Condition monitoring involves taking measurements on an induction motor while it is operating in order to detect faults. For this purpose normally a single sensor type, for example current is used to detect broken rotor bar using fault frequency components only under the full-load condition or a limited number of load cases. The correlations among the different types of sensors and their ability to diagnose single and multiple faults over a wide range of loads have not been the focused in previous research. Furthermore, to detect different faults in machines using any fault frequency components, it is important to investigate the variability in its amplitude to other effects apart from fault severity and load. This area has also often been neglected in the literature on condition monitoring. The stator current and axial flux have been widely used as suitable sensors for detecting different faults i.e. broken rotor bar and eccentricity faults in motors. Apart from detecting the broken rotor bar faults in generalized form, the use of instantaneous power signal has often been neglected in the literature condition monitoring. This thesis aims to improve machine condition monitoring and includes accurate and reliable detection of single and multiple faults (faults in the presence of other faults) in induction machines over a wide range of loads of rated output by using current, flux and instantaneous power as the best diagnostic medium. The research presents the following specific tasks: A comprehensive real database from non–invasive sensor measurements, i.e. vibration measurements, axial flux, 3-phase voltage, 3-phase current and speed measurements of induction motor is obtained by using laboratory testing on a large set of identical motors with different single and multiple faults. Means for introducing these faults of varying severity have been developed for this study. The collected data from the studied machines has been analysed using a custom-written analysis programme to detect the severity of different faults in the machines. This helps to improve the accuracy and reliability in detecting of single and multiple faults in motors using fault frequency components from current, axial flux and instantaneous power spectra. This research emphasises the importance of instantaneous power as a medium of detecting different single and multiple faults in induction motor under varying load conditions. This enables the possibility of obtaining accurate and reliable diagnostic medium to detect different faults existing in machines, which is vital in providing a new direction for future studies into condition monitoring. Another feature of this report is to check the variability in healthy motors due to: test repeatability, difference between nominally identical motors, and differences between the phases of the same motor. This has been achieved by conducting extensive series of laboratory tests to examine fault frequency amplitudes versus fault severity, load, and other factors such as test repeatability and machine phases. The information about the variations in the amplitudes of the fault frequency components is used to check the accuracy and reliability of the experimental set-up, which is necessary for the practical application of the results to reliably detect the different faults in the machines reliably. Finally, this study also considers the detection of eccentricity faults using fault frequency amplitudes as a function of average eccentricity, instead of as a function of load under different levels of loading. This has not been reported in previous studies. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1298314 / Thesis (Ph.D.)-- University of Adelaide, School of Electrical and Electronic Engineering, 2008

Στην παρούσα διπλωματική εργασία ασχοληθήκαμε με το θέμα της μελέτης και διάγνωσης σφαλμάτων στάτη μιας ασύγχρονης μηχανής. Μοντελοποιήθηκε και αναλύθηκε η λειτουργία ενός ασύγχρονου τριφασικού κινητήρα κλωβού, σε υγιή κατάσταση αλλά και υπό συνθήκες σφάλματος. Πιο συγκεκριμένα, με τη χρήση του προγράμματος πεπερασμένων στοιχείων σε δυο διαστάσεις εξομοιώθηκαν πιθανά σφάλματα στο στάτη της ασύγχρονης μηχανής. Αναλυτικότερα, στο πρώτο κεφάλαιο παρουσιάζονται κάποιες βασικές αρχές λειτουργίας των ασύγχρονων μηχανών, μελετώνται τα κυριότερα σφάλματα που συναντώνται σε αυτές τις μηχανές και παρουσιάζονται οι γνωστότερες διαγνωστικές μέθοδοι σφαλμάτων που υπάρχουν στη βιβλιογραφία. Κυρίως αναλύεται η μέθοδος διάγνωσης σφαλμάτων που χρησιμοποιεί την ανάλυση στο πεδίο της συχνότητας του ρεύματος στάτη (MCSA) η οποία και χρησιμοποιείται τελικά για τη διάγνωση των συγκεκριμένων σφαλμάτων. Ειδικότερα, μελετήθηκαν δυο ειδών σφάλματα στάτη, τα οποία είναι πιθανόν να συμβούν σε αυτήν τη μηχανή, το βραχυκύκλωμα κάποιων σπειρών εντός της ίδιας αυλάκωσης μιας φάσης και το βραχυκύκλωμα σπειρών σε δυο γειτονικές αυλακώσεις της ίδιας πάλι φάσης του στάτη τα οποία και μελετώνται αναλυτικά παρακάτω. Στο επόμενο κεφάλαιο εξομοιώθηκαν τα δύο μοντέλα με σφάλμα, όπως προαναφέρθηκε με χρήση του προγράμματος πεπερασμένων στοιχείων κάνοντας όλες τις απαραίτητες τροποποιήσεις στις παραμέτρους των αρχικών μοντέλων ώστε να απεικονιστούν στα μοντέλα τα σφάλματα που επιθυμούμε. Στο τρίτο κεφάλαιο παρουσιάζονται όλες οι γραφικές παραστάσεις των ηλεκτρομαγνητικών μεγεθών για όλες τις περιπτώσεις όπως αυτές προέκυψαν από την ανάλυση με το πρόγραμμα της Opera -2d. Συγκεκριμένα, τα μοντέλα αναλύθηκαν με έναν επιλυτή που λαμβάνει υπ’ όψιν και την κίνηση του δρομέα (RM solver), όπου ενώ η μηχανή στρέφεται με σταθερό αριθμό στροφών στην ονομαστική κατάσταση βλέπουμε τη συμπεριφορά της με και χωρίς τα σφάλματα. Στο τελευταίο κεφάλαιο χρησιμοποιούνται οι διαγνωστικές μέθοδοι για την ανίχνευση των σφαλμάτων και παρατίθενται τα αποτελέσματά τους. Πιο συγκεκριμένα, αναλύθηκε το φάσμα συχνοτήτων των κυματομορφών του ρεύματος στάτη και της ηλεκτρομαγνητικής ροπής (FFΤ), ώστε να εντοπιστούν οι συχνότητες στις οποίες εισάγονται αρμονικές λόγω του σφάλματος. Επίσης μελετήθηκε η μορφή της χρονικής συνάρτησης του μαγνητικού πεδίου σε ορισμένα χαρακτηριστικά σημεία της μηχανής. Τέλος παρατίθενται τα συμπεράσματα , μαζί με την συνεισφορά της εργασίας καθώς και θέματα για διερεύνηση. / In this paper we dealt with the topic of the study and diagnosis of errors stator asynchronous machine. Modeled and analyzed the function of a three-phase asynchronous motor cage, but in a healthy state and under fault conditions.

Os motores de indução desempenham um importante papel na indústria, fato este que destaca a importância do correto diagnóstico e classificação de falhas ainda em fase inicial de sua evolução, possibilitando aumento na produtividade e, principalmente, eliminando graves danos aos processos e às máquinas. Assim, a proposta desta tese consiste em apresentar um multiclassificador inteligente para o diagnóstico de motor sem defeitos, falhas de curto-circuito nos enrolamentos do estator, falhas de rotor e falhas de rolamentos em motores de indução trifásicos acionados por diferentes modelos de inversores de frequência por meio da análise das amplitudes dos sinais de corrente de estator no domínio do tempo. Para avaliar a precisão de classificação frente aos diversos níveis de severidade das falhas, foram comparados os desempenhos de quatro técnicas distintas de aprendizado de máquina; a saber: (i) Rede Fuzzy Artmap, (ii) Rede Perceptron Multicamadas, (iii) Máquina de Vetores de Suporte e (iv) k-Vizinhos-Próximos. Resultados experimentais obtidos a partir de 13.574 ensaios experimentais são apresentados para validar o estudo considerando uma ampla faixa de frequências de operação, bem como regimes de conjugado de carga em 5 motores diferentes. / Induction motors play an important role in the industry, a fact that highlights the importance of correct diagnosis and classification of faults on these machines still in early stages of their evolution, allowing increase in productivity and mainly, eliminating major damage to the processes and machines. Thus, the purpose of this thesis is to present an intelligent multi-classifier for the diagnoses of healthy motor, short-circuit faults in the stator windings, rotor broken bars and bearing faults in induction motors operating with different models of frequency inverters by analyzing the amplitude of the stator current signal in the time domain. To assess the classification accuracy across the various levels of faults severity, the performances of four different learning machine techniques were compared; namely: (i) Fuzzy ARTMAP network, (ii) Multilayer Perceptron Network, (iii) Support Vector Machine and (iv) k-Nearest-Neighbor. Experimental results obtained from 13.574 experimental tests are presented to validate the study considering a wide range of operating frequencies and also load conditions using 5 different motors.

Cette thèse s’inscrit dans le cadre des études des phénomènes électromagnétiques dansles régions frontales des grands turbo-générateurs. L’objectif de la thèse est d’estimer apriori le champ magnétique axial en fonction du point de fonctionnement afin d’éviterles possibles dégradations du circuit magnétique (dus aux points chauds et tensions entretôles, qui sont liés à la composante axiale du champ).Une maquette à échelle réelle a été spécialement conçue et réalisée dans le but d’améliorerla compréhension physique des phénomènes : les pertes, la pénétration du champ magnétiqueet les tensions entre tôles sont analysés.Les simulations par éléments finis sont ensuite utilisées : les avantages et inconvénientsseront discutés, ainsi qu’une comparaison critique des résultats par rapport aux mesuresexpérimentales sur la maquette. La région frontale d’un turbo-alternateur est aussi entièrementmodélisée.Enfin, un modèle simple du flux axial est développé. Ses coefficients sont déterminés àl’aide de simulations par éléments finis, mais il peut ensuite être utilisé en temps réel afind’estimer le flux axial correspondant à un point de fonctionnement quelconque. / This work aims to improve the knowledge of electromagnetic phenomena that occurin the end region of large turbo-generators. The goal of this work is to evaluate theaxial magnetic flux density with regard to the operating conditions (such as active orreactive power) in order to prevent potential deterioration of the stator. Indeed, the axialmagnetic field is known to induce hot points or voltages between laminations that maycause insulation breakdown and thus stator faults.An experimental apparatus in real scale has been designed and built. Its purpose is tostudy precisely the following phenomena: losses, axial magnetic flux density penetration,voltage across adjacent voltages.Finite element simulations (FEM) are also used: their advantages and drawbacks arediscussed, and the results are compared with the experimental measures. The wholeend-region of a turbo-generator is also simulated.Finally, a simple model of the axial magnetic flux is proposed. Its parameters are basedon the results of the FEM model, but it may be used in real time to evaluate the axialmagnetic flux density of any operating point.

Os motores de indução desempenham um importante papel na indústria, fato este que destaca a importância do correto diagnóstico e classificação de falhas ainda em fase inicial de sua evolução, possibilitando aumento na produtividade e, principalmente, eliminando graves danos aos processos e às máquinas. Assim, a proposta desta tese consiste em apresentar um multiclassificador inteligente para o diagnóstico de motor sem defeitos, falhas de curto-circuito nos enrolamentos do estator, falhas de rotor e falhas de rolamentos em motores de indução trifásicos acionados por diferentes modelos de inversores de frequência por meio da análise das amplitudes dos sinais de corrente de estator no domínio do tempo. Para avaliar a precisão de classificação frente aos diversos níveis de severidade das falhas, foram comparados os desempenhos de quatro técnicas distintas de aprendizado de máquina; a saber: (i) Rede Fuzzy Artmap, (ii) Rede Perceptron Multicamadas, (iii) Máquina de Vetores de Suporte e (iv) k-Vizinhos-Próximos. Resultados experimentais obtidos a partir de 13.574 ensaios experimentais são apresentados para validar o estudo considerando uma ampla faixa de frequências de operação, bem como regimes de conjugado de carga em 5 motores diferentes. / Induction motors play an important role in the industry, a fact that highlights the importance of correct diagnosis and classification of faults on these machines still in early stages of their evolution, allowing increase in productivity and mainly, eliminating major damage to the processes and machines. Thus, the purpose of this thesis is to present an intelligent multi-classifier for the diagnoses of healthy motor, short-circuit faults in the stator windings, rotor broken bars and bearing faults in induction motors operating with different models of frequency inverters by analyzing the amplitude of the stator current signal in the time domain. To assess the classification accuracy across the various levels of faults severity, the performances of four different learning machine techniques were compared; namely: (i) Fuzzy ARTMAP network, (ii) Multilayer Perceptron Network, (iii) Support Vector Machine and (iv) k-Nearest-Neighbor. Experimental results obtained from 13.574 experimental tests are presented to validate the study considering a wide range of operating frequencies and also load conditions using 5 different motors.