Recent years have seen a concurrent development of new sensor technologies and high-fidelity modeling capabilities. At the junction of these two topics lies an interesting opportunity for real-time system monitoring and damage assessment of structures. During monitoring, measurements from a structure are used to learn the parameters and equations characterizing a physics-based model of the system; thus enabling damage identification. Since monitored quantities are physical, these methods offer precious insight into the damage state of the structure (localization, type of damage and its extent). Furthermore, one obtains a model of the structure in its current condition, an essential element in predicting the future behavior of the structure and enabling adequate decision-making procedures. This dissertation focuses more specifically on solving some of the challenges associated with the use of online Bayesian learning algorithms, also called sequential filtering algorithms, for damage detection and characterization in nonlinear structural systems. A major challenge regarding online Bayesian filtering algorithms lies in achieving good accuracy for large dimensional systems and complex nonlinear non-Gaussian systems, where non-Gaussianity can arise for instance in systems which are not globally identifiable. In the first part of this dissertation, we show that one can derive algorithmic enhancements of filtering techniques, mainly based on innovative ways to reduce the dimensionality of the problem at hand, and thus obtain a good trade-off between accuracy and computational complexity of the learning algorithms. For instance, for particle filtering techniques (sampling-based algorithms) subjected to the so-called curse of dimensionality, the concept of Rao-Blackwellisation can be used to greatly reduce the dimension of the sampling space. On the other hand, one can also build upon nonlinear Kalman filtering techniques, which are very computationally efficient, and expand their capabilities to non-Gaussian distributions. Another challenge associated with structural health monitoring is the amount of uncertainties and variabilities inherently present in the system, measurements and/or inputs. The second part of this dissertation aims at demonstrating that online Bayesian filtering algorithms are very well-suited for SHM applications due to their ability to accurately quantify and take into account these uncertainties in the learning process. First, these algorithms are well-suited to address ill-conditioned problems, where not all parameters can be learnt from the available noisy data, a problem which frequently arises when considering large dimensional nonlinear systems. Then, in the case of unknown stochastic inputs, a method is derived to take into account in this sequential filtering framework unmeasured stationary excitations whose spectral properties are known but uncertain.

Bridge monitoring studies indicate that the quasi-static response of a bridge, while dependent on various input forces, is affected predominantly by variations in temperature. In many structures, the quasi-static response can even be approximated as equal to its thermal response. Consequently, interpretation of measurements from quasi-static monitoring requires accounting for the thermal response in measurements. Developing solutions to this challenge, which is critical to relate measurements to decision-making and thereby realize the full potential of SHM for bridge management, is the main focus of this research. This research proposes a data-driven approach referred to as temperature-based measurement interpretation (TB-MI) approach for structural performance evaluation of bridges based on continuous bridge monitoring. The approach characterizes and predicts thermal response of structures by exploiting the relationship between temperature distributions across a bridge and measured bridge response. The TB-MI approach has two components - (i) a regression-based thermal response prediction (RBTRP) methodology and (ii) an anomaly detection methodology. The RBTRP methodology generates models to predict real-time structural response from distributed temperature measurements. The anomaly detection methodology analyses prediction error signals, which are the differences between predicted and real-time response to detect the onset of anomaly events. In order to generate realistic data-sets for evaluating the proposed TB-MI approach, this research has built a small-scale truss structure in the laboratory as a test-bed. The truss is subject to accelerated diurnal temperature cycles using a system of heating lamps. Various damage scenarios are also simulated on this structure. This research further investigates if the underlying concept of using distributed temperature measurements to predict thermal response can be implemented using physics-based models. The case study of Cleddau Bridge is considered. This research also extends the general concept of predicting bridge response from knowledge of input loads to predict structural response due to traffic loads. Starting from the TB-MI approach, it creates an integrated approach for analyzing measured response due to both thermal and vehicular loads. The proposed approaches are evaluated on measurement time-histories from a number of case studies including numerical models, laboratory-scale truss and full-scale bridges. Results illustrate that the approaches accurately predicts thermal response, and that anomaly events are detectable using signal processing techniques such as signal subtraction method and cointegration. The study demonstrates that the proposed TB-MI approach is applicable for interpreting measurements from full-scale bridges, and can be integrated within a measurement interpretation platform for continuous bridge monitoring.

Monitoring the integrity of civil engineering structures is an imperative aspect of public safety, since structural failures can pose serious threats to life and property. Periodic inspection performed throughout the life span of these structures is also vital for a nation’s economy. Substantial sums of money may be saved upon detecting structural deterioration in a timely manner. Nondestructive damage evaluation (NDE) offers effective and economically feasible solutions to perform such tasks. Better predictions can be made regarding the current state of structures, and structurally deficient regions that need immediate attention may successfully be narrowed down by utilizing NDE. For these reasons, a considerable amount of research has been conducted in the field of NDE over the past few decades. As a result, many different methodologies are now available, and many new ones continue to emerge as the need for better evaluation techniques prevails. Upon reviewing the NDE methodologies proposed to date, it may be concluded that theories based on the fundamental equations of mechanics and mathematics in conjunction with justifiable assumptions provided the best results compared to the algorithms developed pragmatically. The goal of this study is to provide NDE methodologies that simultaneously identify the location, the extent, and the severity of damage in general beams. By general beams, we mean beyond Euler-Bernoulli beams (i.e. slender beams) to deep beams and stubby beams whose response may be based on the Timoshenko beam theory, and the Theory of Elasticity. After presenting the governing equations of equilibrium and stress-displacement relations of the fundamental beam theories including the Euler-Bernoulli Beam theory, the Timoshenko beam theory, and the beam theory based on linear Elasticity Theory, mathematical expressions which relate physical properties (e.g. stiffness) of the undamaged and damaged structure to measurable response quantities (e.g. displacement, strains, etc.) are developed. We believe that these algorithms will lead to earlier and more accurate prediction of damage in critical structures. The findings of this work will also lead to a better understanding of the limitations of the currently proposed NDE techniques. In addition, it is anticipated that by incorporating the methodologies proposed in this study to the continuous health monitoring of structural systems could reduce the cost of maintenance and offer safer infrastructure networks.

The main types of faults studied in the literature are commonly categorized as electrical faults and mechanical faults. In addition to well known faults, the performance of a diagnostic algorithm and its operational reliability in harsh environments has been another concern. In this work, the reliability of an electric motor diagnosis signal processing algorithm itself is studied in detail under harsh industrial conditions. Reliability and robustness of the diagnosis has especially been investigated under 1) potential motor feedback error; 2) noise interference to a diagnosis-relevant system; 3) ease of implementation; and 4) universal application of diagnostic scheme in industry. Low cost and flexible implementation strategies are also presented. 1) Signature-based diagnosis has been performed utilizing the speed feedback information which is used to determine fault characteristic frequency. Therefore, feedback information is required to maintain high accuracy for precise diagnosis which, in fact, is not the case in a practical industrial environment due to industrial noise interferences. In this dissertation, the performance under feedback error is analyzed in detail and error compensation algorithms are proposed. 2) Fault signatures are commonly small where the amplitude is continuously being interfered with motor noise. Even though a decision is based on the signature, the detection error will not be negligible if the signature amplitude is within or close to the noise floor because the boundary noise level non-linearly varies and, hence, is quite ambiguous. In this dissertation, the effect of noise interference is analyzed in detail and a threshold design strategy is presented to discriminate potential noise content in diagnosis. 3) The compensating procedure of speed feedback errors and electrical machine current noise, characteristics which are basically non-stationary random variables, requires an exhaustive tracking effort. In this dissertation, the effective diagnosis implementation strategy is precisely presented for digital signal processor (DSP) system application. 4) Most of the diagnosis algorithms in the literature are developed assuming specific detection conditions which makes application difficult for universal diagnosis purposes. In this dissertation, by assuming a sinusoidal fault signal and its Gaussian noise contents, a general diagnosis algorithm is derived which can be applied to any diagnostic scheme as a basic tool.

The objective of this research is to develop a theoretical Structural Impairment Detection System (SIDS) for timber railway bridges. Due to fatigue, the timber stringers in timber railway bridges develop shear cracks. These shear cracks lead to higher bridge deflections, higher stresses in the stringers and rail, and shorter fatigue life of the system. A SIDS is proposed which links wheel path accelerations obtained from traversing freight cars to the condition of the bridge. In order to develop the SIDS, two models of timber railway bridges with various levels of structural impairment were developed. The first model was a quasi-static model developed from classical beam theory and implemented in MATLAB. The second model was a dynamic, finite element model created in LS-DYNA. Traversing axle loads were imposed on the models. The results obtained from the model were the wheel paths the axles take as they traverse the bridge. The paths were expressed as vertical displacements as a function of position on the bridge. Wheel path accelerations were obtained by numerically differentiating the vertical displacements. The accelerations were then used to train neural networks to have an input of an acceleration vector and an output of a bridge condition vector. The neural networks were trained on results from both models under three train speeds: 40 mph, 30 mph, and 20 mph. The networks were able to determine the correct bridge condition 90% of the time when the train speed was 40 mph and 70% of the time when the train speed was 30 mph. The networks were not successful in determining bridge condition when the train speed was 20 mph.

As the UH-1 aircraft continue to age, there is growing concern for their structural integrity. With corrosion damage becoming a bigger part of the sustainment picture with increasing maintenance burden and cost, it is becoming increasingly important for corrosion management to be updated with more advanced techniques. The current find-and-fix technique for handling corrosion has many shortfalls, spurring the recent interest in early detection through structural health monitoring. This condition based technique is becoming more prevalent and is recognized for the potential to greatly reduce maintenance cost. Through corrosion monitoring, structural and environmental conditions can be closely observed, preventing excessive maintenance action and saving cost. Searches for corrosion monitoring system designs revealed several commercial companies with prototype systems installed on commercial aircraft, however, details on system design and data analysis were scarce. This study attempted to bridge the gap in literature by providing insight into the development of a corrosion damage prediction model and the design of a corrosion monitoring system. This study attempted to use aircraft maintenance data to make prediction models for determining what corrosion damage an aircraft can expect, given varying operating conditions. Although a reliable prediction model could not be created, trends observed in the data were still valuable for identifying problematic areas of the aircraft. In order to create reliable models, more accurate corrosion data is needed. This can be accomplished through the implementation of a corrosion monitoring system. A custom corrosion monitoring system was designed for the UH-1 aircraft. Commercial off-the-shelf products were fit to the design and a benefits-to-cost analysis was performed for the monitoring system, evaluating the system based on criteria developed from user requirements. The system proved to meet and exceed expectation, making it an ideal choice for the UH-1 aircraft.

Highway bridges are vital links in any transportation network. Identifying the possible safety problems in the approximately 600,000 bridges across the U.S. is generally accomplished through labor-intensive, visual inspections. Wireless monitoring technology seeks to improve current practices by supplementing the visual inspections with real-time evaluation of bridges. To be economically feasible, wireless sensor networks should be able to (a) operate independent of the power grid, and (b) achieve a service life of at least ten years. Novel energy harvesting approaches have been investigated to fulfill these two criteria. In particular, the feasibility of a vibration energy harvester as a long-term power source was assessed. The goal of the research was to process measured acceleration data and analyze the vibrational response of typical highway bridges under truck loads. The effects of ambient temperature, truck traffic patterns, and harvester position on the power content of the vibrations were explored, as well as the effects of linear and nonlinear harvesters. This thesis presents the results of evaluating the response of five steel bridges in Texas and Oregon for energy harvesting applications. / text

The proposed study explores the feasibility of detecting and quantifying corrosion and delamination (physical separation) at the interface between reinforcing steel bars and concrete using ultrasonic guided waves. The problem of corrosion of the reinforcing steel in structures has increased significantly in recent years. The emergence of this type of concrete deterioration, which was first observed in marine structures and chemical manufacturing plants, coincided with the increased applications of deicing salts (sodium and calcium chlorides) to roads and bridges during winter months in those states where ice and snow are of major concern. Concrete is strengthened by the inclusion of the reinforcement steel such as deformed or corrugated steel bars. Bonding between the two materials plays a vital role in maximizing performance capacity of the structural members. Durability of the structure is of concern when it is exposed to aggressive environments. Corrosion of reinforcing steel has led to premature deterioration of many concrete members before their design life is attained. It is therefore, important to be able to detect and measure the level of corrosion in reinforcing steel or delamination at the interface. The development and implementation of damage detection strategies, and the continuous health assessment of concrete structures then become a matter of utmost importance. The ultimate goal of this research is to develop a nondestructive testing technique to quantify the amount of corrosion in the reinforcing steel. The guided mechanical wave approach has been explored towards the development of such methodology. The use of an embedded ultrasonic network for monitoring corrosion in real structures is feasible due to its simplicity. The ultrasonic waves, specifically cylindrical guided waves can propagate a long distance along the reinforcing steel bars and are found to be sensitive to the interface conditions between steel bars and concrete. Ultrasonic transducers are used to launch and detect cylindrical guided waves along the steel bar.In this dissertation, in-situ corrosion monitoring technique for reinforced concrete is developed based on two methods - 1) variation of signal strength and 2) the time-of-flight (TOF) variations as the corroded member is loaded transversely. This is the first attempt ever to monitor corrosion inside concrete by measuring the change in the time of flight of guided waves along reinforcing bars as the concrete beam is subjected to bending. Advantages of corrosion monitoring by TOF change are discussed in the dissertation.

Ageing mechanical, civil, aerospace, marine and offshore structures require continuous and accurate assessment on their integrity to avoid potentially hazardous failures. To further facilitate this crucial demand, a new technical terminology, generally referred to as structural health monitoring (SHM) has been coined in three past decades. SHM involves deployment of a sensory network on such structures in order to gather useful data, such that processing and interpreting the data through specific algorithms would enable one to detect defects and anomalies within the structures. This dissertation presents the results of a series of efforts expended towards the refinement and enhancement of a vibration-based SHM technique, which was originated within our research group. In the adopted damage detection scheme, vibration data are gathered from structures via piezoelectric sensors. Data are processed by a robust signal processing approach, known as the empirical mode decomposition (EMD) in order to establish energy-based damage indices (EMD_EDIs). Interpretation of the damage indices enables detection of onset, location and advancement of defects within structures. A series of adjustments and modifications were devised and implemented to the application of the originally developed methodology, such that, besides increasing the methodology’s robustness and accuracy, they also facilitate a remote vibration-based SHM targeting onshore and offshore pipelines. The integrity of the method in detection of bolt-loosening in a bolted flange joint of a full-scale pipeline was verified through numerical simulations and experimental investigations. The source of a significant inconsistency reported in the previous trials was identified and resolved. Also, for the first time, the remote application of the technique was facilitated by incorporating an advanced wireless data acquisition system. Moreover, the application of the methodology was extended to detection of cracks in girth-welds of offshore pipelines. In this regard, a comprehensive discussion is first provided, which identifies the role of parameters that influence the accuracy of numerical modeling of the dynamic response of submerged structures. The experimental and numerical investigation following the aforementioned modeling efforts presents encouraging results in detection of an advancing notch in the girth-weld of a submerged pipe. The use of a piezoelectric-based excitation technique, incorporated for the first time in the application of the methodology would evidence the enhanced practicality and robustness of the approach. The study concludes with a successful detection of a real-life sharp propagating crack in a beam due to cyclic loadings.

No description available.