Unmanned, as well as manned space missions have seen a high failure rate in the early era of space technology. However, this decreased a lot since technology advanced and engineers learnt from previous experiences and improved critical real time systems with fault detection mechanisms. Fault detection, isolation and recovery, nowadays, is generally available in every flying device. However, the cost of hardware can bottleneck the process of creating such a system that is both robust and responsive. This thesis analyses the possibility of implementing a fault detection, isolation and recovery system inside of a single-threaded, cooperative scheduling operating system. The thesis suggests a cooperative implementation of such a system, where every task is responsible for parts of the fault detection. The analysis is done from both the integration layer, across the operating system and its tasks, as well as from the inside of the detection system, where two key components are implemented and analyzed: debug telemetry and operation modes. Results show that it is possible to implement a fault detection system that is spread across all the components of the satellite and acts cooperatively. Furthermore, the comparison with a traditional, dedicated fault detection system proves that errors can be caught faster with a cooperative mechanism. / Obemannade såväl som bemannade rymduppdrag har sett ett högt misslyckande i rymdteknikens tidiga era. Detta har dock förbättrats mycket sedan ingenjörer började lära sig av sina tidigare erfarenheter och utrustade kritiska realtidssystem med feldetekteringsmekanismer. Idag är alla flygande enheter utrustade med feldetekterings-, isolerings- och återställningsmekanismer. Däremot kan kostnaden för hårdvara vara ett problem för processen att skapa ett sådant system som är både robust och mottagligt. Denna uppsats analyserar möjligheten att implementera ett feldetekterings-, isolerings- och återställningssystem inuti ett enkelgängat samarbetsplaneringssystem. Denna uppsats föreslår ett samarbete för implementering av ett sådant system, där varje uppgift ansvarar för delar av feldetekteringen. Analysen görs från både integrationsskiktet, över operativsystemet och dess uppgifter, samt från insidan av detekteringssystemet, där två nyckelkomponenter implementeras och analyseras. Resultaten visar att det är möjligt att implementera ett feldetekteringssystem som täcker alla satellitkomponenter och som är mottaglig. Dessutom visar jämförelsen med ett traditionellt, dedikerat feldetekteringssystem att fel kan fångas snabbare med en mottagligmekanism. / Misiunile spat,iale cu oameni, atât cât s, i fara oameni, au avut o rata a es, ecurilor destul de ridicata în perioada init,iala a erei tehnologiei spat,iale. În schimb, aceasta a scazut semnificativ odata cu dezvoltarea tehnologiei, dar s, i datorita faptului ca inginerii au învat,at din experient,ele precendente s, i au îmbunatat, it sistemele critice în timp real cu mecanisme de detect,ie a erorilor. Sisteme de detect,ie, izolare s, i recuperare din erori sunt disponibile astazi în aproape toate sistemele spat,iale. Însa, costul echipamentelor poate împiedica crearea unor astfel de sisteme de detect,ie, care sa fie robuste s, i responsive. Aceasta teza analizeaza posibilitatea implementarii unui sistem de detect,ie, izolare s, i recuperare de la erori într-un satelit care este echipat cu un procesor cu un singur fir de execut,ie, care are un sistem de planificare cooperativ în sistemul de operare. Aceasta teza sugereaza o implementare cooperativa a unui astfel de sistem, unde fiecare proces este responsabil de câte o parte din detectarea erorilor. Analiza este realizata atât din perspectiva integrarii în sistemul de operare s, i procesele acestuia, cât s, i din interiorul acestui sistem de detect,ie, unde doua elemente importante sunt implementate s, i analizate: telemetria de depanare s, i modurile de operare. Rezultatele arata faptul ca este posibila implementarea unui sistem de detect,ie care este împart, it în toate componentele sistemului unui satelit s, i se comporta cooperativ. Mai departe, comparat,ia cu un sistem tradit,ional, dedicat, de detect,ie a erorilor arata ca erorile pot fi detectate mai rapid cu un sistem cooperativ.

In the aviation industry, safety and robustness are the number one priorities, which is why they use well-tested systems such as hydraulic actuators. However, drawbacks such as high weight and maintenance have pushed the industry toward newer, electrical, actuators that are more efficient and lighter. Electrical actuators, on the other hand, have some reliability issues. In particular, short circuits in the stator windings of Permanent-Magnet SynchronousMotors (PMSMs), referred to as Inter-Turn Short Faults (ITSFs), are the dominating faults, and is the focus of this thesis. ITSFs are usually challenging to detect and often do not become noticeable until the fault has propagated, and the motor is on the verge of being destroyed. This thesis investigates the possibility of detecting ITSFs in a PMSM, at an early stage when only one turn is shorted. The method is limited to finding the faults using ML algorithms. Both an experiential PMSM and a simulated model of the experimental PMSM, with the ability to induce an ITSF, were used to collect the data. Several Machine Learning (ML) models were developed, and then trained and tested with the collected data. The results show that four of the tested ML models, being: Random Forest, Gaussian SVM, KNN, and the CNN, all achieve an accuracy exceeding 95%, and that the fault can be found at an early stage in a PMSM with three coils connected in parallel in each phase. The results also show that the ML models are able to identify the ITSF when the simulated data is downsampled to the same frequency as the experimental data. We conclude that the ML models, provided in this study, can be used to detect an ITSF in a simulated PMSM, at an early stage when only one turn is shorted, and that there is great potential for them to detect ITSFs in a physical motor as well.

Indiana University-Purdue University Indianapolis (IUPUI) / In a wide-area power system, detecting dynamic events is critical to maintaining system stability. Large events, such as the loss of a generator or fault on a transmission line, can compromise the stability of the system by causing the generator rotor angles to diverge and lose synchronism with the rest of the system. If these events can be detected as they happen, controls can be applied to the system to prevent it from losing synchronous stability. In order to detect these events, pattern recognition tools can be applied to system measurements. In this thesis, the pattern recognition tool decision trees (DTs) were used for event detection. A single DT produced rules distinguishing between and the event and no event cases by learning on a training set of simulations of a power system model. The rules were then applied to test cases to determine the accuracy of the event detection. To use a DT to detect events, the variables used to produce the rules must be chosen. These variables can be direct system measurements, such as the phase angle of bus voltages, or indices created by a combination of system measurements. One index used in this thesis was the integral square bus angle (ISBA) index, which provided a measure of the overall activity of the bus angles in the system. Other indices used were the variance and rate of change of the ISBA. Fitting a polynomial curve to a sliding window of these indices and then taking the difference between the polynomial and the actual index was found to produce a new index that was non-zero during the event and zero all other times for most simulations. After the index to detect events was chosen to be the error between the curve and the ISBA indices, a set of power system cases were created to be used as the training data set for the DT. All of these cases contained one event, either a small or large power injection at a load bus in the system model. The DT was then trained to detect the large power injection but not the small one. This was done so that the rules produced would detect large events on the system that could potentially cause the system to lose synchronous stability but ignore small events that have no effect on the overall system. This DT was then combined with a second DT that predicted instability such that the second DT made the decision whether or not to apply controls only for a short time after the end of every event, when controls would be most effective in stabilizing the system.

The integration of control systems with wireless communication networks has gainedsignificant popularity, leading to the emergence of wireless networked control systems(WNCS). WNCS comprises wireless devices such as sensors, actuators, andcontrollers that work together to ensure system stabilizability. However, communicationamong these wireless devices often relies on relay nodes, which presents a challengein guaranteeing system stabilizability due to potential failures caused by natural eventsor malicious activities targeting and disabling these nodes. This study proposes an approach to enhance the resilience of wireless sensor networks(WSNs) utilizing the WirelessHART communication protocol. The objective is to designthe WSNs and controller that can ensure stabilizability, even in the specific number ofrelay node failures. The study employs two key methodologies: firstly, the analysis ofconditions such as controllability, observability, solvability of fault detection andisolation, and the associated requirements to guarantee system stabilizability. Secondly,MATLAB simulations are employed to test the proposed system. By combiningtheoretical analysis and practical simulations, the study provides valuable insights anddesign strategies that contribute to the advancement of WNCS, enabling their reliabilityand stability in the face of potential relay node failures. / Integrationen av styrsystem med trådlösa kommunikationsnätverk har blivit mycketpopulär, vilket har lett till framväxten av trådlösa nätverksstyrningssystem (WNCS).WNCS omfattar trådlösa enheter som sensorer, aktuatorer och styrenheter som arbetartillsammans för att säkerställa systemets stabilisering. Kommunikationen mellan dessatrådlösa enheter förlitar sig emellertid ofta på relänoder, vilket utgör en utmaning föratt garantera systemets stabilisering på grund av potentiella fel orsakade av naturligahändelser eller skadliga aktiviteter som riktar sig mot och inaktiverar dessa noder. Denna studie föreslår ett tillvägagångssätt för att förbättra motståndskraften hos trådlösasensornätverk (WSNs) med hjälp av kommunikationsprotokollet WirelessHART. Måletär att utforma WSN och styrenheter som kan säkerställa stabilisering även vid ettspecifikt antal relänodfel. Studien använder sig av två huvudsakliga metoder: för detförsta analys av villkor såsom styrbarket, observerbarheten, lösbarheten hos felupptäcktoch isolering metod och de tillhörande kraven för att garantera systemets stabilisering.För det andra används MATLAB-simuleringar för att testa det föreslagna systemet.Genom att kombinera teoretisk analys och praktiska simuleringar ger studien värdefullainsikter och designstrategier som bidrar till framsteg inom WNCS, vilket möjliggörtillförlitlighet och stabilitet även vid potentiella fel på relänoder.

[pt] A eletrocoagulação é um método emergente de tratamento de efluentes que combina os benefícios da coagulação, flotação e eletroquímica. Devido à complexidade inerente às operações de uma estação de tratamento de efluentes, é um desafio reagir com rapidez e precisão às condições dinâmicas necessárias para manter a qualidade do efluente. Portanto, esta tese tem como objetivo identificar a condição operacional de uma estação de tratamento de efluentes que adotou a eletrocoagulação para o tratamento de efluentes de panificação. Três condições operacionais baseadas em clarificação do efluente e lodo da reação foram as variáveis-alvo. A tese está dividida em dois ensaios. O primeiro usou sete métodos de seleção de atributos para selecionar as variáveis mais importantes em um determinado conjunto de dados. O desempenho dos modelos de classificação de redes neurais treinados no conjunto de atributos original foi comparado ao desempenho daqueles que foram treinados em um subconjunto curado usando técnicas de seleção de atributos. O modelo que utilizou a seleção de atributos apresentou o melhor desempenho (F1-score = 0,92) e uma melhoria de mais de 30 por cento na prevenção de falsos positivos. A segunda contribuição trouxe um modelo que poderia detectar o comportamento anômalo do processo usando apenas imagens coloridas da superfície do efluente obtidas através de dois módulos de câmera de tamanho pequeno. O desempenho de vários métodos, incluindo MLP, LSTM, SVM e XGBoost foi avaliado. O modelo LSTM superou os outros em termos de Precisão (84,620 por cento), Recall (84,531 por cento) e F1-score (84,499 por cento), mas o modelo XGBoost vem em segundo lugar com Precisão (83,922 por cento), Recall (82,272 por cento) e F1-score (83,005 por cento). / [en] Electrocoagulation is an emerging wastewater treatment method that combines the benefits of coagulation, flotation, and electrochemistry. As a result of the inherent complexity of processes associated with wastewater treatment plants, it is difficult to respond swiftly and correctly to the dynamic circumstances that are necessary to ensure effluent quality. Therefore, this thesis aims to identify the operational condition of a wastewater treatment plant that has adopted electrocoagulation for treating bakery wastewater. Three operational conditions based on effluent clarification and reaction sludge were the target variables. The thesis is divided into two essays. The first endeavor used seven feature selection methods to select the most important features in a given dataset. The performance of neural network classification models trained on the original feature set was compared to the performance of those that were trained on a subset of features that had been curated using feature selection techniques. The model that utilised feature selection was found to have the best performance (F1-score = 0.92) and an improvement of more than 30 percent in preventing false positives. The second contribution brought a model that could detect anomalous process behavior using only wastewater surface color images from two small-size camera modules. The performance of various methods, including MLP, LSTM, SVM, and XGBoost was assessed. The LSTM model outperformed the others in terms of macro average Precision (84.620 percent), Recall (84.531 percent), and F1-score (84.499 percent), but the XGBoost model comes closely in second with Precision (83.922 percent), Recall (82.272 percent), and F1-score (83.005 percent).

This thesis aims to investigate a fault detection and diagnosis system for automotive cameras using unsupervised learning. 1) Can a front-looking wide-angle camera image dataset be created using Hardware-in-Loop (HIL) simulations? 2) Can an Adversarial Autoencoder (AAE) based unsupervised camera fault detection and diagnosis method be crafted for SPA2 Vehicle Control Unit (VCU) using an image dataset created using Hardware-inLoop? 3) Does using AAE surpass the performance of using Variational Autoencoder (VAE) for the unsupervised automotive camera fault diagnosis model? In the field of camera fault studies, automotive cameras stand out for its complex operational context, particularly in Advanced Driver-Assistance Systems (ADAS) applications. The literature review finds a notable gap in comprehensive image datasets addressing the image artefact spectrum of ADAS-equipped automotive cameras under real-world driving conditions. In this study, normal and fault scenarios for automotive cameras are defined leveraging published and company studies and a fault diagnosis model using unsupervised learning is proposed and examined. The types of image faults defined and included are Lens Flare, Gaussian Noise and Dead Pixels. Along with normal driving images, a balanced fault-injected image dataset is collected using real-time sensor simulation under driving scenario with industrially-recognised HIL setup. An AAE-based unsupervised automotive camera fault diagnosis system using VGG16 as encoder-decoder structure is proposed and experiments on its performance are conducted on both the selfcollected dataset and fault-injected KITTI raw images. For non-processed KITTI dataset, morphological operations are examined and are employed as preprocessing. The performance of the system is discussed in comparison to supervised and unsupervised image partition methods in related works. The research found that the AAE method outperforms popular VAE method, using VGG16 as encoder-decoder structure significantly using 3-layer Convolutional Neural Network (CNN) and ResNet18 and morphological preprocessings significantly ameliorate system performance. The best performing VGG16- AAE model achieves 62.7% accuracy to diagnosis on own dataset, and 86.4% accuracy on double-erosion-processed fault-injected KITTI dataset. In conclusion, this study introduced a novel scheme for collecting automotive sensor data using Hardware-in-Loop, utilised preprocessing techniques that enhance image partitioning and examined the application of unsupervised models for diagnosing faults in automotive cameras. / Denna avhandling syftar till att undersöka ett felupptäcknings- och diagnossystem för bilkameror med hjälp av oövervakad inlärning. De huvudsakliga forskningsfrågorna är om en bilduppsättning från en frontmonterad vidvinkelkamera kan skapas med hjälp av Hardware-in-Loop (HIL)-simulationer, om en Adversarial Autoencoder (AAE)-baserad metod för oövervakad felupptäckt och diagnos för SPA2 Vehicle Control Unit (VCU) kan utformas med en bilduppsättning skapad med Hardware-in-Loop, och om användningen av AAE skulle överträffa prestandan av att använda Variational Autoencoder (VAE) för den oövervakade modellen för felanalys i bilkameror. Befintliga studier om felanalys fokuserar på roterande maskiner, luftbehandlingsenheter och järnvägsfordon. Få studier undersöker definitionen av feltyper i bilkameror och klassificerar normala och felaktiga bilddata från kameror i kommersiella passagerarfordon. I denna studie definieras normala och felaktiga scenarier för bilkameror och en modell för felanalys med oövervakad inlärning föreslås och undersöks. De typer av bildfel som definieras är Lens Flare, Gaussiskt brus och Döda pixlar. Tillsammans med normala bilder samlas en balanserad uppsättning felinjicerade bilder in med hjälp av realtidssensor-simulering under körscenarier med industriellt erkänd HIL-uppsättning. Ett AAE-baserat system för oövervakad felanalys i bilkameror med VGG16 som kodaredekoderstruktur föreslås och experiment på dess prestanda genomförs både på den självinsamlade uppsättningen och felinjicerade KITTI-raw-bilder. För icke-behandlade KITTI-uppsättningar undersöks morfologiska operationer och används som förbehandling. Systemets prestanda diskuteras i jämförelse med övervakade och oövervakade bildpartitioneringsmetoder i relaterade arbeten. Forskningen fann att AAE-metoden överträffar den populära VAEmetoden, genom att använda VGG16 som kodare-dekoderstruktur signifikant med ett 3-lagers konvolutionellt neuralt nätverk (CNN) och ResNet18 och morfologiska förbehandlingar förbättrar systemets prestanda avsevärt. Den bäst presterande VGG16-AAE-modellen uppnår 62,7 % noggrannhet för diagnos på egen uppsättning, och 86,4 % noggrannhet på dubbelerosionsbehandlad felinjicerad KITTI-uppsättning. Sammanfattningsvis introducerade denna studie ett nytt system för insamling av data från bilsensorer med Hardware-in-Loop, utnyttjade förbehandlingstekniker som förbättrar bildpartitionering och undersökte tillämpningen av oövervakade modeller för att diagnostisera fel i bilkameror.

Sensor fault detection is an actively researched area and there are a plethora of studies on sensor fault detection in various applications such as nuclear power plants, wireless sensor networks, weather stations and nuclear fusion. However, there does not seem to be any study focusing on detecting sensor faults in the threaded fastening assembly application. Since the threaded fastening tools use torque and angle measurements to determine whether or not a screw or bolt has been fastened properly, faulty measurements from these sensors can have dire consequences. This study aims to investigate the use of machine learning to detect a subtle kind of sensor faults, common in this application, that are difficult to detect using canonical model-based approaches. Because of the subtle and infrequent nature of these faults, a two-stage system was designed. The first component of this system is given sensor data from a tightening and then tries to classify each data point in the sensor data as normal or faulty using a combination of low-pass filtering to generate residuals and a support vector machine to classify the residual points. The second component uses the output from the first one to determine if the complete tightening is normal or faulty. Despite the modest performance of the first component, with the best model having an F1-score of 0.421 for classifying data points, the design showed promising performance for classifying the tightening signals, with the best model having an F1-score of 0.976. These results indicate that there indeed exist patterns in these kinds of torque and angle multi-sensor signals that make machine learning a feasible approach to classify them and detect sensor faults. / Sensorfeldetektering är för nuvarande ett aktivt forskningsområde med mängder av studier om feldetektion i olika applikationer som till exempel kärnkraft, trådlösa sensornätverk, väderstationer och fusionskraft. Ett applikationsområde som inte verkar ha undersökts är det inom åtdragningsmontering. Eftersom verktygen inom åtdragningsmontering använder mätvärden på vridmoment och vinkel för att avgöra om en skruv eller bult har dragits åt tillräckligt kan felaktiga mätvärden från dessa sensorer få allvarliga konsekvenser. Målet med denna studie är att undersöka om det går att använda maskininlärning för att detektera en subtil sorts sensorfel som är vanlig inom åtdragningsmontering och har visat sig vara svåra att detektera med konventionella modell-baserade metoder. I och med att denna typ av sensorfel är både subtila och infrekventa designades ett system bestående av två komponenter. Den första får sensordata från en åtdragning och försöker klassificera varje datapunkt som antingen normal eller onormal genom att uttnyttja en kombination av lågpassfiltrering för att generera residualer och en stödvektormaskin för att klassificera dessa. Den andra komponenten använder resultatet från den första komponenten för att avgöra om hela åtdragningen ska klassificeras som normal eller onormal. Trots att den första komponenten hade ett ganska blygsamt resultat på att klassificera datapunkter så visade systemet som helhet mycket lovande resultat på att klassificera hela åtdragningar. Dessa resultat indikerar det finns mönster i denna typ av sensordata som gör maskininlärning till ett lämpligt verktyg för att klassificera datat och detektera sensorfel.

No description available.

No description available.

No description available.