| dc.description.abstract | In practically all machines used inside a factory there is an electric motor, responsible for
moving components. As with other mechanisms, engines are subject to operating failures. With
the extensive use of these equipments to meet current manufacturing needs, it is essential that a
maintenance stop, even when scheduled and routine, is an event that is atypical of the day-today
routine of a large company or industry and, as a result, generates economic losses. With the
rise of Industry 4.0, the use of computational methods to predict and prevent unexpected stops
is becoming more and more abundant. Machine learning methods are being developed regularly
to meet the needs for systems that predict equipment failures. One of the main steps for
the development of such techniques is the learning itself. These learnings depend on a training
dataset, that must be as effective as possible to develop efficient and reliable predictive systems.
The objective of this work is to develop a computational model to simulate and extract data from
an electric motor under different operating conditions, in order to study which parameters extracted
from this are the most suitable for the development of an effective database. To achieve
this objective, a study of numerical models of electric motors with five degrees of freedom was
carried out, as well as a study of data statistics to have a better quantitative understanding of the
data extracted from the developed computational system. Simulations were carried out where
the engine was placed under different operating conditions, varied structural characteristics and
different types of data were extracted, and such data evaluated in a quantitative way. For this
experiment, the methodology used covered the open use programming language Python for the
application of numerical models, in addition to validation through bibliographic data for the
proposed model. In it, the result of one of the degrees of freedom developed in the proposed
model did not observe the sensitivity to the structural parameter, two parameters were more
sensitive than the other two, showing the method to be effective in the study and development
of a database, but not used in real cases, due to the different assumptions made in the study.
Keywords: Simulation. | |
