Helicopter rotor system operates in a highly dynamic and unsteady aerodynamic environment leading to severe vibratory loads on the rotor system. Repeated exposure to these severe loading conditions can induce damage in the composite rotor blade which may lead to a catastrophic failure. Therefore, an interest in the structural health monitoring (SHM) of the composite rotor blades has grown markedly in recent years. Two important issues are addressed in this thesis; (1) structural modeling and aeroelastic analysis of the damaged rotor blade and (2) development of a model based rotor health monitoring system. The effect of matrix cracking, the first failure mode in composites, is studied in detail for a circular section beam, box-beam and two-cell airfoil section beam. Later, the effects of further progressive damages such as debonding/delamination and fiber breakage are considered for a two-cell airfoil section beam representing a stiff-inplane helicopter rotor blade. It is found that the stiffness decreases rapidly in the initial phase of matrix cracking but becomes almost constant later as matrix crack saturation is reached. Due to matrix cracking, the bending and torsion stiffness losses at the point of matrix crack saturation are about 6-12 percent and about 25-30 percent, respectively. Due to debonding/delamination, the bending and torsion stiffness losses are about 6-8 percent and about 40-45 percent after matrix crack saturation, respectively. The stiffness loss due to fiber breakage is very rapid and leads to the final failure of the blade. An aeroelastic analysis is performed for the damaged composite rotor in forward flight and the numerically simulated results are used to develop an online health monitoring system. For fault detection, the variations in rotating frequencies, tip bending and torsion response, blade root loads and strains along the blade due to damage are investigated. It is found that peak-to-peak values of blade response and loads provide a good global damage indicator and result in considerable data reduction. Also, the shear strain is a useful indicator to predict local damage. The structural health monitoring system is developed using the physics based models to detect and locate damage from simulated noisy rotor system data. A genetic fuzzy system (GFS) developed for solving the inverse problem of detecting damage from noise contaminated measurements by hybridizing the best features of fuzzy logic and genetic algorithms. Using the changes in structural measurements between the damaged and undamaged blade, a fuzzy system is generated and the rule-base and membership functions optimized by genetic algorithm. The GFS is demonstrated using frequency and mode shape based measurements for various beam type structures such as uniform cantilever beam, tapered beam and non-rotating helicopter blade. The GFS is further demonstrated for predicting the internal state of the composite structures using an example of a composite hollow circular beam with matrix cracking damage mode. Finally, the GFS is applied for online SHM of a rotor in forward flight. It is found that the GFS shows excellent robustness with noisy data, missing measurements and degrades gradually in the presence of faulty sensors/measurements. Furthermore, the GFS can be developed in an automated manner resulting in an optimal solution to the inverse problem of SHM. Finally, the stiffness degradation of the composite rotor blade is correlated to the life consumption of the rotor blade and issues related to damage prognosis are addressed.

In this thesis, we explore the four aspects of fiber Bragg grating sensors: mathematical modeling of Fiber Bragg Grating response/spectral characteristics, fabrication using phase mask, application and interrogation. Applications of fiber Bragg gratings, also known as in-fiber gratings, with emphasis on their sensing capabilities, interrogation of an array of sensors and their performance in structural health monitoring scenario are documented. First, we study the process of photosensitivity phenomenon in glasses, in particular GeO2:SiO2 glasses. For mathematical modeling we consider the 1-D refractive index profile along the propagation axis of an optical fiber drawn from the preform of such glasses. These 1-D index structures exhibit a bandgap for propagation along the fiber axis. We show how the bandgap is dependent on the two structural parameters: index periodicity and effective refractive index. The mathematical model provides the characteristics of three sensor parameters -resonance wavelength also known as the Bragg wavelength (λB ), filter bandwidth (ΔλB ), and reflectivity (R). We show that the evolution of the index structure in germanosilicate glasses is dependent on the inscription parameters such as exposure time, intensity of the laser used for inscribing, the interference pattern, and coherence of the laser system. In particular, a phase mask is used as the diffffacting element to generate the required interference pattern, that is exposed on the photosensitive fiber. We present a mathematical model of the electromagnetic diffraction pattern behind the phase mask and study the effect of the limited coherence of the writing laser on the interference pattern produced by the diffracting beams from the mask. Next, we demostrate the sensing capabilities of the fiber Bragg gratings for measuring strain, temperature and magnetic fields. We report linearity of 99.7% and sensitivity of 10.35pm/◦C for the grating temperature sensor. An array of gratings assigned with non-overlapping spectral windows is inscribed in a single fiber and applied for distributed sensing of structural health monitoring of an aircraft’s composite air-brake panel. The performance of these sensors is compared with the industry standard resistance foil gauges. We report good agreement between the two gauges (FBG and RSG). In some applications it is more desirable to know the spectral content, rather than the magnitude of perturbation. Fiber Bragg gratings sensors can be used to track events that occur in a very small span of time and contain high frequencies. Such applications demand very high speed wavelength demodulation methods. We present two interrogation techniques: wavelength-shift time-stamping (WSTS) and reflectivity division multiplexing (RDM). WSTS interrogation method employs the multiple threshold-crossing technique to quantize the sensor grating fluctuations and in the process produces the time stamps at every level-cross. The time-stamps are assembled and with the a priori knowledge of the threshold levels, the strain signal is reconstructed. The RDM methodology is an extension of the WSTS model to address multiple sensors. We show that by assigning unique reflectivities to each of the sensors in an array, the time-stamps from each of the sensors can be tagged. The time-stamps are collected by virtue of their corresponding pulse heights, and assembled to reconstruct the strain signal of each of the array sensor. We demonstrate that the two interrogation techniques are self-referencing systems, i.e., the speed at which the signals are reconstructed is instantaneous or as fast as the signal itself.

Le thème principal de cette thèse est d'étudier la détection d'anomalies dans des flux de données de grande dimension avec une application spécifique au \emph{Health Monitoring} des moteurs d'avion. Dans ce travail, on considère que le problème de la détection d'anomalies est un problème d'apprentissage non supervisée. Les données modernes, notamment celles issues de la surveillance des systèmes industriels sont souvent des flux d'observations de grande dimension, puisque plusieurs mesures sont prises à de hautes fréquences et à un horizon de temps qui peut être infini. De plus, les données peuvent contenir des anomalies (pannes) du système surveillé. La plupart des algorithmes existants ne peuvent pas traiter des données qui ont ces caractéristiques. Nous introduisons d'abord un algorithme de clustering probabiliste offline dans des sous-espaces pour des données de grande dimension qui repose sur l'algorithme d'espérance-maximisation (EM) et qui est, en plus, robuste aux anomalies grâce à la technique du trimming. Ensuite, nous nous intéressons à la question du clustering probabiliste online de flux de données de grande dimension en développant l'inférence online du modèle de mélange d'analyse en composantes principales probabiliste. Pour les deux méthodes proposées, nous montrons leur efficacité sur des données simulées et réelles, issues par exemple des moteurs d'avion. Enfin, nous développons une application intégrée pour le Health Monitoring des moteurs d'avion dans le but de détecter des anomalies de façon dynamique. Le système proposé introduit des techniques originales de détection et de visualisation d'anomalies reposant sur les cartes auto-organisatrices. Des résultats de détection sont présentés et la question de l'identification des anomalies est aussi discutée.

Structural health monitoring (SHM) is concerned with the continuous, long-term assessment of structural integrity. One commonly investigated SHM technique uses guided ultrasonic waves, which travel through the structure and interact with damage. Measured signals are then analyzed in software for detection, estimation, and characterization of damage. One common configuration for such a system uses a spatially-distributed array of fixed piezoelectric transducers, which is inexpensive and can cover large areas. Typically, one or more sets of prerecorded baseline signals are measured when the structure is in a known state, with imaging methods operating on differences between follow-up measurements and these baselines. Presented here is a new class of SHM spatially-distributed array algorithms that rely on sparse reconstruction. For this problem, damage over a region of interest (ROI) is considered to be sparse. Two different techniques are demonstrated here. The first, which relies on sparse reconstruction, uses an a priori assumption of scattering behavior to generate a redundant dictionary where each column corresponds to a pixel in the ROI. The second method extends this concept by using multidimensional models for each pixel, with each pixel corresponding to a "block" in the dictionary matrix; this method does not require advance knowledge of scattering behavior. Analysis and experimental results presented demonstrate the validity of the sparsity assumption. Experiments show that images generated with sparse methods are superior to those created with delay-and-sum methods; the techniques here are shown to be tolerant of propagation model mismatch. The block-sparse method described here also allows the extraction of scattering patterns, which can be used for damage characterization.

The reliability of modern gas turbine engines is largely due to careful damage tolerant design a method of structural design based on the assumption that flaws (cracks) exist in any structure and will continue to grow with usage. With proper monitoring, largely in the form of periodic inspections at conservative intervals reliability and safety is maintained. These methods while reliable can lead to the early retirement of some components and unforeseen failure if design assumptions fail to reflect reality. With improvements to sensor and computing technology there is a growing interest in a system that could continuously monitor the health of structural aircraft as well as forecast future damage accumulation in real-time. Through the use of two-dimensional and three-dimensional numerical modeling the initial goals and findings for this continued work include: (a) establishing measurable parameters directly linked to the health of the blade and (b) the feasibility of detecting accumulated damage to the structural material and thermal barrier coating as well as the onset of damage causing structural failure.

Many of the existing ambient modal identification methods based on vibration data process information centrally to calculate the modal properties. Such methods demand relatively large memory and processing capabilities to interrogate the data. With the recent advances in wireless sensor technology, it is now possible to process information on the sensor itself. The decentralized information so obtained from individual sensors can be combined to estimate the global modal information of the structure. The main objective of this thesis is to present a new class of decentralized algorithms that can address the limitations stated above. The completed work in this regard involves casting the identification problem within the framework of underdetermined blind source separation (BSS). Time-frequency transformations of measurements are carried out, resulting in a sparse representation of the signals. Stationary wavelet packet transform (SWPT) is used as the primary means to obtain a sparse representation in the time-frequency domain. Several partial setups are used to obtain the partial modal information, which are then combined to obtain the global structural mode information. Most BSS methods in the context of modal identification assume that the excitation is white and do not contain narrow band excitation frequencies. However, this assumption is not satisfied in many situations (e.g., pedestrian bridges) when the excitation is a superposition of narrow-band harmonic(s) and broad-band disturbance. Under such conditions, traditional BSS methods yield sources (modes) without any indication as to whether the identified source(s) is a system or an excitation harmonic. In this research, a novel under-determined BSS algorithm is developed involving statistical characterization of the sources which are used to delineate the sources corresponding to external disturbances versus intrinsic modes of the system. Moreover, the issue of computational burden involving an over-complete dictionary of sparse bases is alleviated through a new underdetermined BSS method based on a tensor algebra tool called PARAllel FACtor (PARAFAC) decomposition. At the core of this method, the wavelet packet decomposition coefficients are used to form a covariance tensor, followed by PARAFAC tensor decomposition to separate the modal responses. Finally, the proposed methods are validated using measurements obtained from both wired and wireless sensors on laboratory scale and full scale buildings and bridges.

This research study presents a new vibration-based non-destructive global structural damage identification and condition monitoring technique that can be used for detection, localization and quantification of damage. A two-stage damage identification process that combines non-model based and model-based damage identification approaches is proposed to overcome the main difficulties associated with the solution of structural damage identification problems. In the first stage, performance assessment of various response parameters obtained from the time-domain, frequency-domain and spectral-domain analysis is conducted using a non model-based damage detection and localization approach. In addition, vibration response parameters that are sensitive to local and global damage and that possess strong physical relationships with key structural dynamic properties are identified. Moreover, in order to overcome the difficulties associated with damage identification in the presence of structural nonlinearity and response nonstationarity, a wavelet transform based damage-sensitive parameter is presented for detection and localization of damage in the space domain. The level of sensitivity and effectiveness of these parameters for detection and localization of damage are demonstrated using various numerical experimental data determined from one-dimensional and two-dimensional plate-like structures.

La corrosion des armatures de renforcement des structures en béton représente un enjeu socio-économique majeur. Sa détection et le suivi de son évolution constituent un défi pour la recherche appliquée. Les techniques standards non destructives de détection de corrosion mettent en œuvre des procédés indirects tels la mesure d’impédance, de potentiels, ou par ultrasons. Leurs capacités d’auscultation sont limitées dans l’espace (notamment en profondeur), leur coût reste élevé dans un contexte de maintenance périodique et elles conduisent à des paramètres d’interprétation complexe. Des progrès sont nécessaires dans la détection et l’analyse fiable de la progression des processus de corrosion. Dans ce travail, nous présentons une nouvelle méthode pour détecter la corrosion et le suivi de son évolution, basée sur l’observation directe des changements intervenant à l’interface fer-béton par Capteur à Fibre Optique (CFO). L'attaque par corrosion de la surface de l'armature dépend de plusieurs paramètres électrochimiques (température, pH, carbonatation, présence de chlorures, contamination biologique, etc.). Deux comportements mécaniques à l'interface fer-béton sont distingués. Dans le premier cas (carbonatation), le produit d'oxydation du métal reste à l'interface et augmente la pression interne, pouvant conduire à la fissuration de la couche de béton extérieure. Dans le second cas (piqures), les ions métalliques sont évacués hors de la structure avec comme conséquence une réduction de section des barres d'armature (affaiblissement du renforcement). Un CFO innovant est proposé dans le but de localiser et quantifier les deux types de corrosion précités. Le principe consiste à observer l’impact direct de la corrosion sur l’état de déformation d’une fibre optique préalablement précontrainte par construction. Deux procédés métrologiques sont étudiés : Bragg et réflectométrie fréquentielle (Optical Frequency-Domain Reflectometry - OFDR). Des tests de corrosion accélérée montrent la faisabilité du procédé. Une procédure de fabrication simplifiée et à coût optimisé est proposée pour la surveillance in situ et répartie des structures de génie civil, dans une perspective future de maintenance conditionnée. / Corrosion of reinforced bars (rebars) in concrete structures remains a major issue in civil engineering works, being its detection and evolution a challenge for the applied research. Usual non-destructive corrosion detection methods involve impedance, potential or ultra-sonic indirect measurements of complex interpretation. Besides, they are restricted to near-surface examinations and the maintenance cost is still high (scheduled maintenance). Many efforts remain to be done to survey the onset and progression of corrosion processes in a reliable way. In this work, we present a new methodology to detect the onset of corrosion and to monitor its evolution, based on the direct observation of rebar–concrete interface changes by the use of an Optical Fiber Sensor (OFS). The corrosion attack over rebar surface depends on several physical, chemical and electrochemical parameters (temperature, pH, presence of chlorides/CO2, biological contamination, etc.). Two types of mechanical behavior and described. In the first case (carbonation), metal oxidation products stay at the interface and increase internal pressure, potentially leading to a crack of the external concrete layer. In the second case (pitting), metal ions are evacuated out of the structure, leading to a reduction of the rebar section (structural weakness). An innovative sensor design is proposed with the purpose of localizing and quantifying the amount of both corrosion types. The basic principle consists in measuring the impact of corrosion over the state of strain of a prestressed optical fiber. Two metrological techniques are used: Fiber Bragg Grating (FBG) and Optical Frequency-Domain Reflectometry (OFDR). Accelerated corrosion tests were performed in electrolytic solutions for both kinds of corrosion types (pitting and carbonation) and provide a proof-of-concept for the technique. A low-cost, simplified manufacturing procedure is proposed with the aim to provide distributed and in situ Structural Health Monitoring (SHM), suitable for future Condition-Based Maintenance (CBM) of civil engineering concrete structures.

L'identification passive d'un système réside dans l'estimation des paramètres qui décrivent ce système uniquement à l'aide de sollicitations ambiantes. Dans le génie civil, cette discipline est appliquée pour le suivi de l'état de santé des structures, on parle de SHM (Structural Health Monitoring) passif. Le SHM passif est généralement réalisé à l'aide d'une instrumentation déployée en surface. La thèse s'est intéressée aux possibilités offertes par une instrumentation qui serait enfouie. Dans une première partie, on établit les résultats associés à l'identification passive des milieux visco-élastiques. L'originalité de ces travaux réside dans la prise en compte d'un modèle de dissipation réaliste, la viscosité, ainsi que du caractère vectoriel des ondes élastiques. Ces résultats théoriques sont validés expérimentalement et démontrent la portabilité du SHM passif en surface au SHM passif en volume. Dans une deuxième partie, on s'intéresse à deux problèmes attachés à l'enfouissement de capteurs: l'estimation passive de leur position (problème SNL, Sensor Network Location problem) et de leur attitude (problème SNA, Sensor Network Attitude problem). Ces problèmes sont résolus grâce à l'identification passive qui fournit, en plus d'information physique sur le milieu, des informations géométriques sur le réseau. En particulier, on peut estimer des distances et des attitudes relatives entre capteurs. A l'aide de ces informations partielles et bruitées, des algorithmes de résolution des problèmes SNL et SNA ont été développés puis validés expérimentalement. Enfin, on synthétise l'apport de la thèse et on identifie les verrous technologiques à lever afin de justifier la faisabilité de l'enfouissement d'un réseau de capteurs dans le but de faire du SHM passif. / Passive identification of a system relies on the estimation of the parameters which describe that system only by using ambient sources. In civil engineering, we can apply this technique to monitor the state of health of structures. This is called passive SHM (Structural Health Monitoring). Passive SHM is generally realised by using an instrumentation distributed on the surface. This thesis focuses on the possibility given by the use of an embedded instrumentation.In the first part, we establish new results associated to passive identification in visco-elastic media. The originality of this work relies on the consideration of a realistic dissipation model, the viscosity, and the vectorial aspect of elastic waves. Those theoretical results, which are experimentally validated, prove the portability of surface passive SHM to volume passive SHM. In the second part, we focus on two problems related to an embedded sensors network: the passive estimation of sensors position (SNL problem - Sensor Network Location problem) and attitude (SNA problem - Sensor Network Attitude problem). Those problems are solved by using passive identification which gives, besides physical information on the medium, geometrical information on the network. In particular, we can estimate the distances and relative attitudes between sensors. With that partial and noisy information we have developed algorithms solving SNL and SNA problems and we have validated them experimentally.At last, we synthesize the contribution of the thesis and we identify the technological locks to release in order to justify the feasibility of passive SHM using an embedded instrumentation.

Pour l’auscultation d’ouvrages, les capteurs à fibre optique sont généralement utilisés puisqu’ils présentent l’avantage de fournir des mesures réparties. Plus particulièrement, le capteur basé sur la technologie Brillouin permet d’acquérir un profil de fréquence Brillouin, sensible à la température et la déformation dans une fibre optique sur une dizaine de kilomètres avec un pas de l’ordre de la dizaine de centimètres. La première problématique est d’obtenir un profil centimétrique sur la même longueur d’auscultation. Nous y répondons en s’appuyant sur des méthodes de séparation de sources, de déconvolution et de résolution de problèmes inverses. Ensuite, nous souhaitons estimer la déformation athermique dans l’ouvrage. Pour cela, plusieurs algorithmes de filtrage adaptatif sont comparés. Finalement, un procédé pour quantifier le déplacement de l’ouvrage à partir des mesures de déformation est proposé. Toutes ces méthodes sont testés sur des données simulées et réelles acquises dans des conditions contrôlées. / For structural health monitoring, optical fiber sensors are mostly used thanks their capacity to provide distributed measurements. Based on the principle of Brillouin scattering, optical fiber sensors measure Brillouin frequency profile, sensitive to strain and temperature into the optical fiber, with a meter spatial resolution over several kilometers. The first problem is to obtain a centimeter spatial resolution with the same sensing length. To solve it, source separation, deconvolution and resolution of inverse problem methodologies are used. Then, the athermal strain into the structure is searched. Several algorithms based on adaptative filter are tested to correct the thermal effect on strain measurements. Finally, several methods are developed to quantify structure displacements from the athermal strain measurements. They have been tested on simulated and controlled-conditions data