With the fast evolution of the Industry 4.0, the increased use of sensors and the rapid development of the Internet of Things (IoT), and the adoption of artificial intelligence methods, smart factories can automate their processes to vastly improve their efficiency and production quality. However, even the most well cared-for machines develop faults eventually. Given that Prognostics and Health Management (PHM) is an indispensable aspect for proper machine performance, Predictive Maintenance (PdM) is an emerging topic within maintenance methodologies whose aim is to predict failure prior to occurrence with the goal of scheduling maintenance only when needed. As data can be collected faster than ever before, deep learning is an effective tool that can leverage big data for data-driven fault diagnosis methodologies. This thesis explores two different fault diagnosis methodologies associated with predictive maintenance: an anomaly detection using an Autoencoder Neural Network, and a fault classifier using a Convolutional Neural Network (CNN). The system under analysis is a 3phase AC induction motor commonly used in industry. Results show great performance and indicate the viability for the implementation of both methods in production applications. / Med den snabba utvecklingen av industri 4.0, den ökade användningen av sensorer och den snabba utvecklingen av Internet of Things samt införandet av metoder för artificiell intelligens kan smarta fabriker automatisera sina processer för att avsevärt förbättra effektiviteten och produktionskvaliteten. Även de mest välskötta maskinerna utvecklar dock fel så småningom. PHM är en oumbärlig aspekt för korrekt maskinunderhåll. PdM är ett nytt ämne inom underhållsmetodik som syftar till att förutsäga fel innan de inträffar, med målet att planera underhållet endast när det behövs. Eftersom data kan samlas in snabbare än någonsin tidigare är djupinlärning ett effektivt verktyg som kan utnyttja stora datamängder för datadrivna metoder för feldiagnostik. I den här uppsatsen undersöks två olika metoder för feldiagnostik i samband med förebyggande underhåll: en anomalidetektion med hjälp av ett neuralt nätverk med autoencoder och en felklassificering med hjälp av ett CNN. Det system som analyseras är en induktionsmotor med 3fas växelström som ofta används inom industrin. Resultaten visar på goda resultat och visar att det är möjligt att genomföra båda metoderna i produktionstillämpningar.

O presente trabalho apresenta a identificação de parâmetros de motor de indução linear tubular para a extração de petróleo, denominado de MATÆOS -- Motor Assíncrono Tubular para Aplicação na Extração de Óleo Subterrâneo. O primeiro passo para a identificação de parâmetros do motor foi a implementação da bancada de testes com o seu sistema de controle e supervisório. O modelo adotado corresponde à situação de operação em regime estacionário e consiste no circuito contendo resistores e indutores tradicionalmente utilizado para representar motores de indução. A identificação dos parâmetros foi realizada por meio de um procedimento de otimização tomando por base o balanço de potência do motor. Para isso, em função dos parâmetros do modelo, definiu-se e minimizou-se uma medida de erro entre as potências medidas e as calculadas pelo modelo para um conjunto pré-fixado de escorregamentos. A solução desse problema de otimização foi obtida por meio de uma busca exaustiva, uma vez que o número de variáveis independentes do problema é pequeno e uma região contendo o ponto de mínimo é conhecida com razoável confiança. Os resultados obtidos indicam que os valores dos parâmetros nominais a quente apresentam algumas diferenças em relação àqueles produzidos pelo método de identificação. Essas diferenças se devem possivelmente a variações nas características térmicas do motor e a erros introduzidos pelos instrumentos de medida. O modelo identificado apresentou uma menor disponibilidade de força e um pior rendimento quando comparado com os valores de projeto do modelo nominal a quente. Apesar disto, verificou-se uma boa proximidade entre a curva característica do modelo e as obtidas experimentalmente. / This work presents the identification of parameters of the tubular linear induction motor for the extraction of oil, called MATÆOS - which is the Portuguese acronym for Asynchronous Tubular Motor for Application to Subterranean Oil Extraction. The first step to the identification of the motor parameters was the implementation of the supervisory and control system for the test bench. The model, which is the usual one containing resistors and inductors, represents the motor operating in steady-state. The parameter identification problem was written as an optimization problem based on the motor power balance. An error between the measured power and the one given by the model was defined and minimized with respect to the model parameters in correspondence to a given set of slipping frequencies. The numerical solution to this problem was obtained by an exhaustive search algorithm. This choice was made because the number of independent variables is small and a region containing the solution is known with a significant degree of confidence. The results obtained show that the nominal values of the model parameters differ slightly from those produced by the identification method. Possibly, such differences are due to the variations in the model caused by temperature changes and to the measurement errors. The identified model presented both a lower available force and a worst efficiency when compared to the nominal design values for the warm motor. Despite this the characteristic curve of the model was quite close to the nominal design one.

This thesis deals with a new solution for medium voltage drives. Load Commutated Inverter (LCI) fed synchronous motor drive is a popular solution for high power drive applications. Though the induction machine is more rugged and cheaper compared to the synchronous machine, LCI fed induction motor drive solution is not available. The basic advantage of a synchronous machine over an induction machine is the fact that the synchronous machine can operate at leading power factor. Due to this property load commutation of SCR switches of the LCI is achievable for synchronous machine. On the contrary an induction machine always draws lagging power factor current; this makes it unsuitable as a drive motor for LCI technology. In this thesis a new LCI fed induction motor drive configuration is developed as an alternative for synchronous motor drives. A new variant of six phase induction motor is proposed in this context. The machine is named as Active Reactive Induction Machine (ARIM). The ARIM contains two sets of three-phase windings with isolated neutral. Both the windings have a common axis. One winding carries the active power and can be wound for higher voltage (say 11kV). The other winding supplies the total reactive power of the machine and can be wound for lower voltage (say 2.2 kV). The rotor is a standard squirrel cage. High power induction machines usually demand lesser magnitude of reactive power compared to the total power rating of the machine ( 20% ). Therefore excitation winding has a smaller fraction of the total machine rating compared to the power winding. A VSI with an LC filter supplies reactive power to the ARIM through the excitation winding and ensures leading power factor at the power winding. This is similar to the excitation control of the LCI fed synchronous machine. The direct VSI connection is possible due to the lower voltage rating for the excitation winding. In this way, the VSI voltage rating does not limit the highest motor voltage that can be handled. An LCI supplies the real power into the ARIM from the power winding. The LCI currents are quasi square wave in shape. Therefore they have rich low order harmonic content. They cause 6th and 12th harmonic torque pulsations in the machine. This is a problem for the LCI fed synchronous machine drive. In the proposed drive, the VSI can compensate these low frequency m.m.f. harmonics inside the machine air gap to remove torque pulsation and rotor harmonic losses. The advantage of the proposed topology is that no transformer is required to drive an 11kV machine. It is always desirable to feed sinusoidal voltage and current to both the power winding and the excitation winding. To address this problem, a second configuration is proposed. A low power three-level VSI is connected in shunt at the power winding with the proposed ARIM drive as discussed above. This VSI compensates the low frequency harmonic currents to achieve sinusoidal motor currents at the motor winding. This VSI acts as a shunt active filter and compensates for the lower order harmonics injected by the LCI. The proposed topologies have LC filters to maintain sinusoidal motor voltages and currents by absorbing the VSI switching frequency components. But the motor terminal voltage oscillates at system resonant frequency due to the presence of LC filters. These resonant components in the terminal voltages are required to be eliminated for smooth terminal voltages and safe load commutation of the thyristors. In this thesis a simple active damping method is proposed to mitigate these issues. The proposed topologies are experimentally verified with an ARIM with 415 V power winding and 220 V excitation winding. The control is carried out on a digital platform having a TMS 320LF 2407A DSP processor and an ALTERA CYCLONE FPGA processor. Results from the prototype experimental drive are presented to show the feasibility and performance of the proposed drive configurations.

Η παρούσα διπλωματική εργασία πραγματεύεται τη μελέτη και κατασκευή ενός μετατροπέα για την οδήγηση ενός μονοφασικού επαγωγικού κινητήρα. Η εργασία αυτή εκπονήθηκε στο Εργαστήριο Ηλεκτρομηχανικής Μετατροπής Ενέργειας του Τμήματος Ηλεκτρολόγων Μηχανικών και Τεχνολογίας Υπολογιστών της Πολυτεχνικής Σχολής του Πανεπιστημίου Πατρών. Σκοπός είναι η μελέτη και κατασκευή ενός τριφασικού αντιστροφέα τάσης για τη λειτουργία και τον έλεγχο των στροφών ενός μονοφασικού επαγωγικού κινητήρα. Αρχικά μελετάται η βασική αρχή λειτουργίας του μονοφασικού επαγωγικού κινητήρα και αναλύονται οι τεχνικές εκκίνησης που χρησιμοποιούνται για σύνδεση του κινητήρα απευθείας στο δίκτυο. Ακόμα παρουσιάζονται τα ισοδύναμα κυκλώματα λειτουργίας του μονοφασικού επαγωγικού κινητήρα, η εξίσωση της ηλεκτρομαγνητικής ροπής του και προσομοιώνεται η λειτουργία του για τη μελέτη της στατικής συμπεριφοράς του. Στη συνέχεια γίνεται μια θεωρητική ανάλυση του κυκλώματος του τριφασικού αντιστροφέα τάσης που κατασκευάστηκε, καθώς και όλων των υπόλοιπων κυκλωμάτων που είναι αναγκαία για τη λειτουργία του. Επιπροσθέτως αναλύεται η μέθοδος παλμοδότησης των διακοπτικών στοιχείων του αντιστροφέα τάσης, που είναι η "Ημιτονοειδής Διαμόρφωση του Εύρους των Παλμών" (SPWM). Στο επόμενο βήμα αναλύονται τα τεχνικά χαρακτηριστικά όλων των κυκλωμάτων που κατασκευάστηκαν, και περιγράφεται ο κώδικας του προγράμματος του μικροελεγκτή, που χρησιμοποιήθηκε για την παραγωγή των παλμών. Τέλος, παραθέτουμε παλμογραφήματα και μετρήσεις που προέκυψαν από τα πειράματα που διενεργήθηκαν μετά την ολοκλήρωση της κατασκευής. / This diploma thesis discourse the analysis and construction of a converter topology for single phase induction motor drives. The project was based in the Laboratory of Electromechanical Energy Conversion of the department of Electrical and Computer Engineering of School Engineering of University of Patras. The objective of this project is the analysis and construction of a three phase voltage inverter to control the speed of a single phase induction motor. The first stage of this work is the study of the basic principle of operation of the single phase induction motor and the analysis of the starting techniques, used for the direct connection to the power grid. The equivalent circuits of the running single phase induction motor and the equation of the electromagnetic torque are also presented in this project. Following, there is a theoretical analysis of the three phase voltage inverter circuit, as well of all the remaining circuit, necessary for its function. Moreover the method of pulse generation for the switching elements of the voltage inverter is analyzed, which is the “Sinusoidal Pulse Width Modulation”. The next step is the analysis of the technical characteristics all of the circuits developed, as well the description of the program code for the microcontroller, used to produce the pulses. Finally oscillograph figures and measurements, occurred from the experiments transacted after the finalization of the construction, are adduced.

Η θερμική ανάλυση είναι μια σημαντική περιοχή μελέτης και γίνεται περισσότερο σημαντική για την σχεδίαση ηλεκτρικών μηχανών εξαιτίας της ανάγκης για μείωση του όγκου των υλικών και του κόστους κατασκευής τους καθώς και για την αύξηση της απόδοσής τους. Είναι εξίσου σημαντική με την ηλεκτρομαγνητική ανάλυση μιας και η θέρμανση της μηχανής θα οριοθετήσει την ονομαστική της ισχύ καθώς και την διάρκεια ζωής της μόνωσης. Στόχος της παρούσας διπλωματικής εργασίας είναι η εύρεση της κατανομής της θερμοκρασίας στο εσωτερικό ενός ασύγχρονου τριφασικού κινητήρα στη μόνιμη κατάσταση λειτουργίας του με τη μέθοδο των συγκεντρωμένων παραμέτρων. Στο πρώτο κεφάλαιο αναφέρονται βασικές έννοιες της θερμοδυναμικής. Γίνεται αναφορά σε διάφορους συντελεστές, παρουσιάζονται οι θερμοδυναμικοί νόμοι και γίνεται σύντομη αναφορά στους μηχανισμούς μετάδοσης θερμότητας. Στο δεύτερο κεφάλαιο δίνεται η αναλυτική περιγραφή των μηχανισμών μετάδοσης θερμότητας και παρουσιάζεται ένα απλό δίκτυο μοντελοποίησης με ισοδύναμες θερμικές αντιστάσεις. Στο τρίτο κεφάλαιο παρουσιάζεται σύντομα η δομή, η αρχή λειτουργίας και οι τύποι μιας ασύγχρονης μηχανής. Εδώ επίσης αναφέρονται και οι διάφορες μορφές απωλειών ενέργειας κατά την λειτουργία μιας τριφασικής ασύγχρονης μηχανής. Παρουσιάζεται ακόμη ο υπό μελέτη κινητήρας και αναφέρονται τα θερμοστοιχεία που χρησιμοποιούνται στην πειραματική διαδικασία. Στο τέταρτο κεφάλαιο περιγράφεται η μέθοδος θερμικής ανάλυσης με χρήση ισοδυνάμου κυκλώματος θερμικών αντιστάσεων για την μόνιμη κατάσταση. Στη συνέχεια δίνεται το προτεινόμενο κύκλωμα και παρουσιάζονται αναλυτικά οι ισοδύναμες θερμικές αντιστάσεις του μοντέλου. Τέλος στο πέμπτο κεφάλαιο παρατίθενται τα αποτελέσματα της θερμικής ανάλυσης, γίνεται σύγκριση με τα πειραματικά δεδομένα θερμοκρασιακών τιμών που πάρθηκαν από τα θερμοστοιχεία και ακολουθεί η διαδικασία της παραμετροποίησης στους διάφορους συντελεστές που χρησιμοποιήθηκαν είτε υπολογίστηκαν κατά την ανάλυση. / Thermal analysis is an important design area and becoming more important part of the electric motor design process due to the push for reduced weights and costs and increased efficiency. Thermal analysis is of equal importance as the electromagnetic design of the machine, because the temperature rise of the machine eventually determines the maximum output power. The purpose of this study is to record the temperature distribution of the internal parts of an induction motor at steady state using an equivalent thermal circuit with lumped parameters. The first chapter is an introduction of the thermodynamic theory. The laws of thermodynamics are described and there is a brief report of heat transfer mechanisms. The second chapter describes analytically the heat transfer mechanisms. Also, an example of modelling using thermal equivalent resistances is given. The third chapter introduces shortly the operational principles of an induction machine. Here are also referred the various losses that occur during the rotation of an induction motor. The studied induction motor, with the modified stator winding to include thermocouples, is shown. The fourth chapter introduces the method of thermal analysis using thermal equivalent circuit with lumped parameters. The proposed model is given and its components are described in particular. At last, in the fifth chapter the results of temperature distribution are given and compared with experimental data of temperature values that are acquired using the thermocouples. Here also takes apart the parameterising of the various coefficients that were used or calculated during this study.

Made available in DSpace on 2014-12-17T14:55:37Z (GMT). No. of bitstreams: 1 SergioPS.pdf: 2376478 bytes, checksum: 7999af148ddd33a9739b28a9fdf05cf3 (MD5) Previous issue date: 2009-04-11 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / Equipment maintenance is the major cost factor in industrial plants, it is very important the development of fault predict techniques. Three-phase induction motors are key electrical equipments used in industrial applications mainly because presents low cost and large robustness, however, it isn t protected from other fault types such as shorted winding and broken bars. Several acquisition ways, processing and signal analysis are applied to improve its diagnosis. More efficient techniques use current sensors and its signature analysis. In this dissertation, starting of these sensors, it is to make signal analysis through Park s vector that provides a good visualization capability. Faults data acquisition is an arduous task; in this way, it is developed a methodology for data base construction. Park s transformer is applied into stationary reference for machine modeling of the machine s differential equations solution. Faults detection needs a detailed analysis of variables and its influences that becomes the diagnosis more complex. The tasks of pattern recognition allow that systems are automatically generated, based in patterns and data concepts, in the majority cases undetectable for specialists, helping decision tasks. Classifiers algorithms with diverse learning paradigms: k-Neighborhood, Neural Networks, Decision Trees and Na?ves Bayes are used to patterns recognition of machines faults. Multi-classifier systems are used to improve classification errors. It inspected the algorithms homogeneous: Bagging and Boosting and heterogeneous: Vote, Stacking and Stacking C. Results present the effectiveness of constructed model to faults modeling, such as the possibility of using multi-classifiers algorithm on faults classification / A manuten??o de equipamentos ? um dos principais fatores de custo no ambiente industrial, sendo de fundamental import?ncia o desenvolvimento de t?cnicas de preven??o de falhas. Os motores de indu??o trif?sicos s?o os equipamentos el?tricos mais utilizados na industria, pois apresentam um baixo custo e boa robustez, entretanto, n?o est?o imunes a diversos tipos de falhas como curto-circuitos nos enrolamentos e quebra de barras rot?ricas. Diversas formas de aquisi??o, processamento e an?lise dos sinais s?o aplicadas para melhorar seu diagn?stico. As t?cnicas mais eficazes utilizam sensores de corrente e a an?lise de sua assinatura. Neste trabalho, s?o apresentadas an?lises a partir destes sensores, sendo esta informa??o processada atrav?s do vetor de Park, que fornece uma boa capacidade de visualiza??o dos padr?es. Visando a obten??o destes padr?es fora do ambiente de opera??o, foi desenvolvida uma metodologia para a constru??o das bases de dados. Para a modelagem da m?quina tamb?m ? aplicada a transforma??o de Park no referencial estacion?rio para solucionar as equa??es diferenciais da m?quina. Detec??o de falhas requer uma an?lise profunda das vari?veis envolvidas e suas influ?ncias, tornando o diagn?stico complexo. Reconhecimento de padr?es permite que sistemas sejam gerados automaticamente, por encontrar padr?es e conceitos nos dados, muitas vezes n?o detectados por especialistas, auxiliando na tomada de decis?es. Algoritmos de classifica??o com diferentes paradigmas de aprendizado como k-vizinhos mais pr?ximos, Redes Neurais, ?rvores de Decis?o e Na?ve-Bayes s?o utilizados para reconhecer os padr?es dos motores. M?todos de multiclassifica??o s?o empregados para melhorar o desempenho na taxa de erro de classifica??o, s?o examinados os seguintes algoritmos homog?neos: Bagging e Boosting e heterog?neos: Vote, Stacking e Stacking C. Nos resultados ? poss?vel notar a efic?cia do modelo constru?do para simular as falhas assim como dos algoritmos multiclassificadores para a classifica??o de falhas

O presente trabalho apresenta a identificação de parâmetros de motor de indução linear tubular para a extração de petróleo, denominado de MATÆOS -- Motor Assíncrono Tubular para Aplicação na Extração de Óleo Subterrâneo. O primeiro passo para a identificação de parâmetros do motor foi a implementação da bancada de testes com o seu sistema de controle e supervisório. O modelo adotado corresponde à situação de operação em regime estacionário e consiste no circuito contendo resistores e indutores tradicionalmente utilizado para representar motores de indução. A identificação dos parâmetros foi realizada por meio de um procedimento de otimização tomando por base o balanço de potência do motor. Para isso, em função dos parâmetros do modelo, definiu-se e minimizou-se uma medida de erro entre as potências medidas e as calculadas pelo modelo para um conjunto pré-fixado de escorregamentos. A solução desse problema de otimização foi obtida por meio de uma busca exaustiva, uma vez que o número de variáveis independentes do problema é pequeno e uma região contendo o ponto de mínimo é conhecida com razoável confiança. Os resultados obtidos indicam que os valores dos parâmetros nominais a quente apresentam algumas diferenças em relação àqueles produzidos pelo método de identificação. Essas diferenças se devem possivelmente a variações nas características térmicas do motor e a erros introduzidos pelos instrumentos de medida. O modelo identificado apresentou uma menor disponibilidade de força e um pior rendimento quando comparado com os valores de projeto do modelo nominal a quente. Apesar disto, verificou-se uma boa proximidade entre a curva característica do modelo e as obtidas experimentalmente. / This work presents the identification of parameters of the tubular linear induction motor for the extraction of oil, called MATÆOS - which is the Portuguese acronym for Asynchronous Tubular Motor for Application to Subterranean Oil Extraction. The first step to the identification of the motor parameters was the implementation of the supervisory and control system for the test bench. The model, which is the usual one containing resistors and inductors, represents the motor operating in steady-state. The parameter identification problem was written as an optimization problem based on the motor power balance. An error between the measured power and the one given by the model was defined and minimized with respect to the model parameters in correspondence to a given set of slipping frequencies. The numerical solution to this problem was obtained by an exhaustive search algorithm. This choice was made because the number of independent variables is small and a region containing the solution is known with a significant degree of confidence. The results obtained show that the nominal values of the model parameters differ slightly from those produced by the identification method. Possibly, such differences are due to the variations in the model caused by temperature changes and to the measurement errors. The identified model presented both a lower available force and a worst efficiency when compared to the nominal design values for the warm motor. Despite this the characteristic curve of the model was quite close to the nominal design one.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-03-21T17:50:55Z No. of bitstreams: 1 thiagoresendedealmeida.pdf: 4974622 bytes, checksum: 0f8c86117d7625e1546498cd6c17f367 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-04-24T02:16:26Z (GMT) No. of bitstreams: 1 thiagoresendedealmeida.pdf: 4974622 bytes, checksum: 0f8c86117d7625e1546498cd6c17f367 (MD5) / Made available in DSpace on 2016-04-24T02:16:26Z (GMT). No. of bitstreams: 1 thiagoresendedealmeida.pdf: 4974622 bytes, checksum: 0f8c86117d7625e1546498cd6c17f367 (MD5) Previous issue date: 2013-10-30 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O principal objetivo deste trabalho é avaliar a influência da variação dos parâmetros da rede elétrica em função da frequência do sistema em estudos relacionados à avaliação da segurança de tensão de Sistemas Elétricos de Potência (SEP). Neste sentido, é proposta uma alteração na formulação denominada Fluxo de Potência com Regulação Primária. Tal formulação representa uma modelagem alternativa do problema de Fluxo de Potência, onde a frequência do sistema pode ser estimada com precisão satisfatória. É importante destacar que a metodologia proposta neste trabalho mantém a formulação full Newton do problema. A partir deste contexto são consideradas as variações dos parâmetros da rede, tais como shunts de barra, shunts de linha e o modelo de Linhas de Transmissão e Transformadores, tornando estes parâmetros dependentes com os desvios de frequência do sistema. É também realizada uma avaliação das cargas tipo motor de indução. Assim como os parâmetros da rede, os parâmetros do motor de indução são modificados de acordo com os desvios de frequência do sistema. Em seguida são desenvolvidos estudos de avaliação da segurança de tensão a partir do desenvolvimento e implementação de um Fluxo de Potência Continuado modificado e de uma formulação alternativa do problema de Fluxo de Potência Ótimo (FPO). A partir dessas ferramentas é possível avaliar a variação da frequência do sistema em função do aumento do seu carregamento. Também são feitas análises a partir dos autovalores da matriz Jacobiana do método proposto. Finalmente são utilizados cinco sistemas de pequeno e médio portes para a validação da metodologia proposta. Além de dois sistemas tutoriais de pequeno porte, é utilizado o sistema 16 Barras, que é um sistema criado a partir de dados reais e condições operativas encontradas no Sistema Interligado Nacional (SIN), os sistemas New England e IEEE-118 Barras, que são sistemas testes normalmente utilizados na literatura. O programa ANATEM, desenvolvido pelo CEPEL, é utilizado de forma a validar os desenvolvimentos realizados neste trabalho. / The main objective of this work is to evaluate the influence of varying the network parameters depending on the system frequency in studies related to the voltage security assessment of Electric Power Systems. In this sense, it is proposed a modification in the formulation of the called Governor Power Flow. Such formulation represents an alternative modeling of the power flow problem, where the system frequency can be estimated with satisfactory accuracy. It is important to highlight that in this work the proposed methodology retains full Newton formulation of the problem. From this context it is considered variations of the network parameters, such as shunt devices (capacitors and reactors) and equivalent models of transmission lines and transformers, making them dependent parameters with frequency system. It is also studied the induction motor modeling. As the network parameters, the parameters of the induction motor are modified according to the deviation of system frequency. Then voltage security assessment studies are carried out through the development and implementation of a Modified Continuation Power Flow and an Optimal Power Flow. Using these tools it is possible to evaluate the system frequency variation due to the increase of the system loading. Analyzes are also made using the eigenvalues of the Jacobian matrix associated with the proposed methodology. Finally six systems of small and medium scale are used in validation process of the proposed methodology. In addition to two tutorials small systems are used a 16 buses test system, which is created from real data and operating conditions found in the Interconnected Brazilian System, the well know New England system and the IEEE-118 buses test systems, which are systems commonly used in the literature. The ANATEM program, developed by CEPEL, is used to validate the developments made in this work.

The inverter dead-time is integral to the safety of a voltage source inverter (VSI). Dead-time is introduced between the complementary gating signals of the top and bottom switches in each VSI leg to prevent shoot-through fault. This thesis reports and investigates dead-time induced sub-harmonic oscillations in open-loop induction motor drives of different power levels, under light-load conditions. The thesis develops mathematical models that help understand and predict the oscillatory behaviour of such motor drives due to dead-time act. Models are also developed to study the impact of under-compensation and over-compensation of dead-time act on stability. The various models are validated through extensive simulations and experimental results. The thesis also proposes and validates active damping schemes for mitigation of such sub-harmonic oscillations. The thesis reports high-amplitude sub-harmonic oscillations in the stator current, torque and speed of a 100-kW open-loop induction motor drive in the laboratory, operating under no-load. Experimental studies, carried out on 22-kW, 11-kW, 7.5-kW and 3.7-kW open-loop induction motor drives, establish the prevalence of dead-time induced sub-harmonic oscillations in open-loop motor drives of different power levels. An experimental procedure is established for systematic study of this phenomenon in industrial drives. This procedure yields the operating region, if any, where the motor drive is oscillatory. As a first step towards understanding the oscillatory behaviour of the motor drive, a mathematical model of the VSI is derived in a synchronously revolving reference frame (SRF), incorporating the of dead-time on the inverter output voltage. This leads to a modified dynamic model of the inverter-fed induction motor in the SRF, inclusive of the dead-time act. While the rotor dynamic equations are already non-linear, dead-time is found to introduce nonlinearities in the stator dynamic equations as well. The nonlinearities in the modified dynamic model make even the steady solution non-trivial. Under steady conditions, the dead-time can be modelled as the drop across an equivalent resistance (Req0) in the stator circuit. A precise method to evaluate the equivalent resistance Req0 and a simple method to arrive at the steady solution are proposed and validated. For the purpose of stability analysis, a small-signal model of the drive is then derived by linearizing the non-linear dynamic equations of the motor drive, about a steady-state operating point. The proposed small-signal model shows that dead-time contributes to different values of equivalent resistances along the q-axis and d-axis and also to equivalent cross-coupling reactance’s that appear in series with the stator windings. Stability analysis performed using the proposed model brings out the region of oscillatory behaviour (or region of small-signal instability) of the 100-kW motor drive on the voltage versus frequency (V- f) plane, considering no-load. The oscillatory region predicted by the small-signal analysis is in good agreement with simulations and practical observations for the 100-kW motor drive. The small-signal analysis is also able to predict the region of oscillatory behaviour of an 11-kW motor drive, which is con consumed by simulations and experiments. The analysis also predicts the frequencies of sub-harmonic oscillations at different operating points quite well for both the drives. Having the validity of the small-signal analysis at different power levels, this analytical procedure is used to predict the regions of oscillatory behaviour of 2-pole, 4-pole, 6-pole and 8-pole induction motors rated 55 kW and 110 kW. The impact of dead-time on inverter output voltage has been studied widely in literature. This thesis studies the influence of dead-time on the inverter input current as well. Based on this study, the dynamic model of the inverter fed induction motor is extended to include the dc-link dynamics as well. Simulation results based on this extended model tally well with the experimentally measured dc-link voltage and stator current waveforms in the 100-kW drive. Dead-time compensation may be employed to mitigate the dead-time and oscillatory behaviour of the drive. However, accurate dead-time compensation is challenging to achieve due to various factors such as delays in gate drivers, device switching characteristics, etc. Effects of under-compensation and over-compensation of dead time are investigated in this thesis. Under-compensation is shown to result in the same kind of oscillatory behaviour as observed with dead-time, but the fundamental frequency range over which such oscillations occur is reduced. On the other hand, over-compensation of dead-time effect is shown to result in a different kind of oscillatory behaviour. These two types of oscillatory behaviour due to under- and over-compensation, respectively, are distinguished and demonstrated by analyses, simulations and experiments on the 100-kW drive. To mitigate the oscillatory behaviour of the drive, an active damping scheme is proposed. This scheme emulates the effect of an external inductor in series with the stator winding. A small-signal model is proposed for an induction motor drive with the proposed active damping scheme. Simulations and experiments on the 100-kW drive demonstrate effective mitigation of light-load instability with this active damping scheme. In the above inductance emulation scheme, the emulated inductance is seen by the sub-harmonic components, fundamental component as well as low-order harmonic components of the motor current. Since the emulated inductance is also seen by the fundamental component, there is a fundamental voltage drop across the emulated inductance, leading to reduced co-operation of the induction motor. Hence, an improved active damping scheme is proposed wherein the emulated inductance is seen only by the sub-harmonic and low-order harmonic components. This is achieved through appropriate altering in the synchronously revolving domain. The proposed improved active damping scheme is shown to mitigate the sub-harmonic oscillation effectively without any reduction in flux.

No description available.