polito.it
Politecnico di Torino (logo)

Predictive maintenance on a Permanent Magnet Synchronous Motor's battery

Paolo Pastore

Predictive maintenance on a Permanent Magnet Synchronous Motor's battery.

Rel. Tania Cerquitelli. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Matematica, 2021

[img]
Preview
PDF (Tesi_di_laurea) - Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives.

Download (4MB) | Preview
[img] Archive (ZIP) (Documenti_allegati) - Altro
Licenza: Creative Commons Attribution Non-commercial No Derivatives.

Download (9MB)
Abstract:

In this thesis, three principal machine learning algorithms are explored for Predictive Maintenance (PdM) purposes in the context of Electric Vehicle (EV), Support Vector Machine (SVM), XGBoost and Random Forest (RF). Another simple statistical model is performed, the linear model regression, whose results will be taken as starting point. Two kind of models are used, regression and classification algorithms, in order to better explain the continuous target variable (for the regression) and the discretized one for classification. The data on which this paper works, don't make available any maintenance information. So, an explorative advanced analysis is carried forward in order to extract from the data which records would have represented a fault in the in Permanent Magnet Synchronous Motor (PMSM). In greater details, the values of the voltage in the PMSM’s battery is analysed, by giving much attention to those values outside a certain interval. By looking at the accuracy, the model that predicts best the continuous values of the mean vector voltage, is the XGBoost, while for the classification problem the Random Forest seems to be the most accurate one, both by tuning the Hyperparameters and a selecting a subset of the entire set of features.

Relatori: Tania Cerquitelli
Anno accademico: 2020/21
Tipo di pubblicazione: Elettronica
Numero di pagine: 84
Soggetti:
Corso di laurea: Corso di laurea magistrale in Ingegneria Matematica
Classe di laurea: Nuovo ordinamento > Laurea magistrale > LM-44 - MODELLISTICA MATEMATICO-FISICA PER L'INGEGNERIA
Aziende collaboratrici: Zirak S.r.l.
URI: http://webthesis.biblio.polito.it/id/eprint/17343
Modifica (riservato agli operatori) Modifica (riservato agli operatori)