This research reviews and extends the impedance-based structural health monitoring technique in order to detect and identify structural damage on various complex structures. The basic principle behind this technique is to apply high frequency structural excitations (typically higher than 30 kHz) through the surface-bonded piezoelectric transducers, and measure the impedance of structures by monitoring the current and voltage applied to the transducers. Changes in impedance indicate changes in the structure, which in turn can indicate that damage has occurred. Several case studies, including a pipeline structure, a composite reinforced aluminum plate, a precision part (gear), a quarter-scale bridge section, and a steel pipe header, demonstrate how this technique can be used to detect damage in real-time. A method to process impedance measurements to prevent significant temperature and boundary condition changes registering as damage has been developed and implemented. Furthermore, the feasibility of using the technique for high temperature structures and for condition monitoring of critical facilities subjected to a severe natural disaster has been investigated. While the impedance-based structural health monitoring technique indicates qualitatively that damage has occurred, more information on the nature of damage is necessary for remote structures. In this research, two different damage identification schemes have been combined with the impedance method in order to quantitatively assess the state of structures. One is based on a wave propagation modeling, and the other is the use of artificial neural networks. A newly developed wave propagation model has been developed and combined with the impedance method in order to estimate the severity of damage. Numerical and experimental investigations on 1-dimensional structures were presented to illustrate the effectiveness of the combined approach. Furthermore, to avoid the complexity introduced by conventional computational methods in high frequency ranges, multiple sets of artificial neural networks were integrated with the impedance-based health monitoring technique. By incorporating neural network features, the technique is able to detect damage in its early stage and to determine the severity of damage without prior knowledge of the model of structures. The dissertation concludes with experimental examples, investigations on a quarter-scale steel bridge section and a space truss structure, in order to verify the performance of the proposed methodology. / Ph. D.

The focus of this dissertation is aimed at advancing autonomous structural health monitoring. All the research is based on developing the impedance method for monitoring structural health. The impedance technique utilizes piezoelectric patches to interrogate structures of interested with high frequency excitations. These patches are bonded directly to the structure, so information about the health of the structure can be seen in the electrical impedance of the piezoelectric patch. However, traditional impedance techniques require the use of a bulky and expensive impedance analyzer. Research presented here describes efforts to miniaturize the hardware necessary for damage detection. A prototype impedance-based structural health monitoring system, incorporating wireless based communications, is fabricated and validated with experimental testing. The first steps towards a completely autonomous structural health monitoring sensor are also presented. Power harvesting from ambient energy allows a prototype to be operable from a rechargeable power source. Aerospace vehicles are equipped with thermal protection systems to isolate internal components from harsh reentry conditions. While the thermal protection systems are critical to the safety of the vehicle, finding damage in these structures presents a unique challenge. Impedance techniques will be used to detect the standard damage mechanism for one type of thermal protection system. The sensitivity of the impedance method at elevated temperatures is also investigated. Sensors are often affixed to structures as a means of identifying structural defects. However, these sensors are susceptible to damage themselves. Sensor diagnostics is a field of study directed at identifying faulty sensors. The influence of temperature on these techniques is largely unstudied. In this dissertation, a model is generated to identify damaged sensors at any temperature. A sensor diagnostics method is also adapted for use in developed hardware. The prototype used is completely digital, so standard sensor diagnostics techniques are inapplicable. A new method is developed to work with the digital hardware. / Ph. D.

A promising structural health monitoring (SHM) method for implementation on real world structures is impedance-based health monitoring. An in-service system is envisioned to include on board processing and perhaps wireless transfer of data. Ideally, a system could be produced as a slap-on or automatically installed addition to a structure. The research presented in this dissertation addresses issues that will help make such a system a reality. Although impedance-based SHM does not typically use an analytical model for basic damage identification, a model is necessary for more advanced features of SHM, such as damage prognosis, and to evaluate system parameters when installing on various structures. A model was developed based on circuit analysis of the previously proposed low-cost circuit for impedance-based SHM in combination with spectral elements. When a three-layer spectral element representing a piezoceramic bonded to a base beam is used, the model can predict the large peaks in the impedance response due to resonances of the bonded active sensor. Parallel and series connections of distributed sensor systems are investigated both experimentally and with the developed model. Additionally, the distribution of baseline damage metrics is determined to assess how the large quantities of data produced by a monitoring system can be handled statistically. A modification of the RMSD damage metric has also been proposed that is essentially the squared sum of the Z-statistic for each frequency point. Preferred excitation frequencies for macro-fiber composite (MFC) active sensors are statistically determined for a long composite boom under development for use in rigidizable inflatable space structures. / Ph. D.

The understanding of how structural components endure loads, in particular variable loads, is that these components gradually, over some period of time depending on the nature of the loading and the material, develop a microcrack. After some additional time and loading, the microcrack grows to a size that might be detected. Beyond that point, the microcrack propagates in a manner that can be reliably predicted by computer analysis codes. Consequently, one can define different stages for the life of a structural component. These are: 1) the period prior to the formation of a microcrack, 2) the period of microcrack growth, and finally 3) the period of crack growth. To date, structural health monitoring approaches that seek to detect cracks offer no insight into the extent of deterioration occurring in the initial stage that is a precursor to the formation of the microcrack or its growth. However, an approach that would facilitate monitoring the extent of the deterioration that takes place during this stage promises to improve life prediction capabilities of structural components. The challenge, thus, is to develop quantitative assessment of damage accumulation from the earliest stages of the fatigue process and to provide a structure's signature that is dependent of the damage stage. One such signature is the structure's response to forced excitation. The realization of such a goal would help in advancing structural health monitoring procedures using interrogative system identification techniques and determine sensitivities of physical parameters to damage. Additionally, vibration-based spectral quantities are related to physical properties of the structure under test. In this thesis, nonlinear response to parametric excitation is exploited for nonlinear system identification of metallic and composite beam-mass systems before damage initiation through intermediate states of damage progression to failure. Parametric identification procedure combines linear and higher order spectral analysis of vibration measurements and perturbation techniques for the derivation of the approximate solution of the system nonlinear governing differential equation. The possibility of using optical Fiber Bragg Grating sensors technology for damage localization is also assessed. Spectral moments and quantities obtained from fiber optic strain measurements are evaluated near and away from cracks to assess the relation between these moments and cracks. Variations in parameters representing natural frequency, damping and effective nonlinearities for different levels of progressive damage in a beam-mass system have been determined. Their percentage variations have been quantified to establish their sensitivities to damage initiation. The results show that damping and effective nonlinearity parameters are more sensitive to damage conditions than the natural frequency of the first mode. Crack localization is assessed by means of optical fiber technology for a composite beam-mass system. The results show that noise levels in fiber optic signals are high in comparison to strain gage signals. Of particular interest, however, is the observation that the nonlinear response is more pronounced near the cracks than away from them. / Ph. D.

Damage detection in mechanical and aerospace structures is critical to maintaining safe and optimal performance. The early detection of damage increases safety and reduces cost of maintenance and repair. Structural Health Monitoring (SHM) integrates sensor networks and structures to autonomously interrogate the structure and detect damage. The development of robust SHM systems is becoming more vital as aerospace structures are becoming more complex. New SHM methods that can determine the health of the structure without using traditional non-destructive evaluation techniques will decrease the cost and time associated with these investigations. The primary SHM method uses the signals recorded on a pristine structure as a reference and compares operational signals to the baseline measurement. One of the current limitations of baseline SHM is that environmental factors, such as temperature and stress, can change the system response so the algorithm indicates damage when there is none. Many structures which can benefit from SHM have multiple components and often have connections and interfaces that also can change under environmental conditions, thus changing the dynamics of the system. This dissertation addresses some of the current limitations of SHM. First the changes that temperature variations and applied stress create on Lamb wave propagation velocity in plates is analytically modeled and validated. Two methods are developed for the analytical derivative of the Lamb wave velocity; the first uses assumes a thermoelastic material while the second expands thermoelastic theory to include thermal expansion and the associated stresses. A model is developed so the baseline measurement can be compensated to eliminate the false positives due to environmental conditions without storage of dispersion curves or baseline signals at each environmental state. Next, a wave based instantaneous baseline method is presented which uses the comparison of simultaneously captured real time signals and can be used to eliminate the influence of environmental effects on damage detection. Finally, wave transmission and conversion across interfaces in prestressed bars is modeled to provide a better understanding of how the coupled axial and flexural dynamics of a non-ideal preloaded interface change with applied load. / Ph. D.

Structural health monitoring (SHM) is an anomaly detection process. Data-driven SHM has gained much attention compared to the model-based strategy, specifically with the current state-of-the-art machine learning routines. Model-based methods require structural information, time-consuming model updating, and may fail with noisy data, a persistent condition in real-time SHM problems. However, there are several hindrances in supervised and unsupervised settings in machine learning-based SHM. This study identifies and addresses such hindrances with the versatility of state-of-the-art deep learning strategies. While managing those complications, we aim at proposing a general, structure-independent (ie requires no prior information) SHM framework. Developing such techniques plays a crucial role in the SHM of smart cities. In the supervised SHM and sensor output validation (SOV) category, data class imbalance results from the lack of data from nuanced structural states. Employing Long Short-Term Memory (LSTM) units, we developed a general technique that manages both SHM and SOV. The developed architecture accepts high-dimensional features, enabling the train of Generative Adversarial Networks for data generation, addressing the complications of data imbalance. GAN-generated SHM data improved accuracy for low-sampled classes from 44.77% to 64.58% and from 73.39% to 90.84% in two SOV and SHM case studies, respectively. Arguing the unsupervised SHM as a practical category since it identifies novelties (ie unseen states), the current application of dimensionality reduction (DR) in unsupervised SHM is investigated. Due to the curse of dimensionality, classical unsupervised techniques cannot function with high-dimensional features, driving the use of DR techniques. Investigations highlighted the importance of avoiding DR in unsupervised SHM, as data dimensions that DR suppresses may contain damage-sensitive features for novelties. With DR, novelty detection accuracy declined up to 60% in two benchmark SHM datasets. Other obstacles in the unsupervised SHM area are case-dependent features, lack of dynamic-class novelty detection, and the impact of user-defined detection parameters on novelty detection accuracy. We chose the fast Fourier transform-based (FFT) of raw signals with no dimensionality reduction to develop the SHM framework. A deep neural network scheme is developed to perform the pattern recognition of that high-dimensional data. The framework does not require prior information, with GAN models implemented, offering robustness to sensor placement in structures. These characteristics make the framework suitable for developing general unsupervised SHM techniques. / Doctor of Philosophy / Detecting abnormal behaviors in structures from the input signals of sensors is called Structural health monitoring (SHM). The vibrational characteristics of signals or direct pattern recognition techniques can be applied to detect anomalies in a data-driven scheme. Machine learning (ML) tools are suitable for data-driven methods; However, challenges exist on both supervised and unsupervised ML-based SHM. Recent improvements in deep learning are employed in this study to address such obstacles after their identification. In supervised learning, the data points for the normal state of structures are abundant, and datasets are usually imbalanced, which is the same issue for the sensor output validation (SOV). SOV must be present before SHM takes place to remove anomalous sensor outputs. First, a unified decision-making system for SHM and SOV problems is proposed, and then data imbalance is alleviated by generating new data objects from low-sampled classes. The proposed unified method is based on the recurrent neural networks, and the generation mechanism is Generative Adversarial Network (GAN). Results indicate improvements in accuracy metrics for data classes in the minority. For the unsupervised SHM, four major issues are identified, including data loss during feature extraction, case-dependency of such extraction schemes. These two issues are solved with the fast Fourier transform (FFT) of signals to be the features with no reduction in their dimensionality. The other obstacles are the lack of discrimination between different novel classes (ie only two classes of damage and undamaged) and the effect of the detection parameters, defined by users, on the SHM analysis. The latter two predicaments are also addressed by online generating new data objects from the incoming signal stream with GAN and tuning the detection system to have the same performance regarding user-defined parameters regarding GAN-generated data. The proposed unsupervised technique is further improved to be insensitive to the sensor placement on structures by employing recurrent neural network layers in the GAN architecture, with the GAN that has overfitted discriminator.

Maintenance of the structural integrity and damage detection are critical for all massive and complicated new and aging structures. A structural health monitoring (SHM) system intends to identify damage on the structure under monitoring, so that necessary action can be taken in advance to avoid catastrophic results. Impedance-based SHM utilizes a piezoelectric ceramic as a collocated actuator and sensor, which measures the electrical impedance of the piezoelectric ceramic over a certain frequency range. The impedance profile of a structure under monitoring is compared against a reference profile obtained from the healthy structure. An existing approach called the sinc method adopts a sinc wave excitation and performs traditional discrete Fourier transform (DFT) based structural condition assessment. The sinc method requires rather intensive computing and a digital-to-analog converter (DAC) to generate a sinc excitation signal. It also needs an analog-to-digital converter (ADC) to measure the response voltage, from which impedance profile is obtained through a DFT. This dissertation investigates system design approaches for impedance-based structural health monitoring (SHM), in which a primary goal is low power dissipation. First, we investigated behaviors of piezoelectric ceramics and proposed an electrical model in order to enable us to conduct system level analysis and evaluation of an SHM system. Unloaded and loaded piezoelectric ceramics were electrically modeled with lumped linear circuit components, which allowed us to perform system level simulations for various environmental conditions. Next, we explored a signaling method called the wideband method, which uses a pseudorandom noise (PN) sequence for excitation of the structure rather than a signal with a particular waveform. The wideband method simplifies generation of the excitation signal and eliminates a digital-to-analog converter (DAC). The system form factor and power dissipation is decreased compared to the previously existing system based on a sinc signal. A prototype system was implemented on a digital signal processor (DSP) board to validate its approach. Third, we studied another low-power design approach which employs binary signals for structural excitation and structural response measurement was proposed. The binary method measures only the polarity of a response signal to acquire the admittance phase, and compares the measured phase against that of a healthy structure. The binary method eliminates the need for a DAC and an ADC. Two prototypes were developed: one with a DSP board and the other with a microcontroller board. Both prototypes demonstrated reduction of power dissipation compared with those for the sinc method and for the wideband method. The microcontroller based prototype achieved an on-board SHM system. Finally, we proposed an analytical method to assess the quality of the damage detection for the binary method. Using our method, one can obtain the confidence level of a damage detection for a given damage distance. / Ph. D.

Structural Health Monitoring (SHM) is the science and technology of monitoring and can assess the condition of aerospace, civil, and mechanical infrastructures using a sensing system integrated into the structure. SHM is capable of detecting, locating, and quantifying various types of damage such as cracks, holes, corrosion, delamination, and loose joints, and can be applied to various kinds of infrastructures such as buildings, railroads, windmills, bridges, and aircraft. A major technical challenge for existing SHM systems is high power consumption, which severely limits the range of its applications. In this thesis, we investigated adoption of acoustic emission detection to reduce power dissipation of SHM systems employing the impedance and the Lamb wave methods. An acoustic emission sensor of the proposed system continuously monitors acoustic events, while the SHM system is in sleep mode. The SHM system is evoked to perform the SHM operation only when there is an acoustic event detected by the acoustic emission sensor. The proposed system avoids unnecessary operation of SHM operations, which saves power, and the system is effective for certain applications such as spacecraft and wind turbine blades. We developed prototype systems using a Texas Instruments TMS320F2812 DSP evaluation board for the Lamb wave method and an MSP430 evaluation board for the impedance method. / Master of Science

Although nondestructive evaluation techniques have been implemented in many industry fields and proved to be useful, they are generally expensive, time consuming, and the results may not always be reliable. To overcome these drawbacks, structural health monitoring (SHM) systems has received significant attention in the past two decades. As structural systems are becoming more complicated and new materials are being developed, new methodologies, theories, and approaches in SHM have been developed for damage detection, diagnosis, and prognosis. Among the methods developed, the guided Lamb wave based SHM can be a promising technique for damage evaluation since it provides reliable damage information through signals propagating over large distance with little loss of amplitude. While this method is effective for damage assessment, the guided Lamb wave contains complicated mode characteristics, i.e. an infinite number of wave modes exist and these modes are generally dispersive. For this reason, a minimum number of wave modes and various signal processing algorithms are implemented to obtain better signal interpretations. Phased array beamsteering is an effective means for damage detection in guided Lamb wave SHM systems. Using this method, the wave energy can be focused at localized directions or areas by controlled excitation time delay of each array element. In this research, two types of transducers are utilized as phased array elements to compare beamsteering characteristics. Monolithic piezoceramic (PZT) transducers are investigated for beamsteering by assuming omnidirectional point sources for each actuator. MacroFiber Composite (MFC) transducers with anisotropic actuation are also studied, considering the wave main lobe width, main lobe magnitude, and side lobe levels. Analysis results demonstrate that the MFC phased arrays perform better than the PZT phased arrays for a range of beamsteering angles and have reduced main lobe width and side lobe levels. Experiments using the PZT and MFC phased arrays on an aluminum plate are also performed and compared to the analysis results. A time-frequency signal processing algorithm coupled with a machine learning method can form a robust damage diagnostic system. Four types of such algorithms, i.e. short time Fourier transform, Wigner-Ville distribution, wavelet transform, and matching pursuit, are investigated to select an appropriate algorithm for damage classification, and a spectrogram based on short time Fourier transform is adopted for its suitability. A machine learning algorithm called Adaboost is chosen due to its effectiveness and high accuracy performance. The classification is preformed using spectrograms and Adaboost for crack and corrosion damages. Artificial cracks and corrosions are created in Abaqus® to obtain the training samples consist of spectrograms. Several beam experiments in laboratory and additional simulations are also performed to get the testing samples for Adaboost. The analysis results show that not only correct damage classification is possible, but the confidence levels of each sample are acquired. / Ph. D.

This work investigates the feasibility of developing a self-powered structural health monitoring (SHM) smart tire using piezoelectric materials. While this work is divided into two components: SHM and energy harvesting, the context of smart tire in this work is defined as the development of a SHM system that (i) has self-powering capabilities, and (ii) addresses the potential of embedding sensors. The use of impedance based SHM on a tire is severely limited due to the low stiffness and high damping characteristics of the tire. This work propose the use of a high voltage impedance analyzer, and the addition of electrical circuit to enhance the damage detection process. Experimental work was conducted on an aluminum beam and on a tire section with commercially available piezoelectric sensors. The use of a high voltage impedance analyzer was demonstrated to provide insight on damage type and damage location. Two sensors were connected in parallel as an effective sensory system, and was shown to reduce interrogation time, but reduce damage identification sensitivity. With added electrical circuits, a belt separation on the tire was successfully detected by the shift in electrical impedance signature. For the energy harvesting portion of this work, a bimorph piezoelectric energy harvester model was derived using extended Hamilton's principle and the linear constitutive relations of piezoelectric materials. Comparison of model with experimental data at increasing loading conditions demonstrated the monotonic increase in voltage output, with linear asymptotes at extreme loading conditions (short-circuit and open-circuit). It also demonstrated the existence of an optimal resistive load for maximum power output. To address the ability to embed sensors, an existing fabrication process to grow arrays of ZnO nanowires in carbon fiber reinforced polymer was used in this work. Comparison of power generation from a composite beam with ZnO nanowires with a composite beam without ZnO nanowires demonstrated the power generation capabilities of the nanowires. A maximum peak voltage of 8.91 mV and peak power of 33.3 pW was obtained. After the application of 10V DC, a maximum of 45 pW was obtained. However, subsequent application of 20V DC reduced the maximum peak power output to 2.5 pW. Several attempts to increase power generation including adding a tip mass and changing the geometry of the composite beam were conducted. Finally, the theoretical voltage frequency response function obtained from the theoretical piezoelectric constant and dielectric constant of a single ZnO nanowire were compared to the experimental voltage frequency response function. The discrepancies were discussed. / Master of Science