Igor Varejão
UFES, Espirito Santo, igor.varejao@edu.ufes.br .
Alexandre Rodrigues Loureiros
UFES,
Espirito Santo, arodrigues.ufes@gmail.com .
Thiago Olivera dos Santos
UFES,
Espirito Santo, todsantos@inf.ufes.br,
Flávio Varejão
UFES,
Espirito Santo, flavio.varejao@ufes.br.
ABSTRACT
Companies
in the industrial sector generally have large investments in modern production
equipment, as well as high maintenance costs for these units. Fast and accurate
detection of failures and problems in industrial equipment makes a crucial
contribution to reducing maintenance costs and improving confidence in
production. Fault diagnosis consists of monitoring the operation of equipment
in order to identify the occurrence of a failure. With the increase in the
number of sensors installed on board in equipment, they have been more used to
monitor the status of these equipment and diagnose their failures or
malfunctions. Advances in research in the area of Artificial Intelligence,
especially in the area of Machine Learning, provide ways to increase the reliability
of intelligent fault diagnosis systems and result in a more reliable
performance of equipment and industry. This article presents an overview of the
vibration data that has been used in several works in the last 30 years, points
out a common problem in the use of these data, presents what needs to be done
to solve it and how the academic community can contribute to this solution.
Keywords: Fault Diagnosis; Machine Learning; Vibrational Data Analysis; Similarity Bias.
Reviewed by Current Scientific Journal
on
janeiro 15, 2024
Rating:
