Various improvements which have been made to the vehicle (reduced engine noise, reducedaerodynamic related NVH), have resulted in tire road noise as the dominant source of thevehicle interior noise. Generally, vehicle interior noise has two main sources, 1) travellinglow frequency excitation below 800 Hz from road surface through a structure- borne pathand 2) the high frequency (above 800 Hz) air-borne noise that is caused by air- pumpingnoise caused by tread pattern.The structure-borne waves of the circumference of the tire are generated by excitation atthe contact patch due to the road surface texture and characteristics. These vibrations arethen transferred from the sidewalls of the tire to the rim and then are transmitted throughthe spindle-wheel interface, resulting in high frequency vibration of vehicle body panels andwindows.The focus of this study is to develop several statistical-based models for analyzing the roadsurface and using them to predict the tire-road noise structure-borne component. In order todo this, a new methodology for sensing the road characteristics, such as asperities and roadsurface condition, were developed using virtual sensing and intelligent tire technology. In ad-dition, the spindle forces were used as an indicator to the structure-borne noise of the vehicle.Several data mining and multivariate analysis-based methods were developed to extractfeatures and to develop an empirical model to predict the power of structure-borne noiseunder different operational and road conditions. Finally, multiple data driven models-basedmodels were developed to classify the road types, and conditions and use them for the noisefrequency spectrum prediction. / Doctor of Philosophy / Multiple data driven models were developed in this study to use the vibration of the tirecontact patch as an input to sense some characteristics of road such as asperity, surface type,and the surface condition, and use them to predict the structure-borne noise power. Also,instead of measuring the noise using microphones, forces at wheel spindle were measuredas a metric for the noise power. In other words, a statistical model was developed that bysensing the road, and using the data along with other inputs, one can predict forces at thewheel spindle.

This master thesis focuses on applying deep neural networks (DNNs) in image-based condition monitoring of air-jet spinning machines, specifically focusing on the spinning pressure parameter. The study aims to develop a sensor system to detect structural defects in yarns and assign them to specific machine conditions. The research explores using DNNs to analyze images of yarns generated at different spinning pressures within the spinning box to create a rich dataset for training deep learning models. The study also evaluates the effectiveness of the DNN-based approach in detecting and classifying structural defects in yarns and determining the corresponding machine conditions. The outcomes of this research could potentially help textile enterprises improve the quality and efficiency of their yarn manufacturing processes.

<p>One of the most often utilized machinery in fluid power applications is the material-handling machines, which includes telehandlers, forklifts, cranes, and scissor lifts that are used from constructions to mining.<br> Counterbalance valves (CBVs), hydraulic components that protect the system from failures and manage the load under overrunning load conditions due to their distinctive design, are used in material-handling devices to ensure both the operators' and most off-road vehicles' safety. However, they present a significant shortcoming: the over-pressurization of the supply line, which leads to constringent energy consumption. The primary motivation for this work is this drawback. In this work, a CBV-based system with an adjustable pilot has been investigated using a truck-mounted hydraulic crane as a reference machine.</p> <p>By analyzing theoretically and experimentally the behavior of this novel hydraulic system, it is possible to achieve up to 90% of energy-saving than a baseline configuration of a load-holding machine by controlling the opening of the CBV by adjusting the pressure at the pilot stage. After exploring the capabilities of the studied system and the possible control strategies to control opening of the CBV, this work suggests two different solutions to control the system: “Smart CBV” and “Smart System” modes. By properly controlling the signal on the pilot stage of the CBV, "Smart CBV" enables energy savings of up to 80%. On the other hand, the "Smart System" mode can save up to 95% of energy by using the CBV as a meter-out element that successfully regulates the flow to the actuator and, consequently, its velocity. To attain these outstanding results, it is essential to maintain proper system control.</p> <p>Moreover, since safety is one of the priorities of this type of machine, a Condition Monitoring (CM) model is developed to ensure the actual functionalities of the novel proposed system. By identifying faulty conditions and preventing breakdowns before they occur, CM can be utilized to improve the safety of these type of machines. However, training a CM model using experimental data is time-consuming and expensive since it requires abundant data with different extent of machine failures from the field test. The solution suggested in this work is to generate faulty and healthy data for the reference machine using a high-fidelity simulation tool to train a CM model.</p> <p>Particular focus is given to the counterbalance valve (CBV), a crucial element for the hydraulic system of material handling machines, and the linear actuator (hydraulic cylinder). The different types of faults on two elements are modeled with an approach validated using experimental tests. Considering that the simulation model provides comparable outcomes to training on empirical data, the CM model is trained in a single fault condition and multi faults conditions using simulated data. Instead, the CM model is tested using the experimental tests in multiple faulty conditions on the chosen components.</p> <p>Moreover, finding the best CM model for this case study is another goal of this work. As a result, several CM models are investigated: Random Forest (RF), K-Nearest Neighbor (KNN), and Support Vector Machine (SVM). In terms of precision and recall, metrics frequently employed in the CM field to assess the performances of the designed CM model, the results generally indicate more than 90% accuracy.</p>

This thesis investigates the current state of tool wear prediction models in machining, focusing on their limitations in accurately incorporating the complex dynamics of cutting fluids and their industrial applicability. It proposes a comprehensive evaluation framework to classify and evaluate a wide range of models, including empirical, physical, computational, and data-driven models. The study identifies the key limitations and strengths of each model category. It proposes enhancements by integrating advanced fluid dynamics and predictive modeling techniques to improve tool wear predictions' accuracy and industrial applicability. A structured literature review was conducted to investigate and evaluate existing tool wear models and their integration with cutting fluid dynamics. This review included defining search criteria, selecting relevant studies, and assessing their quality and relevance. The study uses thematic analysis and model evaluation frameworks to classify and evaluate the models, leading to the identification of critical limitations and strengths. The literature review and model evaluation findings revealed that empirical models, while simple and quick to implement, showed moderate accuracy and limited fluid dynamics integration. Physical models provided high accuracy in specific conditions but were computationally intensive. Computational models, particularly those using techniques like Finite Element Analysis (FEA) and Computational Fluid Dynamics(CFD), offered detailed insights and high accuracy but required significant computational resources. Data-driven models demonstrated exceptional predictive capabilities and comprehensive fluid dynamics integration but relied heavily on data availability and quality.  The proposed enhancements include introducing non-linear elements into empirical models, incorporating simplified fluid models or empirical correlations into physical models, exploring reduced-order models (ROMs) or surrogate models for computational models, and developing robust data preprocessing and augmentation techniques for data-driven models. These enhancements aim to improve the accuracy and applicability of tool wear models in industrial machining processes, ultimately contributing to more efficient and cost-effective machining operations. The study emphasizes the importance of a systematic and holistic approach to model evaluation and enhancement. Future research should focus on validating these proposed enhancements through empirical studies and real-world applications, ensuring their relevance and robustness in diverse industrial settings. This research offers significant potential to advance tool wear modeling, providing valuable insights for both academia and industry.

Die vorliegende Arbeit beschreibt die Entwicklung und Charakterisierung eines mikromechanischen kraftgekoppelten Schwingsystems für die resonante Detektion niederfrequenter Schwingungen. Es wird ein neuartiges Prinzip vorgestellt, das es ermöglicht, niederfrequente Vibrationen frequenzselektiv zu erfassen. Mittels Amplitudenmodulation wird das niederfrequente Signal in einen höheren Frequenzbereich umgesetzt. Durch Ausnutzung der mechanischen Resonanzüberhöhung wird aus dem breitbandigen Signal ein schmales Band herausgefiltert, die anderen Frequenzbereiche werden unterdrückt. Auf diese Weise wird direkt die spektrale Information des niederfrequenten Signals gewonnen. Eine Fourier-Transformation ist hierbei nicht notwendig. Die Abstimmung des Sensors erfolgt über eine Wechselspannung und führt dadurch zu einer einfachen Auswertung. Die Schwerpunkte der Arbeit liegen in den theoretischen Untersuchungen zum neuartigen Sensorprinzip, in der Entwicklung einer mikromechanischen Sensorstruktur zum Einsatz des neuen Prinzips sowie in der Entwicklung und Charakterisierung eines Messsystems zur Detektion niederfrequenter mechanischer Schwingungen mit dem neuen Sensor. / This thesis describes the development and characterization of a micromechanical force coupled oscillator system for the resonant detection of low frequency vibrations. It presents a novel working principle that enables spectral measurements of low frequency vibrations. The low frequency spectral content is converted into a higher frequency range by means of amplitude modulation. Due to the mechanical resonance a narrow band is filtered out of the wide band vibration signal. The remaining frequency content is suppressed. Hence, the spectral information is directly obtained with the sensor system without a fast Fourier transform. The tuning is done with an AC voltage resulting in a simple analysis. The main focuses of the work are the theoretical analysis of this novel sensor principle, the development of the micromechanical sensor structure for the use of the novel principle as well as the development and characterization of a measurement system for the spectral detection of low frequency mechanical vibrations with the developed sensor system.

The organic materials used for stator winding insulation are subject to deterioration from thermal, electrical, and mechanical stresses. Stator winding insulation breakdown due to excessive thermal stress is one of the major causes of electric machine failures; therefore, prevention of such a failure is crucial for increasing machine reliability and minimizing financial loss due to motor failure. This work focuses on the development of an efficient and reliable stator winding temperature estimation scheme for small to medium size mains-fed induction machines. The motivation for the stator winding temperature estimation is to develop a sensorless temperature monitoring scheme and provide an accurate temperature estimate that is capable of responding to the changes in the motors cooling capability. A discussion on the two major types of temperature estimation techniques, thermal model-based and parameter-based temperature techniques, reveals that neither method can protect motors without sacrificing the estimation accuracy or motor performance. Based on the evaluation of the advantages and disadvantages of these two types of temperature estimation techniques, a new online stator winding temperature estimation scheme for small to medium size mains-fed induction machines is proposed in this work. The new stator winding temperature estimation scheme is based on a hybrid thermal model. By correlating the rotor temperature with the stator temperature, the hybrid thermal model unifies the thermal model-based and the parameter-based temperature estimation techniques. Experimental results validate the proposed scheme for stator winding temperature monitoring. The entire algorithm is fast, efficient and reliable, making it suitable for implementation in real time stator winding temperature monitoring.

I dagsläget finns det ingen standardiserad metod för att mäta en enhets tillstånd medhjälp av dofter. Vid tillståndsövervakning av rullningslager är vibrationsmätning denmest dominanta metoden. I samband med vibrationsmätning används i vissa falltemperaturövervakning för att få en bättre insikt på rullningslagrets tillstånd. I det härarbetet undersöks de om en elektronisk näsa kan avgöra ett rullningslagers tillstånd.Innan några mätningar påbörjas monterades en elektronisk näsa ihop i ett hölje sombestår av ett kretskort, metalloxid-sensorer och en fläkt för att styra dofter med ettkonstant flöde mot sensorerna. Den elektroniska näsan styrs av en Arduino Nanomikrokontroller. Utöver e-näsan sättes en enhet ihop tillhörande två temperaturgivareoch en luftfuktighetsgivare som styrs av en Arduino UNO. Enhetens syfte är att kunnakontrollera de rådande förhållandena vid mätningar och för att leta någon form avkorrelation mot e-näsan vid eventuella utslag. Förstörande prover av kullager utfördesför att se om e-näsan reagerar innan ett lagerhaveri. Testerna gjordes i en öppen samtsluten miljö och tre stycken olika oljor används för att smörja lagret. Detta för att seom e-näsan reagerar olika beroende på vilken olja som används. En undersökningutförs ifall den elektroniska näsan kan separera på de tre oljorna som används ilagertesterna. För att utvärdera mätresultaten används Excel och Minitab, därprincipalkomponentanalyser genomförs på all mätdata. Efter att alla lagerprover harverkställts utfördes en uppföljning av rullningslagrena för att studera deras tillstånd,detta genom ett optiskt mikroskop.Det framgår i rapporten att med hjälp av analysmetoden PCA syns det att denelektroniska näsan kunde skilja på hydraulolja, motorolja och växellådsolja. Utslag iPCA för de olika mätserierna blev inte identiska men det blev tydligaklusterindelningar hos samtliga mätserier. Genomförd studie visade att med delagerhaveri samt temperaturer går det inte att avgöra ett kullagers tillstånd med hjälpav en elektronisk näsa. Eftersom att de specifika gas-sensorerna som användes till enäsaninte gav någon form av utslag vid mätningarna. Den elektroniska näsanreagerade däremot vid totalhaveri av kullager, vilket är för sent i ett förebyggandeunderhållsperspektiv. Detta medförde att den elektroniska näsan inte kan användas förtillståndsövervakning av det specifika kullagret som användes vid denna studie. / At present, there is no standardized method of measuring a device's condition with thehelp of odors. In condition monitoring of rolling bearings, vibration measurement isthe most dominant method. In case of vibration measurement, temperature monitoringis used in some cases to get a better insight into the condition of the bearing. In thiswork, it is investigated whether an electronic nose can determine the condition of arolling bearing.Before any measurements began, an electronic nose is assembled in a housingconsisting of a circuit board, metal oxide sensors and a fan for stearing odors with aconstant flow towards the sensors. The electronic nose is controlled by an ArduinoNano which is a microcontroller. In addition to the e-nose, a unit is connected to twotemperature sensors and a humidity sensor controlled by an Arduino UNO. The unit'spurpose is to monitor the status and to look for any kind of correlation with the e-nosein case of any possible findings. Destructive specimens of ball bearings are performedto see if the e-nose responds prior to a bearing failure. Tests are conducted in an openand closed environment and three different oils are used to lubricate the bearings.This to see if the e-nose acts differently depending on the oil that is used. Aninvestigation is conducted if the electronic nose can separate the three different typesof oils that is used in the destructive bearing tests. To evaluate the measurementresults, Excel and Minitab are used, where principal component analysis is performedon all measurement data. After all bearing tests have been performed, a follow-up ofthe rolling bearings condition is performed, this through an optical microscope.The report shows that using the PCA analysis method, it appears that the electronicnose could distinguish between hydraulic oil, engine oil and gear oil. In the PCA forthe different measurement series the results did not become identical, but clusterdivisions became clear in all measurement series. Completed study showed that withthese bearing failures and temperatures, it is not possible to determine the condition ofthis ball bearer using an electronic nose. Because the specific gas sensors used for thee-nose did not give any kind of impact during the measurements. On the other hand,the electronic nose responded to a total failure of a ball bearing, which is too late in apreventative maintenance perspective. Therefore, the electronic nose cannot be usedfor condition monitoring of the specific ball bearing used in this study.

L’objectif de ces travaux de thèse est de développer des architectures fiables de surveillance et de détection des défauts d’une machine asynchrone basées sur des techniques paramétriques de traitement du signal. Pour analyser et détecter les défauts, un modèle paramétrique du courant statorique en environnement stationnaire est proposé. Il est supposé être constitué de plusieurs sinusoïdes avec des paramètres inconnus dans le bruit. Les paramètres de ce modèle sont estimés à l’aide des techniques paramétriques telles que les estimateurs spectraux de type sous-espaces (MUSIC et ESPRIT) et l’estimateur du maximum de vraisemblance. Un critère de sévérité des défauts, basé sur l’estimation des amplitudes des composantes fréquentielles du courant statorique, est aussi proposé pour évaluer le niveau de défaillance de la machine. Un nouveau détecteur des défauts est aussi proposé en utilisant la théorie de détection. Il est principalement basé sur le test du rapport de vraisemblance généralisé avec un signal et un bruit à paramètres inconnus. Enfin, les techniques paramétriques proposées ont été évaluées à l’aide de signaux de courant statoriques expérimentaux de machines asynchrones en considérant les défauts de roulements et les ruptures de barres rotoriques. L’analyse des résultats expérimentaux montre clairement l’efficacité et la capacité de détection des techniques paramétriques proposées. / This Ph.D. thesis aims to develop reliable and cost-effective condition monitoring and faults detection architectures for induction machines. These architectures are mainly based on advanced parametric signal processing techniques. To analyze and detect faults, a parametric stator current model under stationary conditions has been considered. It is assumed to be multiple sinusoids with unknown parameters in noise. This model has been estimated using parametric techniques such as subspace spectral estimators and maximum likelihood estimator. A fault severity criterion based on the estimation of the stator current frequency component amplitudes has also been proposed to determine the induction machine failure level. A novel faults detector based on hypothesis testing has been also proposed. This detector is mainly based on the generalized likelihood ratio test detector with unknown signal and noise parameters. The proposed parametric techniques have been evaluated using experimental stator current signals issued from induction machines under two considered faults: bearing and broken rotor bars faults.Experimental results show the effectiveness and the detection ability of the proposed parametric techniques.

Dans le cas des installations photovoltaïques, l’onduleur est le premier élément défaillant dont il est difficile d’anticiper la panne, et peu d’études ont été faites sur la fiabilité de ce type de convertisseur. L'objectif de cette thèse est de proposer des outils et méthodes en vue d'étudier le vieillissement des modules de puissance dans ce type d'application en se focalisant sur les phénomènes de dégradation liés à des aspects thermomécaniques. En règle générale, le vieillissement accéléré des modules de puissance est effectué dans des conditions aggravées de courant (Cyclage Actif) ou de température (Cyclage Passif) pour accélérer les processus de vieillissement. Malheureusement, en appliquant ce type de vieillissement accéléré, des mécanismes de défaillances qui ne se produisent pas dans la vraie application peuvent être observés et, inversement, d'autres mécanismes qui se produisent habituellement peuvent ne pas apparaître. La première partie de la thèse se focalise donc sur la mise en place d'une méthode de vieillissement accéléré des composants semi-conducteurs des onduleurs photovoltaïques. Cela est fait en s’appuyant sur l’analyse des profils de mission du courant efficace de sortie des onduleurs et de la température ambiante, extraits des centrales photovoltaïques situées au sud de la France sur plusieurs années. Ces profils sont utilisés pour étudier les dynamiques du courant photovoltaïque, et sont introduites dans des modèles numériques pour estimer les pertes et les variations de la température de jonction des semi-conducteurs utilisés dans les onduleurs, en utilisant l’algorithme de comptage de cycles "Rainflow". Cette méthode est ensuite mise en œuvre dans deux bancs expérimentaux. Dans le premier, les composants sous test sont des modules IGBT. Les composants sont mis en œuvre dans un banc de cyclage utilisant la méthode d'opposition et mettant en œuvre le profil de vieillissement défini précédemment. Un dispositif in-situ de suivi d'indicateurs de vieillissement (impédance thermique et résistance dynamique) est également proposé et évalué. Le deuxième banc est consacré à l'étude de modules de puissance à base de MOSFET SiC. Le vieillissement est effectué dans les mêmes conditions que pour les modules IGBT et de nombreux indicateurs électriques sont monitorés mais, cette fois ci, en extrayant les composants de l'onduleur de cyclage. Les résultats obtenus ont permis de déterminer des indicateurs de vieillissement d’IGBT et de MOSFET SiC utilisés dans un onduleur photovoltaïque / In the case of photovoltaic installations, the DC/AC inverter has the highest failure rate, and the anticipation of its breakdowns is still difficult, while few studies have been done on the reliability of this type of inverter. The aim of this PhD is to propose tools and methods to study the ageing of power modules in this type of application, by focusing on ageing phenomena related to thermo-mechanical aspects. As a general rule, the accelerated ageing of power modules is carried out under aggravated conditions of current (Active Cycling) or temperature (Passive Cycling) in order to accelerate the ageing process. Unfortunately, when applying this type of accelerated ageing tests, some failure mechanisms that do not occur in the real application could be observed, while inversely, other mechanisms that usually occur could not be recreated. The first part of the PhD focuses on the implementation of an accelerated ageing method of the semiconductor devices inside photovoltaic inverters. This is accomplished by analyzing the mission profiles of the inverter’s output current and ambient temperature, extracted over several years from photovoltaic power plants located in the south of France. These profiles are used to study photovoltaic current dynamics, and are introduced into numerical models to estimate losses and junction temperature variations of semiconductors used in inverters, using the cycle counting algorithm “Rainflow”. This method is then performed in two experimental test benches. In the first one, the devices under test are IGBT modules, where the accelerated ageing profile designed is implemented using the opposition method. Moreover, an in-situ setup for monitoring ageing indicators (thermal impedance and dynamic resistance) is also proposed and evaluated. The second bench is devoted to study the ageing of SiC MOSFET power modules. The accelerated ageing test is carried out under the same conditions as for the IGBT modules with more monitored electrical indicators, but this time by disconnecting the semiconductor devices from the inverter. The results obtained allowed to determine several potential ageing indicators of IGBTs and SiC MOSFETs used in a photovoltaic inverter

The main objective of the presented thesis is to contribute to the development of diagnostic methods for piezoelectric sensor testing. The thesis describes the methods for piezoelectric sensors microcracks identification and diagnostics. The core of the thesis presents the development of a knock sensor prototype, design of suitable methods for the knock sensors diagnosis and evaluation of developed methods with focus on detection of microcracks in the sensor piezoceramic. The last part of the thesis deals with the influence of cracks and splits on the measured data from the piezoelectric transducer. The presented methods are focusing on impedance measurements and sensors frequency response measurements. Known properties of used piezoelectric material as an information source for measurement and diagnosis are provided. The main result of the work is the evaluation of the methods developed for the piezoelectric sensors self-diagnosis.