In this paper we report the development of a fiber optic corrosion sensing system that complements and/or surpasses the capabilities of conventional corrosion sensing techniques. The sensing system is based on evanescent wave phenomena and in the configured sensor allows for the detection of general corrosion on and within materials. Based on the authors' experience installing may kilometers of fiberoptic sensors into large civil structures such as multistory buildings, hydroelectric dams, and railway/roadway bridges, we are (currently) embedding these sensors into bridge test members -- limited structures that are being subjected to accelerated corrosion testing conditions. Three Vermont Agency of Transportation bridges, one in a low salt use region, one in a medium salt use region, and the third in a high salt use region, are being selected and will be instrumented with these embedded fiber optic corrosion sensors. Monitoring of chloride penetration and rebar corrosion status will be measured during the course of a longitudinal study. The specific sensing mechanism and design for robustness (to allow survival of the embedding process during repaving of the bridges) are discussed and laboratory and initial field results are presented.

The application of a state-of-the-art fiber optic sensing system for the quantitative analysis of strain in strands used in prestressed concrete is proposed. Compressive stress in concrete is used to counterbalance any tensile force due to loading, which might lead to cracking or deflection. In pre-tensioning prestressed concrete, a tendon is tensioned before concrete is placed and the prestress is transferred to the concrete after it has cured by releasing the tension on the tendon. In linear prestressing it is often required to determine the axial strain on the tendon during the initial procedure of pre-tensioning, so that the required longitudinal force to achieve maximum concrete strength, can be accurately determined. Conventional techniques for this purpose involve the use of conventional foil strain gages, which are not only expensive to use, but are also known for their failure rate in high strain environments. We discuss the absolute extrinsic Fabry-Perot interferometer (AEFPI) fiber optic sensing system for monitoring strain in pretensioned tendons while this tendon is being loaded. The experiments performed at the Turner Fairbanks Federal Highway Administration at McLean, Virginia exhibit the survivability of the EFPI sensor at strain in excess of 12,000 (mu) (epsilon) while being attached to the tendon surface. The results are compared to those obtained from a collocated foil strain gage and excellent correlation is obtained. Applications of the AEFPI system to high performance smart materials and structures are analyzed and future work in this area is discussed.

This paper describes preliminary laboratory strain monitoring results of a 15-ft tall (2 ft in diameter) circular flexural lap-splice concrete column wrapped with a carbon-composite jacket. Optical fiber time domain (OFTD) strain sensors were embedded in the composite jacket in the hoop direction to monitor strains during testing. Optical fiber time domain methods are based on measuring the time-of-flight of optical signals launched into optical fibers which are embedded in or attached to flexible structures. This fiber-optic-based method allows in-line sensor multiplexing for distributed sensing along a single optical fiber strand. Time-domain multiplexing stands out as the most reliable and cost-effective approach to monitor strains within large civil structures over prolonged periods of time. In this study, the system utilizes novel fiber optic strain gauge patches. The strain values measured by the fiber optic strain gauge patches were compared to the values measured by conventional surface- mounted strain gauges. Excellent correlation between the fiber optic sensor data and the strain gauge data was obtained during the test.

Monitoring of the structure condition with the long-term goal of a `self-supervising' building presupposes sensor techniques which yield measuring information about the building characteristic and open up possibilities to detect beginning damage. This paper is to demonstrate the activities of the IEMB in Berlin for use of fiber optic sensors in order to get more efficient information about quality parameters of buildings and structure components. The necessary steps for a successful integration of fiber sensors into structures -- from lab experiments via tests on large parts of structures in the test hall to applications on real structures -- are described. The preconditions for a sensor design which is suitable for embedment into mortar bodies for construction-accompanying measurements as well as for a handable application on steel structures on-site are presented. Examples of applications of fiber optic sensors for experimental investigations on large elements, and for measurements during construction and on specially test loaded bridges are described.

Structural health monitoring systems have been designed and evaluated in the laboratory for installation in several bridges and commercial buildings. The systems employ solid-state sensor elements which experience a strain-dependent phase transformation from a metastable, nonmagnetic, austenitic phase to the stable, ferromagnetic, martensitic phase. The irreversible phase transformation is useful for indicating the level of peak structural strain experienced in a particular monitored location. Some of the sensor material characteristics and details related to the phase transformation are discussed as applied to structural health monitoring. The design of representative systems for bridges and commercial buildings is included. Important system(s) features and design capabilities are discussed. Finally, the evaluation of passive and active systems in the laboratory is discussed. The results of experiments detailing the behavior of these systems under uniaxial tension and cyclic loading conditions are presented.

A prototype optical fiber sensor for monitoring corrosion on large steel structures has been designed and tested with favorable results. The sensor works by pulling a multimode fiber into a tight bend and securing it with a `corrosion fuse.' When the fuse corrodes, it eventually breaks and allows the fiber to straighten. The resulting difference in optical intensity emerging from the fiber is measurable using an OTDR or other optical detector. Initial experiments were carried out to determine the effect of bending fibers in a small radius and showed the feasibility of the device. Following, tests were performed on three in-line sensors in a simulated corrosive atmosphere and showed that this cheap and easily implemented monitoring scheme could be used to infer the presence of corrosion at different locations, and/or the degree of corrosion at a single location.

As damage accumulates in structure, the stiffness of the structure changes. Changes in stiffness are reflected in changes in the frequencies and mode shapes vibration of the structure. A theory of structural damage evaluation in which changes in the dynamic characteristics of a structure are used to predict the location and severity of damage has been developed by Stubbs et al. The objective of this study is to investigate the feasibility of using such technique for detecting loss of prestress in a prestressed concrete bridge. In order to achieve this goal the following tasks were performed: (1) a review of the state of the art modeling of prestressed concrete using finite elements, (2) a finite element model of a 3-D prestress beam using a commercial finite element code (ABAQUS/PATRAN) was developed, (3) dynamic modal analyses of the undamaged and damaged (less prestress) models are performed, (4) evaluation of an existing damage location algorithm for predicting location and severity of damage. The inability of the algorithm to predict loss of prestress is observed and explained. An improved version of the algorithm is qualitatively presented.

An analysis of artificial neural networks on damage assessment of an aluminum cantilever beam was conducted. The neural networks were trained and tested with deterministic data of resonant frequency information to test their ability in determining the magnitude, location and type of damage on the beam. Being a preliminary study, no experimental data has been included, since no information was found in the literature where neural networks were used in determining the type of damage on a structure. This paper includes a discussion on the theory of neural network and the process involved in developing the architecture for three layer backpropagation neural networks for damage assessment. The neural networks were tested for three types of damage using four damage magnitudes.

The merits and limitations of vibrationally based non-destructive (NDD) methods such as frequency response only, displacement time history, system identification, acoustic emission, and work at Texas A&M University which includes the damage sensitivity or damage index method are reviewed with respect to their applicability in detecting damage to a floating structure and the findings stated. One of these methods, namely the damage sensitivity method, is reported to be suitable for detecting structural damage, with a slight adjustment, for floating platforms. Due in part to the fact that floating structures depend on the surrounding aqueous environment for their vertical support, the detection of flooding, conceivable as mass damage, is stated to be remaining outside the scope of the existing methods. A sensitivity based mass damage index is formulated for the latter case. The performance feasibility of the stiffness and mass damage indicators when applied to a floating structure, as exemplified by a bridge, is demonstrated.

Effects of experimental uncertainties on the detectability of structural damage using a frequency sensitivity method are examined. An existing theory of global damage detection based on resonant frequency changes is modified to account for the uncertainties in the frequency measurements. The effects of the uncertainties are studied by simulating a number of NDE inspections on a cantilever beam subjected to a single damage at several locations using Monte Carlo techniques. The statistics of the final damage predictions are obtained through the simulation results. The probability of detecting various levels of damage along the beam is examined by defining successful detection events.

A comprehensive market survey and laboratory evaluation was conducted for the Ohio Department of Transportation and the Federal Highway Administration to identify the most promising sensors and data-acquisition systems for infrastructure application. A pilot system for highway bridge monitoring was implemented on a typical steel-stringer bridge in Cincinnati for high-speed traffic and long-term environmental monitoring. Static tests were performed with known truck loads to confirm monitoring results and to calibrate finite-element and section analysis models of the bridge. A complete and accurate characterization of the as- is structural condition has been related to the structural capacities and reliability. This multi- disciplinary research improves our understanding of the actual loading environment and the corresponding bridge responses. Instrumented monitoring is expected to complement inspection methods, provide an objective measure of the state-of-health, and alert officials to bridge deterioration or failure.

In this paper, a robust algorithm to detect and localize damage in full-scale plate-girder bridges by utilizing modal response parameters of undamaged (as-built) and damaged structures is presented. In the first part, a damage detection algorithm which yields information on the location of damage directly from changes in mode shapes of structures is outlined. In the next part, the algorithm is implemented to detect and localize damage in a real highway bridge (on the US Highway I-40) located in Bernalillo County, New Mexico, USA. In this damage detection exercise, pre-damage and post-damage modal parameters of only a single mode measured from the test structure are utilized to localize the damage.

This paper deals with determination of damage location and magnitude evaluation by measuring the natural frequencies of a structure. A uniform transverse crack, located on a steel cantilever prismatic beam of rectangular cross section, is used as the model for damage. The crack on the beam is assumed to be open during the transverse vibration of the structure, thus the effect of closing and opening of the crack is ignored in the analysis. Euler-Bernoulli beam behavior is modeled and the damage effect (crack location and depth) is modeled as a torsional spring. The equations of motion of the beam with the crack are developed from Hamilton's principle. The natural frequencies for various crack depths and locations are calculated to establish a collection of failure sets. The response from an actual beam is then compared to the collection of failure sets. The failure set that most nearly approximates the actual response is used to identify the crack location and depth. This method is confirmed with the experimental vibration of a steel cantilever beam. This beam has a single damage location and the damage (crack depth) is increased by fatigue loading. The method is valid as long as the response of any of the failure sets do not overlap into the response of other failure sets. The accuracy of this procedure depends upon the type of failure sets used, (mathematical model to predict response) and the response of the system to the actual failure. The advantages and disadvantages of the method are discussed.

Current analytical damage detection schemes for Euler-Bernoulli beams are based on the modal parameters generated from a one dimensional finite element model. However, since damage has been limited to an effective loss in bending stiffness at the damage location, the interpretation of the damage has not been addressed in any great detail. The objective of this paper is to investigate the effects of damage size and location in a beam on the apparent damage inflicted on adjacent locations.

Fundamental principles of mechanics have recently been brought to bear on problems concerning very large structures. Fields of study include tectonic plate motion, nuclear waste repository vault closure mechanisms, the flow of glacier and sea ice, and highway bridge damage assessment and residual life prediction. Quantitative observations, appropriate for formulating and verifying models, are still scarce however, so the need to adapt new methods of experimental mechanics is clear. Large dynamic systems often exist in environments subject to rapid change. Therefore, a simple field technique that incorporates short time scales and short gage lengths is required. Further, the measuring methods must yield displacements reliably, and under oft-times adverse field conditions. Fortunately, the advantages conferred by an experimental mechanics technique known as speckle photography nicely fulfill this rather stringent set of performance requirements. Speckle seemed to lend itself nicely to the application since it is robust and relatively inexpensive. Experiment requirements are minimal -- a camera, high resolution film, illumination, and an optically rough surface. Perhaps most important is speckle's distinct advantage over point-by-point methods: It maps the two dimensional displacement vectors of the whole field of interest. And finally, given the method's high spatial resolution, relatively short observation times are necessary. In this paper we discuss speckle, two variations of which were used to gage the deformation of a reinforced concrete bridge structure subjected to bending loads. The measurement technique proved to be easily applied, and yielded the location of the neutral axis self consistently. The research demonstrates the feasibility of using whole field techniques to detect and quantify surface strains of large structures under load.

Electronic shearography (ES) is a laser based non-destructive testing method that has the potential to be developed into a commercial bridge monitoring technique. The primary advantage of ES over other similar techniques like electronic speckle pattern interferometry (ESPI) is its decreased sensitivity to in-plane rigid body movement and vibrations. Bridge inspection with ES has proven to be a daunting task so far. The main problem has been the inability of the method to handle the large deflections and vibrations that might be expected in a typical bridge subjected to normal service loads. Earlier research has shown that the extent of in-plane movement that can be tolerated by the system is dependent on the speckle size. The speckle size also affects the fringe quality by imposing resolution requirements on the imaging device. This article shall undertake the study of speckle size as a function of the focal length of the imaging lens, object distance and illumination wavelength using high resolution holographic film and a high magnification optical microscope.

The current quality of our nations bridges is on a decline. There are roughly half a million highway bridges in the United States and out of the half a million more than 200,000 are deficient. With catastrophic failure of bridges causing the loss of life and property, the need for bridge inspection and maintenance is evident. When the Silver Bridge that crossed the Ohio River collapsed in December 1967, 46 people were killed. The failure to prevent the disaster was attributed to the poor inspection techniques used by the bridge inspectors. Current inspection techniques depend on humans being able to recognize structural imperfections without the aid of instrumentation. The Federal-Aid Highway Act of 1968 mandated both national bridge inspection standards and training for bridge inspectors. This act has encouraged the development of instruments that would allow inspectors to perform more complete inspections of bridges. To improve the quality of inspection and data, there is a great need for proven methods and instruments used to acquire data. The Laser Optical Displacement System (L.O.D.S.) developed at New Mexico State University by the Optical and Materials Science Lab is such a device. The L.O.D.S. has been tested and proven in both laboratory situations and in the field. This paper describes some of the methods that are now being used to measure deflections in bridges. Then, a description of the development and application of the L.O.D.S. unit is given.

The impact of model uncertainty on the damage localization accuracy of a nondestructive damage detection technique when applied to a full-scale plate-girder bridge is assessed. First, a nondestructive damage detection algorithm to localize damage from a few mode shapes of structures is outlined. Next, damage localization exercises are performed for a real highway bridge (on the US Highway I-40) located in Bernalillo County, New Mexico, USA. Finally, the accuracy of damage localization results for the test structure is assessed as a function of the model uncertainty of the damage localization procedure for the structure.

An algorithm originally used to locate errors in finite element models is applied to a full scale bridge damage detection experiment. The method requires experimental frequency response function data measured at discrete locations along the major bridge load paths. In the bridge damage application the algorithm is most effective when applied to static flexibility shapes estimated with a truncated set of six mode shapes rather than individual mode shapes. The algorithm compares `before damage' and `after damage' data to locate physical areas where significant stiffness changes have occurred. A damage indicator shows whether damage is detectable. Damage is correctly located in the two most significant damage cases using the driving point static flexibility estimates. Limitations of the technique are addressed. The damage detection experiment was performed on a three span steel girder bridge that was 425 feet long. This bridge was part of Interstate 40 across the Rio Grande. The New Mexico State University Department of Civil Engineering organized the experiment. The frequency response functions were collected by Los Alamos National Laboratories personnel. The bridge excitation was provided by Sandia National Laboratories.

A non-destructive evaluation technique based on full-field dynamic strain distribution information is presented in this paper. It is well known that the fatigue damage in advanced composites typically manifests itself in the form of a gradual reduction in local stiffness. The reduction of local stiffness will change the strain distribution at the corresponding area. By comparing the strain distribution over the surface of the structure, the fatigue damage in the structure can be detected and located. In the present study, a numerical simulation has been carried out to show that the fatigue damage in the structure can be detected and located by observing the changes in the strain distribution over the structural surface. The changes of the strain distribution are usually related to the damage pattern. It is found that when the damage is in the form of local stiffness reduction, strain distributions for all strain components will change in the same order of magnitude. When the damage is in the form of a crack, only the strain component which is perpendicular to the crack has a significant change. This suggests that it is possible to determine the damage pattern by examining strain distribution changes in all strain components.

We describe the design and performance of a prototype fiber Bragg grating demodulation system based on the use of a scanning fiber Fabry-Perot filter. The computer driven system is capable of demodulating several arrays of wavelength division multiplexed gratings at various scanning rates for real-time strain display and data logging. The instrument represents a new measurement tool which should be useful in a variety of structural health monitoring applications. Results obtained by the system in several applications are presented and system performance limitations are discussed.

This paper describes the use of short gage length optical fiber sensors for the measurement of relative strain in rebar-type and cable-type bolts used in the mining industry to support unstable material or to keep rock masses together. For research purposes, and in order to develop numerical models for analyzing rockmass behavior, these bolts are instrumented with strain gages. For cable-type bolts, for example, traditional gages, such as foil strain gages, have proven to be inefficient or unreliable and new sensing methods have to be developed. Optical fiber-based sensors, which were surface mounted, using an epoxy resin adhesive, inside a small groove of the king wire of a 7-strand cable, could offer an attractive option for such applications. Tests include the fatigue loading of the instrumented rebar and cable bolt specimen. Stress-strain curves for both the rebar-type and cable-type bolts are provided. Practical applications and limitations of such sensor methods for bolt analysis in the future are described as well.

Structures, however well designed, may experience excessive cycling or variations of design loads. These could result in higher stress levels causing accelerated deterioration or `wear' in these structures. Eventually, these types of deterioration may cause defects or damage in the structure. In case of a bridge, if not detected on time, these types of deterioration may result in catastrophic failures. This paper proposes to develop techniques that will determine the prevailing condition of a structure in a way that will indicate its structural integrity in a quantifiable manner. These techniques will assist in evaluating the existing condition of and will help to design a structure better as well. The performance of a structure can also be predicted by these methods. These techniques will include modeling of a structure, nondestructive testing, and utilization of statistical estimation methods. Most importantly, the methods will provide a `tool' for personnel in charge of the inspection of such structures, bridge inspectors for example, to perform inspections in a more objective and scientific way. Presently, these inspections are generally carried out visually and in a subjective manner. By taking a deflection and/or strain measurement at a point, a bridge inspector, with the help of a set of curves, will be able to quantify the existing condition of that structure compared to its original or new state. Equipped with this tool and periodic checking, an inspector will then be able to assess deterioration of a structure over time. Historical data will then be available for the inspection team to evaluate the performance of the structure. If the data show more than expected `wear,' appropriate action can be taken to avoid occurrences of damages or defects and in extreme cases total replacements of a structure. These actions will not only save resources such as money, material, and time, it may save lives as well.

A nondestructive vibratory technique has been used to record the vibratory signature of electric pylons. It evidences the principal modes of vibrations. The vibration magnitude depends on the impact location and the loading amplitude. The appearance of energy dissipation mechanisms, corresponding to a damaging process, is readily detected by a change of characteristics of the recorded acceleration signals. A method based on Volterra series provides a ready mean to analyze the nonlinear response and permits us to monitor the durability of the pylon in service. The presented technique relies on the evaluation of a scalar called `nonlinear contrast,' indicating the significance of nonlinear behavior of the structure under impact loading, in comparison with a reference state. The calculation requires the identification of Volterra kernels through a multi-amplitude procedure. The fact that an unscrewed bolt has been evidenced shows that this technique is very sensitive. The second major experimental result is that the nonlinear contrast appears to be independent on the accelerometer location and direction.

Representation of on-orbit microgravity environment in a 1-g environment is a continuing problem in space engineering analysis, procedures development and crew training. A way of adequately depicting weightlessness in the performance of on-orbit tasks is by a realistic (or real-time) computer based representation that provides the look, touch, and feel of on-orbit operation. This paper describes how a facility, the Systems Engineering Simulator at the Johnson Space Center, is utilizing recent advances in computer processing power and multi- processing capability to intelligently represent all systems, sub-systems and environmental elements associated with space flight operations. It first describes the computer hardware and interconnection between processors; the computer software responsible for task scheduling, health monitoring, sub-system and environment representation; control room and crew station. It then describes, the mathematical models that represent the dynamics of contact between the Mir and the Space Shuttle during the upcoming US and Russian Shuttle/Mir space mission. Results are presented comparing the response of the smart, active pilot-in-the-loop system to non-time critical CRAY model. A final example of how these systems are utilized is given in the development that supported the highly successful Hubble Space Telescope repair mission.

Shannon's sampling theorem is used here for the first time to reconstruct the mode shapes, resulting from equidistantly spaced sampling points obtained in the field, associated with building structures. The feasibility of reconstructing mode shapes of a structural system using a minimum number of sensors is investigated. To evaluate the feasibility of the method, the first three mode shapes of a simply supported beam and a mode shape represented by a parabolic equation are reconstructed. We find that the error between the true mode shapes and the reconstructed mode shapes reduces significantly as the number of the repeats of sampling points increases. The comparisons between exact and reconstructed mode shapes are presented, in addition to values for the modal assurance criteria (MAC) for the exact and reconstructed modes. The practicality of using these reconstructed mode shapes obtained via Shannon's sampling theorem is demonstrated using a recently developed nondestructive damage localization algorithm. The results show that the reconstructed mode shapes, generated using a minimum number of sensors, can indeed be used to localize damage.

An optical technique has been developed whereby two angles and linear displacement can be simultaneously measured in a non-contact manner. The method depends upon the usage of a diffraction grating with linear variation of period along its length. The grating is attached to a structure at a point of interest while all other system components are placed at a remote location. Evaluation of this measurement technique has been demonstrated on a laboratory- based structure which simulated conditions found at deep trench (or tunnel) walls or bracing systems. In a construction site configuration, this sensor allows the user to determine if the walls are undergoing structural deformation. In addition, the magnitude of deformation may be measured and alarm conditions may be monitored. Experimental results obtained using this technique are presented and compared with theory.

Routine acquisition of data from instrumented civil structures need not require onsite presence. We have developed a sensor interrogation method which utilizes the Internet global computer network as the information conduit from sensor(s) to user. It is therefore possible for the data monitoring to be performed at a remote location with the only requirements for data acquisition being Internet accessibility. Such a system may prove quite advantageous when a single user, or small group, is required to acquire and analyze data from several instrumented structures which are geographically very separated. In this developed system, it is possible to remotely acquire raw sensor data as well as realtime video/audio images of the current status at the instrumented structure.

Bridge integrity is recognized as a complex issue, with visual inspection currently being the prime technique for determining structural health. For instrumentation to contribute to such determination, high cost effectiveness is mandatory. Potential bridge instrumentation functions are outlined with a cost comparison between conventional strain gages and a distributed fiber optic system. Because of the lack of a static model to assess overall bridge integrity from localized strain/displacement data, collaboration is suggested with dynamic methods to mutually determine structural health on as a 3-D or global basis. Examples of several types of bridges are presented, with suggested fiber optic instrumentation. Known fiber optic bridge installations are reviewed and two new fiber optic sensing developments are presented as uniquely applicable to cost-effective bridge integrity determination.