One of the most commonly used mechanical systems is the internal combustion engine. Internal combustion engines dominate the automotive industry, and have numerous other applications in generation, transportation, etc. This thesis presents the development of a fault detection and diagnosis (FDD) system for use with an internal combustion engine valve train. A FDD system was developed with a focus on the valve impact amplitudes. Engine cycle averaging and band-pass filtering methods were tuned and utilized for improving the signal to noise ratio. A novel feature extraction method was developed that included a local RMS sliding window method and an adaptive threshold. Faults were seeded in the form of deformed valve springs, as well as abnormal valve clearances. The engine’s manufacturer specifies that a valve spring with 3 mm or more of deformation should be replaced. This thesis investigated the detection of a relatively small 0.5mm spring deformation. Valve clearance values were adjusted 0.1mm above and below the nominal clearance value (0.15mm) to test large clearance faults (0.25mm) and small clearance faults (0.05mm). The performance of the FDD system was tested using an instrumented diesel engine test bed. A comparison of numerous signal processing techniques and classification methods was performed. / Master of Applied Science (MASc)

A number of non-invasive fault detection and diagnosis (FDD) techniques have been researched and have proven to have worked well in classifying faults in internal combustion engines (ICE) and other mechanical and electrical systems. These techniques are an integral step to creating more robust and accurate methods of determining where or how a fault has or will occur in such systems. These FDD techniques have the potential to not only save time avoiding a tear-down of a costly machine, but could potentially add another layer of safety in detecting and diagnosing a fault much earlier than was possible before. Looking at the previous research methods and the systems they used to acquire this data, it is a natural progression to try and make a system which is able to encapsulate all of these ideologies into one inexpensive module capable of integrating itself into the advanced set of FDD. This thesis follows along with the development of a new wireless sensor that is developed specifically for the use in FDD for ICE and other mechanical systems. A new set of software and firmware is created for the system to be able to be incorporated into previously designed algorithms. After creating and manufacturing the sensor it is put to the test by incorporating it into several Artificial Neural Networks (ANN) and comparing the results to previous experiments done with previous research equipment. Using vibration data acquired from a running engine to train a neural network, the wireless sensor was able to perform equally as well as its expensive counter parts. It proved to have the ability to achieve 100% accuracy in classifying specific engine faults. The performance of three ANN training algorithms, Levenberg-Marquardt (LM), extended Kalman Filter (EKF), and Smooth Variable Structure filter (SVSF), were tested and compared. Adding to the feasibility of a standalone system the wireless sensor was tested in a live environment as a method of instant ICE fault detection. / Thesis / Master of Applied Science (MASc)

The diesel engine in heavy-duty trucks is a complex system with many components working together, and a malfunction in any of these components can impact engine performance and result in increased emissions. Fault detection and diagnosis have therefore become essential in modern vehicles, ensuring optimal performance and compliance with progressively stricter legal requirements. One of the most common faults in a diesel engineis faulty injectors, which can lead to fluctuations in the amount of fuel injected. Detecting these issues is crucial, prompting a growing interest in exploring additional signals beyond the currently used signal to enhance the performance and robustness of diagnosing this fault. In this work, an investigation was conducted to identify signals that correlate with faulty injectors causing over- and underfueling. It was found that the NOx, O2, and exhaust pressure signals are sensitive to this fault and could potentially serve as additional diagnostic signals. With these signals, two different diagnostic methods were evaluated to assess their effectiveness in detecting injector faults. The methods evaluated were data-driven residuals and Random Forest classifier. The data-driven residuals, when combined with the CUSUM algorithm, demonstrated promising results in detecting faulty injectors. The O2 signal proved effective in identifying both fault instances, while NOx and exhaust pressure were more effective at detecting overfueling. The Random Forest classifier also showed good performance in detecting both over- and underfueling. However, it was observed that using a classifier requires more extensive data preprocessing. Two preprocessing methods were employed: integrating previous measurements and calculating statistical measures over a defined time span. Both methods showed promising results, with the latter proving to be the better choice. Additionally, the generalization capabilities of these methods across different operating conditions were evaluated. It was demonstrated thatthe data-driven residuals yielded better results compared to the classifier, which requiredtraining on new cases to perform effectively.

Machine health and condition monitoring are billion-dollar concerns for industry. Quality control and continuous improvement are some of the most important factors for manufacturers to consider in order to maintain a successful business. When work floor interruptions occur, engineers frequently employ “Band-Aid” fixes due to resource, timing, or technical constraints without solving for the root cause. Thus, a need for quick, reliable, and accurate fault detection and diagnosis methods are required. Within complex rotating machinery, a fundamental component that accounts for large amounts of downtime and failure involves a very basic yet crucial element, the rolling-element bearing. A worn-out bearing constitutes to some of the most drastic failures in any mechanical system next to electrical failures associated with stator windings. The cyclical motion provides a way for measurements to be taken via vibration sensors and analyzed through signal processing techniques. Methods will be discussed to transform these acquired signals into usable input data for neural network training in order to classify the type of fault that is present within the system. With the wide-spread utilization and adoption of neural networks, we turn our attention to the growing field of sequence-to-sequence deep learning architectures. Language based models have since been adapted to a multitude of tasks outside of text translation and word prediction. We now see powerful Transformers being used to accomplish generative modeling, computer vision, and anomaly detection -- spanning across all industries. This research aims to determine the efficacy of the Transformer neural network for use in the detection and classification of faults within 3-phase induction motors for the automotive industry. We require a quick turnaround, often leading to small datasets in which methods such as data augmentation will be employed to improve the training process of our time-series signals. / Thesis / Master of Applied Science (MASc)

In today's industrial landscape, the pursuit of operational excellence has driven organizations to seek innovative approaches to ensure the uninterrupted functionality of machinery and equipment. Predictive maintenance (PM) provides a pivotal strategy to achieve this goal by detecting faults earlier and predicting maintenance before the system enters a critical state. This thesis proposed a fault detection and diagnosis (FDD) method for predictive maintenance using particle filter resampling and a particle tracking technique. To develop this FDD method, particle filter and hidden Markov model efficiency in the forecasting system state variables are studied on a hydraulic wind power transfer system with different noise levels and system faults. Furthermore, a particle tracker is developed to analyze the particle filter's resampling process and study the particle selection process. After that, the proposed FDD method was developed and validated through three simulation tests employing system degradation models. Furthermore, the system's remaining useful life (RUL) is estimated for those simulation tests.

[pt] Os processos industriais enfrentam novos desafios com o avanço da Indústria 4.0 e a crescente demanda por melhorias na detecção de falhas. A detecção de falha fundamenta-se em diversas técnicas de métodos estatísticos e aprendizado de máquina. Embora sejam eficazes, possuem algumas desvantagens, tais como simplificação do processo, baixa capacidade em lidar com ruído, baixa capacidade em lidar com sistemas complexos não lineares, alta demanda computacional e risco de de overf itting. Em resposta a essas limitações, este trabalho apresenta uma abordagem inovadora na área da polimerização empregando redes neurais siamesas (SNNs) e células long short-term memory (LSTM) para a detecção precoce de falhas na polimerização de estireno. Foi realizado a modelagem da polimerização do estireno em reator CSTR utilizando o método dos momentos para o balanço de massa e energia e, neste sistema, foi adicionado controle proporcional-integral-derivativo (PID) para simular uma situação real de controle de processo no contexto de um processo industrial. A partir do modelo foi possível obter treze simulações, das quais cinco são processos sem falha e oito são processos com falhas. Esses dados foram tratados e serviram para treinar as redes siamesas. Com a capacidade de classificar se esses dados de entrada são semelhantes ou diferentes, foi possível realizar a detecção de falha. Os resultados encontrados demonstram uma taxa de detecção de falhas com uma acurácia de até 100 por cento, demonstrando a capacidade desse modelo em detectar falhas em processos químicos complexos, dinâmicos e não-lineares. Este estudo representa um avanço significativo no campo da detecção de falhas, oferecendo oportunidades valiosas para futuras investigações e aprimoramentos em sistemas inteligentes de detecção de falhas na indústria química. / [en] Industrial processes face new challenges with the advancement of Industry 4.0 and the increasing demand for improvements in fault detection. Fault detection is based on various techniques of statistical methods and machine learning. Although effective, they have some disadvantages, such as process simplification, low capacity to deal with noise, low capacity to deal with complex nonlinear systems, high computational demand, and risk of overfitting. In response to these limitations, this work introduces an innovative approach on the polymerization field that employs siamese neural networks (SNNs) and long short-term memory (LSTM) cells for early detection of faults in styrene polymerization. The modeling of styrene polymerization in a CSTR reactor was carried out using the method of moments for mass and energy balance, and in this system, proportional-integral-derivative (PID) control was added to simulate a real process control situation in the context of an industrial process. From the model, it was possible to obtain thirteen simulations, of which five are non-fault processes and eight are processes with faults. These data were processed and used to train the siamese networks. With the ability to classify whether these input data are similar or dissimilar, it was possible to perform fault detection. The results found demonstrate a fault detection rate with an accuracy of up to 100 percent, demonstrating the capability of this model in detecting faults in complex, dynamic, and nonlinear chemical processes. This study represents a substantial advance in the field of fault detection and also offers valuable findings for future investigations and improvements in intelligent fault detection systems in the chemical industry.

The major contribution of this dissertation is the development of a new integrated measurement and control system for intermixed biogas and photovoltaic systems to achieve safe and optimal energy usage. Literature and field studies show that existing control methods fall short of comprehensive system optimization and fault diagnosis, hence the need to re-look these control methods. The control strategy developed in this dissertation is a considerable enhancement on existing strategies as it incorporates intelligent fuzzy logic algorithms based on C source codes developed on the MPLABX programming environment. Measurements centered on the PIC18F4550 microcontroller were carried out on existing biogas and photovoltaic installations. The designed system was able to accurately predict digester stability, quantify biogas output and carry out biogas fault detection and control. Optimized battery charging and photovoltaic fault detection and control was also successfully implemented. The system optimizes the operation and performance of biogas and photovoltaic energy generation. / Electrical Engineering / M. Tech. (Electrical Engineering)

Ce document synthétise mes travaux de thèse de doctorat en Automatique Productiqueà Grenoble INP (Institut National Polytechnique), thèse préparée au sein dudépartement automatique du laboratoire GIPSA-lab (Grenoble Image Parole Signal etAutomatique). Ce travail s’inscrit dans le cadre du contrôle local et de la supervisiondes systèmes de trafic routier. Les principales contributions portent sur la modélisation,la supervision et la commande locale des systèmes de trafic routier.La contribution apportée à la modélisation du trafic est l’ajout d’un modèle d’incertitudesur le modèle CTM (Cell Transmission Model [Daganzo, 1994]). Ce nouveaumodèle permet de prendre en compte les incertitudes sur différents paramètres dumodèle pour in-fine proposer de nouvelles stratégies de commandes commutées robustes.Outre cette approche de modélisation, nous proposons un niveau de supervisionpermettant d’une part d’estimer en temps réel le mode de fonctionnement et d’autrepart de détecter, localiser et estimer certaines fautes sur le système. L’estimation dynamiquede mode de fonctionnement nous permet de connaître l’état de congestion (ou denon-congestion) de l’aménagement routier considéré. Nous sommes en mesure de détecterdes fautes telles que des chutes de vitesse ou des chutes de capacité survenant sur la route.Enfin, nous proposons deux lois de commandes locales basées sur la théorie dessystèmes à commutations. Ainsi, le schéma de contrôle s’adaptera dynamiquementaux changements de propriétés du système. Ces lois de commande ont pour objet des’insérer dans un schéma de régulation hiérarchique. / This document synthesizes my Phd thesis work in Automatic Control in Grenoble-INP. This thesis has been prepared in the automatic control department of thelaboratory GIPSA-lab. This work is situated in the area of traffic systems control andsupervision. Our contributions are about modeling, supervision and local traffic control.The CTM traffic model has been extended with a model of uncertainties. Thisnews model allows us to take into account the uncertain parameters of the model, topropose new robust switched control law.In addition to this modeling approach, we propose some developments on supervisionof trafic systems. On one hand, we can estimate the operating mode of thesystem in real time and on the other hand to estimate some faults on the system. Thedynamical estimation of the operating mode allows us to know the state of congestion(or non congestion) of the road. We are able to estimate faults such as speed fall andcapacities drop that may appear.Finally, we propose two control laws based on switching systems control. The developedcontrollers adapt their geometry to the properties of the system. The purposeof these controllers is to be inserted in a hierarchic control scheme.

[pt] Calibrações manuais periódicas asseguram o funcionamento correto de um instrumento durante certo período de tempo, mas não garantem que sensores defeituosos permaneçam calibrados por outros períodos, além de eventualmente serem desnecessárias. Em plantas industriais, a análise dos sinais fornecidos pelos sensores que monitoram os processos de produção é uma tarefa difícil em virtude da grande dimensão dos dados. Portanto, uma estratégia de monitoramento e correção online para múltiplos sensores se faz necessária. Este trabalho propõe a utilização de dois modelos: um que emprega Redes Neurais Auto-Associativas com Treinamento Robusto Modificado (RNAAM), e outro que emprega o Teste Seqüencial da Razão de Probabilidade (SPRT) e Lógica Fuzzy. Esses modelos são aplicados em um sistema de monitoramento para auto-correção online e auto-validação das medições realizadas por um grande número de sensores. Diferentemente dos modelos existentes, o sistema proposto faz uso de apenas uma RNAAM para reconstruir os sinais dos sensores com falha. Além disso, os estados do sensor e do modelo de auto-correção são validados por meio dos módulos de confiabilidade e de classificação, respectivamente. Para avaliação do modelo, faz-se uso de uma base de dados com medidas de sensores industriais que controlam e realizam o monitoramento de um motor de combustão interna, instalado em um caminhão de mineração. Os resultados obtidos mostram a capacidade do modelo proposto de mapear e corrigir, com um baixo nível de erro, os sinais dos sensores que apresentam falhas, além de fornecer ao especialista uma ferramenta para a realização de cronogramas de revisão de sensores. / [en] Periodical manual calibrations assure the correct operation of an instrument for a certain period of time, but do not guarantee that faulty sensors remain calibrated for other periods, besides being occasionally unnecessary. In industrial plants the analysis of signals from sensors that monitor a plant is a difficult task due to the high-dimensionality of data. Therefore an online strategy for monitoring and correcting multiple sensors is required. This work proposes the use of two models: one that employs Auto Associative Neural Networks with a Modified Robust Training, and another that employs the Sequential Probability Ratio Test (SPRT) and Fuzzy Logic. These models are applied to an online monitoring system for self-correction and selfvalidation of measurements generated by a large number of sensors. Unlike existing models, the proposed system makes use of only one AANN to reconstruct signals from faulty sensors. Moreover, the states of the sensor and of the self correction model are validated through the reliability and classification modules, respectively. The model is evaluated with a database containing measurements of industrial sensors that control and carry out the monitoring of an internal combustion engine installed in a mining truck. Experimental results show the ability of the proposed model to map and correct signals from faulty sensors with a low error, and to provide a tool for sensor review scheduling.

Rolling element bearings are a vital part in many rotating machinery, including vehicles. A defective bearing can be a symptom of other problems in the machinery and is due to a high failure rate. Early detection of bearing defects can therefore help to prevent malfunction which ultimately could lead to a total collapse. The thesis is done in collaboration with Scania that wants a better understanding of how external sensors such as accelerometers, can be used for condition monitoring in their gearboxes. Defective bearings creates vibrations with specific frequencies, known as Bearing Characteristic Frequencies, BCF [23]. A key component in the proposed method is based on identification and extraction of these frequencies from vibration signals from accelerometers mounted near the monitored bearing. Three solutions are proposed for automatic bearing fault detection. Two are based on data-driven classification using a set of machine learning methods called Support Vector Machines and one method using only the computed characteristic frequencies from the considered bearing faults. Two types of features are developed as inputs to the data-driven classifiers. One is based on the extracted amplitudes of the BCF and the other on statistical properties from Intrinsic Mode Functions generated by an improved Empirical Mode Decomposition algorithm. In order to enhance the diagnostic information in the vibration signals two pre-processing steps are proposed. Separation of the bearing signal from masking noise are done with the Cepstral Editing Procedure, which removes discrete frequencies from the raw vibration signal. Enhancement of the bearing signal is achieved by band pass filtering and amplitude demodulation. The frequency band is produced by the band selection algorithms Kurtogram and Autogram. The proposed methods are evaluated on two large public data sets considering bearing fault classification using accelerometer data, and a smaller data set collected from a Scania gearbox. The produced features achieved significant separation on the public and collected data. Manual detection of the induced defect on the outer race on the bearing from the gearbox was achieved. Due to the small amount of training data the automatic solutions were only tested on the public data sets. Isolation performance of correct bearing and fault mode among multiplebearings were investigated. One of the best trade offs achieved was 76.39 % fault detection rate with 8.33 % false alarm rate. Another was 54.86 % fault detection rate with 0 % false alarm rate.