Citas bibligráficas
García, F., (2018). Clasificación de fallas en rodamientos de máquinas rotativas utilizando aprendizaje de máquinas [Tesis, Universidad Nacional de Ingeniería]. http://hdl.handle.net/20.500.14076/17625
García, F., Clasificación de fallas en rodamientos de máquinas rotativas utilizando aprendizaje de máquinas [Tesis]. : Universidad Nacional de Ingeniería; 2018. http://hdl.handle.net/20.500.14076/17625
@mastersthesis{renati/709977,
title = "Clasificación de fallas en rodamientos de máquinas rotativas utilizando aprendizaje de máquinas",
author = "García Mora, Félix Antonio",
publisher = "Universidad Nacional de Ingeniería",
year = "2018"
}
Rotary machines have become one of the key equipment in many sectors of the industry, energy systems and aerospace engineering. Due to the extreme application conditions to which the equipment is working and the increasing demand for reliability, availability and safety, we have as critical components of a rotating machine, the bearings, the gearbox, the rotor, between others, which are easily subject to failure, which can cause the machine to break down and produce different types of losses. Until now, there are a variety of condition-based maintenance techniques (CBM), which is why vibration analysis has been accepted as an important diagnostic tool, because vibration signals can be easily obtained and contain a wealth of information about the conditions of the machines. This thesis used several current scientific articles that served as a basis to obtain the extraction of characteristics of the faults, in order to analyze and choose the best adapted for the present study. From this extraction of characteristics, machine learning techniques were used to obtain the classification of the faults in rolling machine bearings. Machine learning (ML) was used, through five different classification techniques: Support vector machine (SVM), random forest (RF), gradient boosting (GB), extra trees (ET) and XGBoost (XGB), in order to present a comparative framework, and thus find the most efficient method. As a result, it was found that the XGBoost method outperforms the other classifiers in terms of recognition accuracy, especially with a small set of training data. In comparison with traditional methods, machine leaming techniques outperform traditional fault classification techniques. These results suggest that machine learning using XGBoost is a promising method and offers a great contribution to practica! engineering.
Este ítem está sujeto a una licencia Creative Commons Licencia Creative Commons