Instrumentation & Measurement Magazine 24-7 - 70

Table 1 - Methodology performance [5](BRB - Broken Rotor Bar, HBRB - Half Broken Rotor Bar)
Work (year)
D. Matié et al.
(2012)
E. Germen et al.
(2014)
A. Glowacz
(2014)
B. Bessam et al.
(2015)
Applied Methodology
1. Hilbert transform is used for feature extraction.
2. Support vector machine is used for feature
classification.
1. Wavelet transform is used for feature extraction.
2. Self-organizing maps method is used for features
classification
1. MUSIC method is used for feature extraction.
2. Bayes method is used for features classification
1. Wavelet and Hilbert transform are used for
feature extraction.
2. Linear discriminant method is used for feature
classification.
1. EMD is used for decomposing the measured
signal.
Valle-Novo et. al.
(2015) [10]
2. The zero crossing number and the time between
successive zero crossing of the decomposed signal
are used as features.
3. Gaussian probability density function is used for
feature classification.
R. Romero et al.
(2016)
CamarenaMartinez
et al.
(2016)
A. Naha et al.
(2016)
AmezquitaSanchez
et al.
(2017)
LizarragaMorales
et al.
(2017)
RangelMagdaleno
et
al. (2017)
Morales-Perez
et al. [5]
1. MUSIC method is used for feature extraction.
2. Variation between the two pseudospectra is used
for feature classification.
1. Wavelet transform is used for feature extraction.
2. Pearson Correlation method is used for feature
classification.
1. Extended Kalman Filter method is used to
estimate and track the fundamental component of
the measured signal.
2. MUSIC method is used for fault classification.
1. Fractal dimensión is used for feature extraction.
2. Fuzzy logic methods are used for features
classification.
1. Homogeneity analysis is used for feature
extraction.
2. Gaussian probability density function is used for
feature classification.
1. Hilbert-Huang method is used for feature
extraction.
2. Gaussian probability density function is used for
feature classification.
1. Fourier Transform and orthogonal matching
pursuit method are used for feature extraction.
2. Majority decision method is used for fault
classification.
Industrial Experimental Setup
Fig. 8c shows a real example in a stone crushing factory,
where experimental tests are carried out. Installing the failure
detection system in this industry is a challenge for the
70
50
10
HBRB and
1-BRB
HBRB, 1-BRB,
and 2-BRB
25
HBRB, 1-BRB,
and 2-BRB
50 and 100
3mm, 5mm,
7mm, and
1-BRB
0, 50 and 100
1mm to 1-BRB
100
25, 50, and 75
50
1-BRB
HBRB, 1-BRB,
and 2-BRB
100
95
50 and 75
HBRB and
1-BRB
98.2
10,50, and 100 1-BRB
0, 25, 50, and
100
50
10
3-BRB
1- and 2-BRB
1- and 2-BRB
100
100
100
100
Load (%)
Level of Fault
Effectiveness
(%)
95
99
99
93
instrumentation, due to the adverse conditions that are presented,
mainly the dust generated by the crushing of the
stones, outdoor conditions, the climate of the area, the amount
of vibration generated by the crusher, the current amplitude
IEEE Instrumentation & Measurement Magazine
October 2021
Instrumentation & Measurement Magazine 24-7

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 24-7
