Resumo:
The importance of electric motors in industry is undeniable. Many of these machines play
a critical role in the processes in which they are integrated. Thus, an unplanned outage of
one of these devices can disrupt the operation of the entire system, leading to operational
risks and consequent financial losses. Among the causes of faults in three-phase induction
motors, we observe that between 26% and 36% are caused by insulation failures.
The literature suggests the Frequency Response Analysis (FRA) technique as a method
capable of gradually monitoring insulation aging, offering early diagnosis compared to traditional
monitoring techniques. However, current methodologies for online analysis - with
the machine in operation - besides being invasive, present some limitations, such as: lack
of repeatability in measurements, influence of parallel loads to the monitored asset, and
reduced spectrum analysis area. In this context, this work proposes an analysis methodology
to address current challenges highlighted in the literature. The study also proposes
a sweep frequency system through inductive coupling, ensuring a non-invasive analysis
system, a condition that simplifies its field installation and consequent market adoption.
The proposed system uses the Absolute Sum of Logarithmic Error (ASLE) statistical indicator
to monitor motor’s insulation condition. A laboratory aging test of a low-voltage
motor was carried out and the proposed system was used to monitor it, noting the ASLE
index evolution from 0.31 to 1.28, indicating the progression of dielectric degradation. In
contrast, the monitoring performed with a megometer showed no variations in the insulation
resistance parameter, indicating values between 300 𝐺W and 400 𝐺W. An aging test
of a medium-voltage motor was also conducted at the Laboratório de Alta Tensão (LAT)
of the Universidade Federal de Itajubá (UNIFEI). In this test, the proposed equipment
pointed out variations in the ASLE indicator from 0.09 to 0.17 in the first week, reaching
0.21 at the end of the first month, and 1.03 by the end of the test. The test was monitored
with various market equipment, among them, the partial discharge measurement system,
which indicated variations in the discharge rate only after 24 days of testing.
