditions is presented. The results obtained illustrate how it is difficult to interpret a weight measured in dynamic conditions and correlate the value with the static weight.

Navigation in polar waters presents a formidable challenge to ships' propulsion systems as large ice pieces impinging on their propeller blades sometimes result in stresses exceeding the yield strength of the blade material. Damage to propellers is costly and can also spell disaster if a ship becomes disabled in a remote area. To prevent such situations, design practice must be improved and theoretical models of propeller/ice interaction must be validated against experimental data. The blade shape requires that the load be monitored at many locations in order to obtain an accurate picture of the stress and load distribution. Fiber optic sensors are ideally suited for such an application, owing to their small size, stability over time, immunity to electro-magnetic interference, resistance to corrosion and chemical attack by sea water and hydraulic oil. We report the full-scale instrumentation of an icebreaker propeller blade with 54 Fabry-Perot based fiber optic strain gauges and shaft-mounted electronics. The instrumentation design and installation procedures are described. Additional data gathered from the propulsion control system and the ship's navigation equipment is presented and the data fusion performed with underwater video imagery of the instrumented blade is also discussed. An overview of the noise-free data obtained during the Antarctic trials is given. We finally discuss the sensors behavior and long term response, presenting their applicability to smart structures.

The problem of synthesizing large amounts of sensor data into a meaningful form represents one of the key challenges in making effective use of smart sensors/actuators that are distributed throughout a structure. This paper develops an engineering approach for addressing this problem, focusing on how large sets of neuromusculoskeletal measurements can be synthesized with approximate reasoning by experts and trained human observation to help extract and prioritize the most salient diagnostic findings, given a reasonably large set of strategic performance tasks. A key objective is to create an environment for intimate human- computer interaction, that optimally uses the capabilities of each. The best of two key conceptual frameworks are synthesized: initial design via rule-based fuzzy expert critic modules, followed by a gradual transition toward fuzzy neuro-optimization and neuro- classification modules. It is suggested that this provides a more reasonable approach not only for interactive near-real-time medical diagnosis assisted by a 'smart' computer, but also for developing the types of robust adaptive critics needed for advanced studies of principles underlying neuromotor control and skill acquisition.

Despite many well-intentioned attempts to utilize state-of-the-art advanced control systems technology to design contact devices such as powered orthoses, there have been more failures than successes. In part this is due to our limited understanding of neuromechanical function, and of how to optimally design human-technology interfaces. This paper develops a theoretical foundation for mechanical impedance and postural stability for large-scale human systems, and for the analysis and design of human-technology contact interfaces. We start with four basic presuppositions: redundancy is a fundamental feature of biosystem design, muscle actuators possess intrinsic nonlinear stiffness which can be modulated, mechanical interaction between the human and an environment is fundamentally bicausal, and objects with certain properties can become almost a natural extension of the human body. We then develop the key concepts of intimate contact and extended proprioception, and provide examples of how these principles can be applied to practical problems in orthotics, focusing on posture-assist technologies. Finally, suggestions are put forward for applying smart materials and structures to innovative orthotic design.

Using functional electrical stimulation (FES), muscles of spinal-cord injured patients can be activated by externally generated electrical currents in order to restore function. As for gait, the question arises when during the gait cycle and two what extent individual muscles should be stimulated. Computer simulation provides the designer with a tool to evaluate the performance of different muscle stimulation patterns without the need to test patients at every stage of system development. The goals of this paper are: first, to identify, using computer simulation, multi-channel stimulation patterns that are capable of reproducing normal gait kinematics for a full gait cycle, without relying on sensory feedback (open-loop control); second, to briefly assess the stability of the gait obtained. A two-dimensional musculo-skeletal model was developed, based on mathematical representations of muscle properties (including force-length and force velocity characteristics and muscle activation dynamics). A visco-elastic model, including non-linear heel-pad properties, was used to describe the foot-ground interaction. A seven segment skeletal model was actuated by 8 major muscle groups in each leg. Rectangular muscle stimulation patterns were defined by 3 parameters: onset, termination and level of stimulation. Thus, the minimization of the differences between simulated and measured normal gait kinematics was a 24 (3 by 8) parameter optimization problem. Although a good agrement was found between simulated and measured kinematics (rms difference equals 6.5 degrees), stable cyclic locomotion was not achieved. At this point it is concluded that muscle properties do not provide sufficient stability to permit cyclic locomotion with sixteen channels of muscle stimulation, and that incorporation of sensory feedback control will be necessary to achieve this goal.

The advent of the emerging field of smart sensors suggests new applications for implantable microelectronic devices in neural prostheses. Optimal use of miniature and subminiature (thin- film) electronic sensors in implanted systems will depend upon the nature of the power and communication link to the sensor. Microtelemetry technology is under current development to meet this need. Microtelemetry techniques can be used to provide operating power and bi- directional communication for a microimplant through a common, wireless, magnetic link. Owing to the extremely unfavorable geometry, i.e. the size of the implant relative to the size of the extracorporeal transmitter, the design of such links is highly parametric. Magnetic circuit parameters must be closely matched to the implant's integrated-circuit power usage. In addition, the bandwidth of the communication channel must be adequate to meet the data collection requirements. This paper describes on-going R&D work for the design and fabrication of smart sensors based upon microtelemetry technology. Presently, sensor designs for two applications are in progress -- EMG and joint angle position.

A silicon-based tactile sensor is fabricated and tested on human subjects. The sensor is packaged in a flexible polyimide-based skin which allows the sensor to be mounted on the finger. The sensor is fabricated using anisotropic and isotropic etching, silicon-to-silicon bonding, and other standard microfabrication technologies. The sensor was characterized on the bench. The sensor is capable of withstanding loads in excess of 100 N. Typical output sensitivity is 1.4 mV/V/N. In addition, a dynamic calibration and a tracking experiment were performed on human subjects. For the human subject experiments the sensor was mounted on the distal phalangeal pad of the thumb on the dominant hand. A dual beam strain gage dynamometer was used to measure the actual forces. The force range for the dynamic calibration was based on the maximum acceptable pinch force of the subjects. Sensor design and operation are described. Statistical analysis of the human subject experiment is presented and future work is discussed.

The halo vest is a head and neck immobilization system that is often used on patients that are recovering from cervical trauma or surgery. The halo vest system consists of a rigid halo that is firmly attached to the skull, an upright support structure for stabilization and immobilization, and a torso-enveloping vest. The main purpose of this study was to measure the forces that are carried by the halo-vest structure as the subject undergoes various activities of daily living and external loading for different vest designs. A tethered strain gage load cell based instrumentation system was used to take these load measurements on ten different subjects. Three different halo-vest systems were evaluated. The primary difference between the vests was the amount of torso coverage and the use of shoulder straps. The loads were measured, analyzed and used to compare the vests and to create a model of halo-vest-neck mechanics. Future applications of this technology to standalone data logging, pin-load measuring and biofeedback applications are discussed.

The human musculoskeletal system represents one of the ultimate manifestations of smart structures capabilities. It can sense, actuate, and heal for periods occasionally in excess of one hundred years. As a natural consequence, research and treatment regimes for a variety of musculoskeletal disorders use technologies that are in many respects very similar to that of the more traditional aerospace and civil smart structures technologies. This paper presents an overview of the technologies that are currently in use in orthopaedic practice and research that mimic these more traditional smart structures. This includes a wide variety of instrumentation that can measure loads and motions within the musculoskeletal system, within prostheses (artificial joints and limbs), and within orthoses (devices that limit the motion of joints and limbs). Included are discussions about the instrumentation of spine and hip implants and the use of telemetry to transmit the data, the measurement of spinal motions through goniometers and surface-attached lordosimeters, and the forces involved in ambulation. In addition to the systems that can measure loads and motions, there are some devices that can measure these quantities and respond in such as way as to control the motions or loads. A 'virtual corset' that provides audio and/or tactile feedback to patients to prevent excessive trunk flexion is described.

The work presented in this session is part of a project to develop an arm-control system based on neuronal activity recorded from the cerebral cortex. This will make it possible for amputees or paralyzed individuals to move a prosthetic arm or, using functional neural stimulation, their own limbs as effortlessly and with as much skill as intact individuals. We are developing and testing this system in monkeys and hope to have a prototype working in the next couple of years. This project has been made more feasible because we have been able, in the last 15 years to extract, from the brain, a signal that represents arm trajectory accurately. In this paper, we describe how this technique was developed and how we use this as the basis for our control signal. An alternative approach using a self-organizing feature map, an algorithm to deduce arm configuration given an endpoint trajectory and the development of a telemetry system to transmit the neuronal data is described in subsequent papers.

The population vector algorithm has been developed to combine the simultaneous direction- related activities of a population of motor cortical neurons to predict the trajectory of the arm movement. In our study, we consider a self-organizing model of a neural representation of the arm trajectory based on neuronal discharge rates. Self-organizing feature mapping (SOFM) is used to select the optimal set of weights in the model to determine the contribution of individual neuron to the overall movement. The correspondence between the movement directions and the discharge patterns of the motor cortical neurons is established in the output map. The topology preserving property of the SOFM is used to analyze real recorded data of a behavior monkey. The data used in this analysis were taken while the monkey was drawing spirals and doing the center out movement. Using such a statistical model, the monkey's arm moving directions could be well predicted based on the motor cortex neuronal firing information.

It was hypothesized that the configuration of the upper limb during a hand static positioning task could be predicted using a dynamic musculoskeletal model and an optimal control routine. Both rhesus monkey and human upper extremity models were formulated, and had seven degrees of freedom (7-DOF) and 39 musculotendon pathways. A variety of configurations were generated about a physiologically measured configuration using the dynamic models and perturbations. The pseudoinverse optimal control method was applied to compute the minimum cost C at each of the generated configurations. Cost function C is described by the Crowninshield-Brand (1981) criterion which relates C (the sum of muscle stresses squared) to the endurance time of a physiological task. The configuration with the minimum cost was compared to the configurations chosen by one monkey (four trials) and by eight human subjects (eight trials each). Results are generally good, but not for all joint angles, suggesting that muscular effort is likely to be one major factor in choosing a preferred static arm posture.

This paper investigates the use of waveform-based acoustic emission (AE) techniques for location of low velocity impact in composite plates. An eight-ply [0 degrees/90 degrees/0 degrees/90 degrees], laminated glass/epoxy plate is investigated. The plate is impacted with a steel ball using a drop tower. Three broadband AE sensors are mounted on the surface of the composite plate. The response signals of the AE transducers are amplified by broadband AE preamplifiers and fed into a fracture wave detector made by Digital Wave Corporation. The signals are instantaneously sampled and stored in a pentium computer. The digitized AE signals are processed in the time and frequency domains. The raw AE signals were pre- processed to remove reflections from the plate boundaries that cause location error. The Gaussian cross-correlation method and Hilbert transformation are used to obtain arrival time differences among sensors and to overcome dispersion problems. Theoretical flexural wave velocities are calculated using the first order shear deformation plate theory. The calculated arrival time differences among sensors and the flexural velocities are used to determine impact location. The experimental results show that the flexural wave is the dominant mode for low velocity impact of the composite plate. Little or no extensional waves were observed. The locations of impact points are successfully determined. The location error varied from 0.00984 to 0.4807 radial inch for the 11.5' by 11.5' composite plate investigated. The method of using the Gaussian cross-correlation technique to determine AE arrival time differences among AE sensors is suitable for AE wave propagation detection due to low velocity impact.

The accuracy of optical encoders is usually limited by interpolation errors. A new design for eliminating these errors is presented. Its distinguished features are: (1) Gratings having sine function transmissivity are used as the index gratings, and as a result the harmonic errors of encoder signals are greatly reduced with the total amount of harmonic components less than 1%. A new method for producing sine index grating of various kinds is also investigated. (2) Non-orthogonal error, unequal amplitudes in the two interpolation signals and their residual dc voltages are corrected in real time by a specially designed error compensator. Non-orthogonal error is reduced to less than plus or minus 1 degree, inequality in amplitudes of the two signals less than plus or minus 2% and residual dc voltages less than plus or minus 20 mv. (3) An easy method for evaluating the interpolation errors without using reference instruments is proposed. Experiments show that a rotary encoder by using the sine index grating and the error compensator could meet the accuracy requirement for interpolation to 0.26% of the encoder's constant.

The thesis introduces one kind of three dimensional scanning probe system for measuring oil tube threads, discourses on measuring theory, mechanical structure, analog and control circuitry and the software system of the system. It has resolved the problem of measuring the threads in high precision with a common coordination machine for the user. The thesis describes the classification and development of 3D coordination probes. In view of the practical problem of the user, the thesis offers an original scheme of adopting a special 3D scanning probe. According to the measuring principle, we carry out mechanical design, fulfill the designing of sensor analog circuit, probe control circuit and communication system, and program corresponding software. The probe adopts an improved three stories form so that it is closely arranged and has complete function; it adopts high precision sensors, an electric and magnetic zero fixing system and a unique spring force-testing drive system. The fine designing has made an on analog circuitry and control system in order to ensure the property and accuracy. The total probe system simultaneously has the principal function of a micro-metrical probe and the special function of the thread measuring. The description of this article is divided into several sections for the system. Except the contents of the article, in the designing of mechanical structure, we also carry out the designing of Z direction self-weight balance and damping structure, force-testing drive and precision zeroing structure, precision positioning and releasing structure and safety protection structure. Those structures adopt a unique version for decreasing volume and enhancing the precision. We design the sensor and oscillatory circuits, phase detecting and impulse trigger circuits. The distinctive features are that the fore- amplifier is mounted in the probe and the total system adopts the suitable chips with high accuracy and reliability. The control and communication systems adopt the intelligence system that is based on INTER 8098 chip microcomputer, the cost is cheap and the reliability is good. In this article, the hardware and software are described individually.

This was an investigation into the feasibility of using liquid core optical fibers for the detection and self repair of cracking in cement or polymer materials generated by dynamic or static loading. These experiments relied on our current research sponsored by the National Science Foundation. That work on the concept of internal adhesive delivery from hollow fibers for repair was here combined with the nondestructive fiber optic analysis of crack location and volume. The combination of the ability to remotely measure crack occurrence in real time and determine the location and volume of crack damage in the matrix is unique in the field of optic sensors. The combination of this with crack repair, rebonding of any detached or broken fibers, and replenishment of liquid core chemicals, when necessary, make this a potentially powerful sensing and repair tool. Work on this research topic was sponsored by the University of Illinois.

An algorithm has been developed to determine deformations of a cantilever honeycomb plate under arbitrary loading conditions. The algorithm utilizes strain information from a set of sixteen fiber Bragg grating sensors, mounted on the plate so that all sensors measure strains along the clamped-free direction. The sensors are interrogated using a wavelength-division multiplexing scheme. A two-dimensional bi-polynomial function which represents the strain field is created using a least-squares algorithm. This function is integrated twice with the known boundary conditions applied to yield the deformation field for the plate. Maximum differences between finite-element solutions and least-squares estimates did not exceed 29.0 percent for any of the 16 investigated load scenarios. However, when considering areas of maximum deflection, the least-squares estimates did not exceed 13.3 percent difference. The algorithms used to interrogate the sensors, perform the strain-displacement calculations, and generate a real-time (approximately 4 Hz) mesh of displacement are encoded in a C program.

We describe an instrumentation system which provides the capability to monitor a large number of Bragg gratings using a common source and a scanning narrowband filter. The system described has the capability to monitor 12 FBG sensors along each of 5 fibers for a total of 60 sensor elements. We demonstrate the use of this system to address multiple sensors embedded in and attached to a quarter scale bridge span.

This paper looks at both the technical and socioeconomic factors that challenge the growth of smart materials and structures (SMS). Topics like human interaction and teaming, the litigious nature of society, the impact 'smart' philosophies are having on commercial sectors, future computing technologies, and others are discussed. This paper necessarily includes some conjecture, with which the reader may or may not agree. In the end, the goal of this paper is to stimulate thought on both the technical and non-technical requirements of developing the technology known as smart structures. The information contained in this paper intentionally draws from sources not commonly seen by practitioners in the SMS field in an effort to broaden awareness and to avoid rehashing technologies and concepts that are discussed time and again in our own literature. Through defendable engineering concepts, popular opinion, and a touch of personal bias, this paper simply tries to divine what it takes to be 'smart.'