Governmental and private institutions rely heavily on reliable computer networks for their everyday business transactions. The downtime of their infrastructure networks may result in millions of dollars in cost. Fault management systems are used to keep today’s complex networks running without significant downtime cost, either by using active techniques or passive techniques. Active techniques impose excessive management traffic, whereas passive techniques often ignore uncertainty inherent in network alarms,leading to unreliable fault identification performance. In this research work, new algorithms are proposed for both types of techniques so as address these handicaps. Active techniques use probing technology so that the managed network can be tested periodically and suspected malfunctioning nodes can be effectively identified and isolated. However, the diagnosing probes introduce extra management traffic and storage space. To address this issue, two new CSP (Constraint Satisfaction Problem)-based algorithms are proposed to minimize management traffic, while effectively maintain the same diagnostic power of the available probes. The first algorithm is based on the standard CSP formulation which aims at reducing the available dependency matrix significantly as means to reducing the number of probes. The obtained probe set is used for fault detection and fault identification. The second algorithm is a fuzzy CSP-based algorithm. This proposed algorithm is adaptive algorithm in the sense that an initial reduced fault detection probe set is utilized to determine the minimum set of probes used for fault identification. Based on the extensive experiments conducted in this research both algorithms have demonstrated advantages over existing methods in terms of the overall management traffic needed to successfully monitor the targeted network system. Passive techniques employ alarms emitted by network entities. However, the fault evidence provided by these alarms can be ambiguous, inconsistent, incomplete, and random. To address these limitations, alarms are correlated using a distributed Dempster-Shafer Evidence Theory (DSET) framework, in which the managed network is divided into a cluster of disjoint management domains. Each domain is assigned an Intelligent Agent for collecting and analyzing the alarms generated within that domain. These agents are coordinated by a single higher level entity, i.e., an agent manager that combines the partial views of these agents into a global one. Each agent employs DSET-based algorithm that utilizes the probabilistic knowledge encoded in the available fault propagation model to construct a local composite alarm. The Dempster‘s rule of combination is then used by the agent manager to correlate these local composite alarms. Furthermore, an adaptive fuzzy DSET-based algorithm is proposed to utilize the fuzzy information provided by the observed cluster of alarms so as to accurately identify the malfunctioning network entities. In this way, inconsistency among the alarms is removed by weighing each received alarm against the others, while randomness and ambiguity of the fault evidence are addressed within soft computing framework. The effectiveness of this framework has been investigated based on extensive experiments. The proposed fault management system is able to detect malfunctioning behavior in the managed network with considerably less management traffic. Moreover, it effectively manages the uncertainty property intrinsically contained in network alarms,thereby reducing its negative impact and significantly improving the overall performance of the fault management system.

This dissertation presents a model-based reasoning architecture with a two fold purpose: to detect and classify component faults from observable system behavior, and to generate fault propagation models so as to make a more accurate estimation of current operational risks. It incorporates a novel approach to system level diagnostics by addressing the need to reason about low-level inaccessible components from observable high-level system behavior. In the field of complex system maintenance it can be invaluable as an aid to human operators. The first step is the compilation of the database of functional descriptions and associated fault-specific features for each of the system components. The system is then analyzed to extract structural information, which, in addition to the functional database, is used to create the structural and functional models. A fault-symptom matrix is constructed from the functional model and the features database. The fault threshold levels for these symptoms are founded on the nominal baseline data. Based on the fault-symptom matrix and these thresholds, a diagnostic decision tree is formulated in order to intelligently query about the system health. For each faulty candidate, a fault propagation tree is generated from the structural model. Finally, the overall system health status report includes both the faulty components and the associated at risk components, as predicted by the fault propagation model.

Continuous integration/delivery and deployment consist of many automated tests, some of which may fail leading to faulty software. Similar faults may occur in different stages of the software production lifecycle and it is necessary to identify similar faults and cluster them into fault types in order to minimize troubleshooting time. Pretrained transformer based language models have been proven to achieve state of the art results in many natural language processing tasks like measuring semantic textual similarity. This thesis aims to investigate whether it is possible to cluster and identify new fault types by using a transformer based model to create context aware vector representations of fault records, which consists of numerical data and logs with domain specific technical terms. The clusters created were compared against the clusters created by an existing system, where log files are grouped by manual specified filters. Relying on already existing fault types with associated log data, this thesis shows that it is possible to finetune a transformer based model for a classification task in order to improve the quality of text embeddings. The embeddings are clustered by using density based and hierarchical clustering algorithms with cosine distance. The results show that it is possible to cluster log data and get comparable results to the existing manual system, where the cluster similarity was assessed with V-measure and Adjusted Rand Index. / Kontinuerlig integration består automatiserade tester där det finns risk för att några misslyckas vilket kan leda till felaktig programvara. Liknande fel kan uppstå under olika faser av en programvarans livscykel och det är viktigt att identifiera och gruppera olika feltyper för att optimera felsökningsprocessen. Det har bevisats att språkmodeller baserade på transformatorarkitekturen kan uppnå höga resultat i många uppgifter inom språkteknologi, inklusive att mäta semantisk likhet mellan två texter. Detta arbete undersöker om det är möjligt att gruppera och identifiera nya feltyper genom att använda en transformatorbaserad språkmodell för att skapa numeriska vektorer av loggtext, som består av domänspecifika tekniska termer och numerisk data. Klustren jämförs mot redan existerande grupperingar som skapats av ett befintligt system där feltyper identifieras med manuellt skrivna filter. Det här arbetet visar att det går att förbättra vektorrepresenationerna skapade av en språkmodell baserad på transformatorarkitekturen genom att tilläggsträna modellen för en klassificeringsuppgift. Vektorerna grupperas med hjälp av densitetsbaserade och hierarkiska klusteralgoritmer. Resultaten visar att det är möjligt att skapa vektorer av logg-texter med hjälp av en transformatorbaserad språkmodell och få jämförbara resultat som ett befintligt manuellt system, när klustren evaluerades med V-måttet och Adjusted Rand Index.

The report is a study of error management and localization in underground cable network with very little overhead line. Error management is treated with the theoretical troubleshooting model and theoretical model with practical feedback. The work relates to the fact that troubleshooting has been complicated when the supply safety in the electricity grid increases. Difficult troubleshooting is followed by a difficult reset work in case of malfunctioning in the network and it leads to longer interruptions.The grid is built with main feed and radial feeds. In the cable stretches, an interval of the short-circuit current occurs with a minimum and a maximum current below the distance. Depending on how the network is built, multiples of same error current is created, that create harder error-handling. The analysis examines which locations in the grid, which help increase the number of alternative malfunctions and if there is any alternative method to solve the problem. The report deals with the connection between currents in main feed and in radial feeds along the line. There appears to be a connection between the emergences of currents with similar current values in several places. A network consisting of a main feed and a plurality of radial feeds, the occurrence of streams in several places is a disadvantage in the troubleshooting task and the work is adversely affected and the troubleshooting work becomes more difficult. / Rapporten är en studie om felhantering och fellokalisering i ett markbelagt kabelnät med en väldigt liten del friledning. Felhanteringen be-handlas med både dataprogramvara och teoretiska beräkningar. Arbetet relaterar till att felsökningen försvåras när leveranssäkerheten i elnätet ökar. Svårare felsökning är följt av ett försvårat återställningsar-bete vid fel i nätet som leder till längre avbrottstider. Elnätsområdet är uppbyggt med huvudmatning samt radiella matning-ar. I kabelsträckorna uppkommer ett intervall på kortslutningsström-men med en minsta och en högsta ström under sträckan. Beroende på hur nätet är uppbyggt skapas multiplar av felställen som skapar svårare felhantering. Analysen undersöker vilka ställen i elnätet som bidrar till att antalet alternativa felställen ökar och om det finns någon alternativ metod att lösa problemet. Rapporten behandlar sammanbandet mellan strömmar i huvudmatning och i radiella matningar längs ledningens sträcka. Det visar sig vara ett samband mellan uppkomsten av strömmar med liknande strömvärden på ett flertal ställen. Ett nät som består av en huvudmatning och ett flertal radiella matningar är uppkomsten av samma kortslutningsströmmar på flera ställen. Det är en nackdel i felsökningsarbetet och arbetet påverkas negativt och fel-sökningsarbetet blir svårare.

Blends of poly butylene terephthalate (PBT) and polycarbonate (PC) form a very important class of commercial blends in numerous applications requiring materials with good chemical resistance, impact resistance even at low temperatures, and aesthetic and flow characteristics. PC and PBT are usually blended in a twin screw extruder (TSE). Product melt volume flow rate (MVR) is a property used to monitor product quality while blending the PC/PBT in a twin screw extruder. It is usually measured off line in a quality control laboratory using extrusion plastometer on samples collected discretely during the compounding operation. Typically a target value representing the desired value of the quality characteristics for an in-control process, along with upper and lower control limits are specified. As long as the MVR measurement is within the control limits, the sample is approved and the whole compounded blend is assumed to meet the specification. Otherwise, the blend is rejected. Because of infrequent discrete sampling, corrective actions are usually applied with delay, thus resulting in wasted material. It is important that the produced PC/PBT blend pellets have consistent properties. Variability and fault usually arise from three sources: human errors, feed material variability, and machine operation (i.e. steady state variation). Among these, the latter two are the major ones affecting product quality. The resulting variation in resin properties contributes to increased waste products, larger production cost and dissatisfied customers. Motivated by this, the objective of this project was to study the compounding operation of PC/PBT blend in a twin screw extruder and to develop a feasible methodology that can be applied on-line for monitoring properties of blends on industrial compounding operations employing available extruder input and output variables such as screw speed, material flow rate, die pressure and torque. To achieve this objective, a physics-based model for a twin screw extruder along with a MVR model were developed, examined and adapted for this study, and verified through designed experiments. This dynamic model for a TSE captures the important dynamics, and relates measurable process variables (screw speed, torque, feed rates, pressure etc.) to ones that are not being measured (material holdups and compositions at the partially and filled section along a TSE barrel). This model also provides product quality sensors or inferential estimation techniques for prediction of viscosity and accordingly MVR. The usefulness of the model for inferential MVR sensing and fault diagnosis was demonstrated on experiments performed on a 58 mm co-rotating twin-screw extruder for an industrial compounding operation at a SABIC Innovative Plastics plant involving polycarbonate – poly butylene terephthalate blends. The results showed that the model has the capability of identifying faults (i.e., process deviation from the nominal conditions) in polymer compounding operations with the twin screw extruder. For instance, the die pressure exhibited a change as a function of changes in raw materials and feed composition of PC and PBT. In the presence of deviations from nominal conditions, the die pressure parameters are updated. These die pressure model parameters were identified and updated using the recursive parameter estimation method. The recursive identification of the die pressure parameters was able to capture very well the effects of changes in raw material and/or composition on the die pressure. In addition, the developed MVR model showed a good ability in monitoring product MVR on-line and inferentially from output process variables such as die pressure which enables quick quality control to maintain products within specification limits and to minimize waste production.

Blends of poly butylene terephthalate (PBT) and polycarbonate (PC) form a very important class of commercial blends in numerous applications requiring materials with good chemical resistance, impact resistance even at low temperatures, and aesthetic and flow characteristics. PC and PBT are usually blended in a twin screw extruder (TSE). Product melt volume flow rate (MVR) is a property used to monitor product quality while blending the PC/PBT in a twin screw extruder. It is usually measured off line in a quality control laboratory using extrusion plastometer on samples collected discretely during the compounding operation. Typically a target value representing the desired value of the quality characteristics for an in-control process, along with upper and lower control limits are specified. As long as the MVR measurement is within the control limits, the sample is approved and the whole compounded blend is assumed to meet the specification. Otherwise, the blend is rejected. Because of infrequent discrete sampling, corrective actions are usually applied with delay, thus resulting in wasted material. It is important that the produced PC/PBT blend pellets have consistent properties. Variability and fault usually arise from three sources: human errors, feed material variability, and machine operation (i.e. steady state variation). Among these, the latter two are the major ones affecting product quality. The resulting variation in resin properties contributes to increased waste products, larger production cost and dissatisfied customers. Motivated by this, the objective of this project was to study the compounding operation of PC/PBT blend in a twin screw extruder and to develop a feasible methodology that can be applied on-line for monitoring properties of blends on industrial compounding operations employing available extruder input and output variables such as screw speed, material flow rate, die pressure and torque. To achieve this objective, a physics-based model for a twin screw extruder along with a MVR model were developed, examined and adapted for this study, and verified through designed experiments. This dynamic model for a TSE captures the important dynamics, and relates measurable process variables (screw speed, torque, feed rates, pressure etc.) to ones that are not being measured (material holdups and compositions at the partially and filled section along a TSE barrel). This model also provides product quality sensors or inferential estimation techniques for prediction of viscosity and accordingly MVR. The usefulness of the model for inferential MVR sensing and fault diagnosis was demonstrated on experiments performed on a 58 mm co-rotating twin-screw extruder for an industrial compounding operation at a SABIC Innovative Plastics plant involving polycarbonate – poly butylene terephthalate blends. The results showed that the model has the capability of identifying faults (i.e., process deviation from the nominal conditions) in polymer compounding operations with the twin screw extruder. For instance, the die pressure exhibited a change as a function of changes in raw materials and feed composition of PC and PBT. In the presence of deviations from nominal conditions, the die pressure parameters are updated. These die pressure model parameters were identified and updated using the recursive parameter estimation method. The recursive identification of the die pressure parameters was able to capture very well the effects of changes in raw material and/or composition on the die pressure. In addition, the developed MVR model showed a good ability in monitoring product MVR on-line and inferentially from output process variables such as die pressure which enables quick quality control to maintain products within specification limits and to minimize waste production.

Made available in DSpace on 2014-06-11T19:27:14Z (GMT). No. of bitstreams: 0 Previous issue date: 2004-02-20Bitstream added on 2014-06-13T18:55:44Z : No. of bitstreams: 1 furtado_rm_me_ilha.pdf: 1436216 bytes, checksum: 09e5f73855e5a468589756fca572b577 (MD5) / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / A proposta deste trabalho é a obtenção de uma metodologia robusta para identificação de falhas estruturais combinando as vantagens de duas metodologias, que não são baseadas em modelos matemáticos, ou seja: impedância elétrica obtida com atuador e sensor piezocerâmico(materiais inteligentes) e redes neurais artificiais. O termo materiais inteligentes (smart materials) conhecido também por material ativo é dado a uma classe de material que exibe propriedades não encontradas em materiais convencionais. Alguns destes materiais são: compostos de materiais piezelétricos, eletrorresistivo e magnetorresistivo, fluidos e sólidos electro-reológicos, e outros. Uma das principais características do PZT (Titanato Zirconato de Chumbo), que permite utilizá-lo como sensor e atuador, é o efeito piezelétrico, ou seja, a aplicação de um campo elétrico resulta em deformação do material (efeito inverso), enquanto, a aplicação de tensão mecânica resulta no surgimento de um campo elétrico (efeito direto). Estas características associadas ao conceito de impedância elétrica e ao conceito de falha métrica permitem a localização e o monitoramento da falha. Esta técnica utiliza altas freqüências e excita os modos locais, proporcionando, assim, o monitoramento de qualquer mudança da impedância mecânica estrutural na região de influência do PZT. Redes neurais artificiais (RNA) fazem parte de um amplo conceito chamado inteligência artificial. Redes neurais têm sua base associada ao funcionamento do cérebro humano, que após treinamento possuem a capacidade de aprender. Esta ciência é objeto de estudo em diversos centros de pesquisa e, embora já tenha grande aplicabilidade, o sucesso de sua utilização depende do caso em que está sendo aplicada e de certa sutileza do projetista, uma vez que o processo ainda é empírico e teorias ainda... . / The proposal of this work is the obtaining of a robust methodology for identification of structural faults combining the advantages of two methodologies, which are not based on mathematical models. The methodology applies electric impedance technique, obtained with actuator and sensor piezoceramic (smart materials), and artificial neural networks. The term smart materials is given for a material class that not exhibits properties found in conventional materials. Some of these materials are: composed of piezoelectric material, electrostrictive and magnetostrictive, electrorheological fluids and solids shape memory alloys, and others. One of the main characteristics of PZT (Lead Zirconate Titanate), that allows to use it as sensor and actuator, is the piezoelectric effect, where the application of an electric field results in deformation of the material (inverse effect), while the application of mechanical tension results in the appearance of an electric field (direct effect). These characteristics associated to the concept of electric impedance and the concept of metric fault allow the location and the monitoring of the fault. This technique uses high frequencies and low voltage and it excites local modes, providing, the monitoring of any change on the structural mechanical impedance in the area of influence of the PZT. Artificial Neural Networks (ANN) are part of a wide concept called artificial intelligence. Neural networks has its base associated to the operation of the human brain, that after training possess the capacity to learn. This science is a study object in several research centers and, although it already has great application. The success of its use depends of the case and planner's certain keenness, once the process is still empiric and theories are still being formulated. Several conceptions of neural networks... (Complete abstract, click electronic address below).

[pt] Quando ocorre algum tipo de falta em uma Linha de Transmissão (LT), sua localização exata é essencial para uma rápida recomposição do Sistema Elétrico. Métodos que utilizam tensão e corrente de apenas um terminal contêm simplificações que podem acarretar erros. Esta dissertação investiga a aplicação de Redes Neurais Artificiais (RNA) na obtenção de uma nova forma de identificar o tipo do curto- circuito e determinar a sua localização, utilizando dados obtidos em somente um terminal. O trabalho consiste de 4 partes principais: estudo bibliográfico da área de Redes Neurais; simulações de faltas para a obtenção de padrões; definição e implementação dos modelos de Redes Neurais para identificação e localização da falta; e estudos de casos. Na fase do estudo bibliográfico sobre RNA, foi verificado que as topologias de redes mais usuais são as Feed- Forward, que podem ter uma ou mais camadas de Elementos Processadores (EP), sendo as redes com múltiplas camadas (Multi-Layer) a configuração mais completa. Para treinamento da rede, o algoritmo de aprendizado mais utilizado é o Back Propagation. Como fruto deste estudo bibliográfico é apresentado neste trabalho um resumo sobre RNA. Nas simulações de faltas para obtenção dos padrões de treinamento e teste, foi utilizado um sistema automático que, através da combinação de vários parâmetros do sistema elétrico, gera situações diferentes de falta. Este sistema utiliza como base o programa Alternative Transient Program - ATP. Neste trabalho o sistema elétrico está representado por uma LT de 345 KV, com fontes equivalentes representando um sistema real de Furnas Centrais Elétricas. Todos o sinais de tensão e corrente utilizados são representados por fasores de 60 Hz, obtidos através da Transformada Discreta de Fourier (TDF). Os modelos de RNAs para identificação e localização de falta foram implementados com sub-rotinas de redes neurais do programa MATLAB ver. 6.0, representados por Redes Perceptron Multicamadas (Multi Layer Perceptron), treinadas com algoritmo Back Propagation com taxa de aprendizado adaptativa e o termo momentum fixo. Um modelo único de RNA identifica quais as fases (A, B, C e T) envolvidas, classificando o tipo de falta, que pode ser: Monofásica; Bifásica; Bifásica-Terra ou Trifásica. Para a localização da falta, foram definidas 4 arquiteturas de RNA, uma para cada tipo de falta. A ativação de cada topologia de RNA para localização é definida em função do tipo de falta classificada no modelo de identificação com RNA. Na etapa de estudo de casos testou-se o desempenho de cada modelo de RNA utilizando casos de testes em outras situações de falta, diferentes dos conjuntos de treinamento. A RNA de identificação de falta foi avaliada para situações de faltas envolvendo outras LTs, com diferentes níveis de tensão. Os resultados das 4 RNAs de localização da falta foram comparados com os resultados obtidos utilizando o método tradicional, tanto para os casos simulados quanto para algumas situações reais de falta. A utilização de RNAs para a identificação e a localização de falta mostrouse bastante eficiente para os casos analisados, comprovando a aplicabilidade das redes neurais nesse problema. / [en] When a kind of fault occurs in a Transmission Line, its exact location is essential for a fast reclosing of the Electric System. Methods that use voltages and currents from only one terminal contain simplifications that can to cause mistakes. This paper presents an investigation about application of Artificial Neural Network (ANN) obtaining a new way of identification for the type of the short circuit and its location, using data obtained only in one terminal. The work consists on the following 4 main parts: bibliographical study of Neural Network`s area; simulations of faults in order to obtain of patterns; definition and implementation of Neural Network`s models for identification and location of the fault; and studies of cases. In the bibliographical study step on ANN, it was verified that the topologies for the more usual nets are Feed- Forward,that can have one or more layers of Processor Elements (PE), being the nets with multiple layers the most complete configuration. For the net training, the more used learning algorithm is Back Propagation. Product of this bibliographical study presents in this work a summary about ANN. In the faults simulations in order to obtain the training patterns and test, it was used an automatic system that, through the combination of several parameters of the electric system, generates different fault situations. This system uses as base the program Alternative Transient Program - ATP. In this work the electric system is represented by a Transmission Line of 345 KV, with equivalent sources representing a real system of Furnas Centrais Elétricas. All the voltages and currents signs used are represented by fasors of 60 Hz, obtained from Discret Fourier Transformer (DFT). The ANN models for identification and fault location were implemented with subroutines of neural network of the program MATLAB version 6.0, represented by Multi Layer Perceptron, with algorithm Back Propagation with tax of adaptive learning and the term momentum fixed. Only one model of ANN identifies which phases (A, B, C and T) are involved, classifying the fault type, that can be: Singlephase; Phase-to-Phase; Double Phase-to-Ground or Three-phase. For the fault location, they were defined 4 architectures of ANN, one for each type of fault. The activation of each topology of ANN for location is defined depending on of the fault type classified in the identification model with ANN. In the stage of cases study the representation of each model of ANN was tested using cases of tests in other fault situations, different from the training groups. The ANN of fault identification was evaluated for situations of faults involving other Transmission Line, with different voltage levels. The results of 4 ANNs of fault location were compared with the obtained results using the traditional method, so much for the simulated cases as for some real situations of fault. The use of ANNs for the identification and the fault location has shown quite efficient for the analyzed cases, proving the applicability of the neural networks in that problem.

Les travaux développés durant cette thèse concernent la détection et l'identification des défauts simples et multiples d'ouverture des transistors dans un convertisseur statique associé à une machine asynchrone. Pour aborder cette problématique, nous avons commencé par l'analyse des potentialités, des faiblesses et des incertitudes des techniques qui ont initiés notre démarche. Ensuite, nous avons présenté deux méthodologies permettant d'analyser les performances du moteur asynchrone en présence des défauts dans une ou plusieurs cellules de commutation. Cette étude préliminaire nous a permis ainsi de proposer deux nouvelles stratégies de diagnostic sans référence basées sur l'approche signal. Les signaux électriques (courants ou tensions) disponibles à la sortie du convertisseur statique sont utilisés pour alimenter le processus de diagnostic. La première stratégie retenue est basée sur l'analyse qualitative des tensions de sortie entre phases du convertisseur et des signaux de commande appliqués aux transistors pendant les instants de commutation. Grâce à une représentation instantanée de ces grandeurs, à l'échelle de la période de découpage, nous avons pu mettre en évidence des caractéristiques favorables à la détection des défauts simples et multiples d'ouverture des transistors. L'implémentation pratique de cette première approche a été réalisée au moyen d'une technologie analogique permettant ainsi de minimiser le temps de retard à la détection jusqu'à quelques dizaines de microsecondes. / The main goal of this thesis concerns the detection and identification of simple and multiple open-circuit faults in voltage source inverters (VSIs)-fed induction motor drives. In first step, the potentialities, the weaknesses as well as the uncertainties of the previously published works have been discussed. The second step was dedicated to the study of the inverter faults impact on the induction motor. For this purpose, we have proposed two methodologies permitting the characterization of the electromagnetic torque behaviour as well as the electric variables of the induction motor under the open- and short-circuit faults. These preliminary studies allowed to propose two novel signal-based approaches for open-circuit fault diagnosis in voltage source inverter. The measured outputs inverter voltages and currents have been used as the input quantities for the fault detection and identification (FDI) process. The first approach consists in analyzing the pulse-width modulation (PWM) switching signals and the line-to-line voltage levels during the switching times, under both healthy and faulty operating conditions. For this purpose, we have adopted an instantaneous representation of these variables, which permits their analysis over one switching period. The fault diagnosis scheme is achieved using simple analog device. This circuit allows an accurate single and multiple faults diagnosis, and a minimization of the fault detection time which becomes about a few tens of microseconds.

Self-healing systems promise operating cost reductions in large-scale computing environments through the automated detection of, and recovery from, faults. However, at present there appears to be little known empirical evidence comparing the different approaches, or demonstrations that such implementations reduce costs. This thesis compares previous and current self-healing approaches before demonstrating a new, unsupervised approach that combines artificial neural networks with performance tests to perform fault identification in an automated fashion, i.e. the correct and accurate determination of which computer features are associated with a given performance test failure. Several key contributions are made in the course of this research including an analysis of the different types of self-healing approaches based on their contextual use, a baseline for future comparisons between self-healing frameworks that use artificial neural networks, and a successful, automated fault identification in cloud infrastructure, and more specifically virtual machines. This approach uses three established machine learning techniques: Naïve Bayes, Baum-Welch, and Contrastive Divergence Learning. The latter demonstrates minimisation of human-interaction beyond previous implementations by producing a list in decreasing order of likelihood of potential root causes (i.e. fault hypotheses) which brings the state of the art one step closer toward fully self-healing systems. This thesis also examines the impact of that different types of faults have on their respective identification. This helps to understand the validity of the data being presented, and how the field is progressing, whilst examining the differences in impact to identification between emulated thread crashes and errant user changes – a contribution believed to be unique to this research. Lastly, future research avenues and conclusions in automated fault identification are described along with lessons learned throughout this endeavor. This includes the progression of artificial neural networks, how learning algorithms are being developed and understood, and possibilities for automatically generating feature locality data.