With the increasing electrification of industrial and domestic processes and the need for efficient, reliable and environment friendly solutions Synchronous Reluctance Motor stands out as viable solution for all the requirements needed for a transition towards a sustainable society. SynRMs can be an alternative to both induction motors (having higher efficiency) and permanent magnets synchronous motors (with the enormous advantage of not needing rare-earth material which presents severe drawback from an environmental and geo-political point of view). A fundamental part of adopting a technology in all kinds of processes is ensuring continuous and reliable operation and this is accomplished by carrying out condition monitoring and fault diagnostics. Several techniques exist to carry out the just named operations. The focus of this thesis is on developing the model of a synchronous reluctance motor and to simulate different cases of healthy and faulty states of such motor and then to adapt and apply a machine learning model in order to carry out fault diagnostic and classification. The classification is done with a gradient boost classifier and the proposed approach is tested by current data.
Con la crescente elettrificazione dei processi industriali e domestici e la necessità di soluzioni efficienti, affidabili e rispettose dell'ambiente, il motore sincrono a riluttanza si pone come soluzione valida per soddisfare tutti i requisiti necessari per una transizione verso una società sostenibile. I SynRM possono essere un'alternativa sia ai motori a induzione (presentando una maggiore efficienza) sia ai motori sincroni a magneti permanenti (con l'enorme vantaggio di non aver bisogno di materiale di terre rare che presenta gravi svantaggi dal punto di vista ambientale e geopolitico). Una parte fondamentale dell'adozione di una tecnologia in tutti i tipi di processi è garantire un funzionamento continuo e affidabile, e questo si ottiene realizzando il monitoraggio delle condizioni e la diagnostica dei guasti. Esistono diverse tecniche per realizzare le operazioni appena citate. L'obiettivo di questa tesi è sviluppare il modello di un motore sincrono a riluttanza e simulare diversi casi di stati sani e guasti di tale motore, per poi adattare e applicare un modello di machine learning al fine di effettuare la diagnosi e la classificazione dei guasti. La classificazione è fatta con un classificatore gradient boost e l'approccio proposto è testato con dati relativi alle correnti assorbite dal motore.
Condition monitoring and fault diagnostic of a Synchronous Reluctance Motor using Artificial Intelligence techniques
MINCHIO, SEBASTIANO
2021/2022
Abstract
With the increasing electrification of industrial and domestic processes and the need for efficient, reliable and environment friendly solutions Synchronous Reluctance Motor stands out as viable solution for all the requirements needed for a transition towards a sustainable society. SynRMs can be an alternative to both induction motors (having higher efficiency) and permanent magnets synchronous motors (with the enormous advantage of not needing rare-earth material which presents severe drawback from an environmental and geo-political point of view). A fundamental part of adopting a technology in all kinds of processes is ensuring continuous and reliable operation and this is accomplished by carrying out condition monitoring and fault diagnostics. Several techniques exist to carry out the just named operations. The focus of this thesis is on developing the model of a synchronous reluctance motor and to simulate different cases of healthy and faulty states of such motor and then to adapt and apply a machine learning model in order to carry out fault diagnostic and classification. The classification is done with a gradient boost classifier and the proposed approach is tested by current data.File | Dimensione | Formato | |
---|---|---|---|
Minchio_Sebastiano.pdf
accesso riservato
Dimensione
3.9 MB
Formato
Adobe PDF
|
3.9 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/29020