Mustapha ICEEDT07
Mustapha ICEEDT07
Mustapha ICEEDT07
net/publication/235999800
CITATIONS READS
2 550
3 authors:
SEE PROFILE
All content following this page was uploaded by Mustapha Messaoudi on 08 February 2017.
Abstract--This paper deals with the diagnosis of (EPVA), which allows the detection of inter-turn short
induction motors (IM) with the so-called motor circuits in the stator winding [13]. The EPVA is
current signature analysis (MCSA). The MCSA is one appropriate for the stator windings monitoring. Çalis and
of the efficient techniques for the detection and the Çakir used the 2.s.fs spectral component in the stator
localization of electrical and mechanical failures, in current zero crossing times (ZCT) spectrum as an index of
which faults become apparent by harmonic rotor bar faults [2]. However, the major deficiency for this
components around the supply frequency. This paper fault indicator, for low slip IM operating at no load
presents a summary of the most frequent faults and its condition it may then be difficult to read its value. In [5]
consequences on the stator current spectrum. A three- authors studied the diagnosis of IM by pattern recognition
phase IM model was used for simulation taking into methods. This method consists in extracting features from
account in one hand the normal healthy operation and the combination of the stator currents and voltages. Some
in the other hand the broken rotor bars, the voltage of these features could be irrelevant or redundant.
unbalance between phases of supply and the abnormal Therefore, the Sequential Backward Selection (SBS)
behaviour of load. The MCSA is used by many algorithm is applied to the complete set of features to
authors for faults detection in IM. But the major select the most relevant. Then they used the k-Nearest
contribution of this work is to prove the efficiency of Neighbors (kNN) rules to monitor the IM functioning
this diagnosis methodology to detect more than one states. This rule is applied with reject options in order to
fault simultaneously, in normal and abnormal avoid automatic classifications and diagnosis errors.
conditions. Experimental results validate the Didier et al., [4] employed the Fourier Transform of the
simulation work. stator current and they analyzed its phase. It is shown that
the basically calculated phase gives good results when the
Index Terms--Induction motors, Fault diagnosis, motor operates near its nominal load. For weak load, the
Motor current signature analysis. results obtained are not robust enough for the detection of
an incipient rotor fault. In [1] the authors used the
1. INTRODUCTION vibration monitoring methodology to detect incipient
failures in IM. Vibration monitoring system requires
Induction motors play an important role in industry for
storing of a large amount of data. Vibration is often
the rotating machine practice because of their hardiness
measured with multiple sensors mounted on different
low costs and quasi-absence of maintenance.
parts of the machine. The examination of data can be
Nevertheless, it arrives that this machine presents an
tedious and sensitive to errors. Also, fault related machine
electric or mechanical defect. The faults of these machines
vibration is usually corrupted with structural machine
are varied. However the most frequent are [3]-[10]:
vibration and noise from interfering machinery. To
opening or shorting of one or more of a stator phase
overcome these problems Pöyhönen et al., used the
winding, broken rotor bar or cracked rotor end rings,
Independent Component Analysis (ICA) to compress
static or dynamic air-gap irregularities, and bearing
measurements from several channels into a smaller
failures. In order to avoid such problems, these faults have
amount of channel combinations and to provide a robust
to be detected to prevent a major failure from occurring. It
and reliable fault diagnostics routine for a cage IM [7].
is well known that a motor failure may yield an
This paper is focused on the Motor Current Signature
unexpected interruption at the industrial plant, with
Analysis (MCSA) approach. This technique utilizes
consequences in costs, product quality, and safety. During
results of spectral analysis of the stator current (precisely,
the past twenty years, there has been a substantial amount
the supply current) of an IM to spot an existing or
of fundamental research into the creation of condition
incipient failure of the motor or the drive system. It is
monitoring and diagnostic techniques for IM drives.
claimed that MCSA monitoring is the most reliable
Different detection approaches proposed in the literature,
method of assessing the overall health of a rotor system
those based on the Extended Park’s Vector Approach
80 80 80
40 40 40
20 20 20
0 0 0
40 50 60 40 50 60 40 50 60
frequency (Hz) frequency (Hz) frequency (Hz)
Fig. 1. Stator current spectrum under 75% operational load: (a) healthy motor, (b) Rr = Rr + 20%Rr,
(c) Rr = Rr + 40%Rr.
100 100
48.81 Hz
51.19 Hz
Stator current (dB)
52.38 Hz
47.62 Hz
50 50
0 0
-50 -50
30 40 50 60 70 30 40 50 60 70
(a) frequency (Hz) (b) frequency (Hz)
100 100
54.85 Hz
45.15 Hz
47.7 Hz
52.3 Hz
Stator current (dB)
40.3 Hz
45.4 Hz
59.7 Hz
54.6 Hz
50 50
0 0
-50 -50
30 40 50 60 70 30 40 50 60 70
(c) frequency (Hz) (d) frequency (Hz)
Fig. 2. Stator current spectrum with Rr = Rr +40% Rr: (a) with zero load condition, (b) with 25% load condition,
(c) with 50% load condition, (d) with full load condition.
TABLE I
used as an efficient indicator of faults degree if the load
effect is taken into account. The frequencies of the first
Harmonic components of the stator current two sidebands around the fundamental frequency for an
IM operating with a fixed number of broken rotor bars are
Frequencies 1,2 * Rr 1,4 * Rr 48.81 Hz and 51.19 Hz under 25 % of rated load, 47.7 Hz
(1 - 6.s).fs 44,75 Hz and 52.3 Hz under 50 % of rated load and 45.15 Hz and
(1 - 4.s).fs 46,70 Hz 46,50 Hz 54.85 Hz under full load. This means that the motor slip
(1 - 2.s).fs 48,35 Hz 48,25 Hz varied proportionally with the load.
(1 + 2.s). fs 51,65 Hz 51,75 Hz
(1 + 4.s). fs 53,30 Hz 53,50 Hz B. Motor current signature analysis for broken rotor
(1 + 6.s). fs 55,25 Hz bars and unbalanced supply voltage detection *
For an IM with a constant number of broken rotor bars
operated under different levels of operational load, it is In order to simulate unbalanced supply faults, a V0 drop
clearly seen in Fig. 2 that the amplitudes of the sidebands in one phase condition was created. Hence, the stator
around the supply frequency increase as the load torque voltage supply equations were changed as follows:
rises. Thus, this indicator component is sensitive to the ⎛ 2π ⎞
load variation. Therefore, it is not exactly reflecting the Vas ( t ) = 2 (V − V0 ) cos (ωt ) ; Vbs ( t ) = 2 V cos ⎜ ωt − ⎟;
⎝ 3 ⎠
fault degree. But, the amplitude of the sidebands can be
⎛ 2π ⎞ supply voltages. Fig. 3 (b) and (c), clearly, show that
V cs ( t ) = 2 V cos ⎜ ω t + ⎟.
(6)
⎝ 3 ⎠ symptoms of both unbalanced stator voltage supply
In Fig. 3 (a), new frequencies (1+2k).fs: (150 Hz; 250 (1+2.k).fs and broken rotor bars (1±2.k.s).fs appear in the
Hz) emerge in the current spectrum, due to the unbalanced stator current spectrum.
100
fs
3fs
Stator current (dB)
50
0 5fs
-50
-100
0 50 100 150 200 250 300
(a) frequency (Hz)
100
fs
(1-2ks)fs (1+2ks)fs
Stator current (dB)
50 3fs
0 5fs
-50
-100
0 50 100 150 200 250 300
(b) frequency (Hz)
100
fs
(1-2ks)fs (1+2ks)fs
Stator current (dB)
50 3fs
0 5fs
-50
-100
0 50 100 150 200 250 300
(c) frequency (Hz)
Fig. 3. Stator current spectrum for 20V drop of one phase voltage: (a) healthy rotor; (b) broken rotor
bars Rr = Rr + 20%Rr; (c) broken rotor bars Rr = Rr + 40%Rr.
C. Motor current signature analysis for faults detection in Fig. 4 depicts the current spectrum of an IM operating
motor operating under variable load condition under three different operational conditions (described in
TABLE II):
In this section, to show the behavior of the motor
operating under variable load condition, constant motor TABLE II
load used in the model is replaced with sinusoidal
changing load as shown in (7). Motor operating conditions
The load amplitude is equal to the rated load and the - Symmetrical rotor - Asymmetrical rotor - Asymmetrical rotor
resistance. resistance. resistance.
load frequency is chosen as the assumption in section 2.3. - Variable load. - Variable load. - Variable load.
Current spectrum is given in Fig. 4. It is shown that the - Balanced supply - Balanced supply - Unbalanced supply
load variation produces sidebands around the supply voltage. voltage. voltage.
frequency equal to frequencies of fs ± k.fr components.
100
fs
fs-fr fs+fr
Stator current (dB)
50
fs+2fr
fs-2fr
0 fs+3fr
-50
-100
0 20 40 60 80 100 120 140 160 180 200
(a) frequency (Hz)
100
(1-2ks)fs fs (1+2ks)fs
fs-fr fs+fr
Stator current (dB)
50
fs+2fr
fs-2fr
0 fs+3fr
-50
-100
0 20 40 60 80 100 120 140 160 180 200
(b) frequency (Hz)
100
(1-2ks)fs fs (1+2ks)fs
fs-fr fs+fr 3fs
Stator current (dB)
50
fs+2fr fs+3fr
fs-2fr
0
-50
-100
0 20 40 60 80 100 120 140 160 180 200
(c) frequency (Hz)
Fig. 4. Stator current spectrum for motor operating under variable load condition: (a) healthy motor; (b)
motor with broken rotor bars (Rr = Rr +40% Rr); (c) motor with broken rotor bars (Rr = Rr +40% Rr) and
for 44 V drop in one phase of the supply voltage.
In Fig. 4 (a), only the frequencies due to load The simulation results, done with different kind of
variations appeared in the stator current spectrum. In Fig. faults and different degrees of severity, shows the
4 (b), with the frequencies components due to load effectiveness of the MCSA method to spot an existing or
variations, it appeared also sidebands around the incipient failure of the motor.
fundamental at frequencies (1±2.k.s).fs produced by the
An experimental setup is built around a dspace 1104
rotor asymmetry. In Fig. 4 (c), frequencies due to load
layout board (DSPtms320) to validate the simulation
variation, rotor asymmetry and unbalanced supply voltage
results. Fig. 5 illustrates the experiment results of the
appeared simultaneously in the stator current spectrum.
stator current spectrum when the IM is operated under
The major efficiency of the MCSA was its ability to different unbalanced supply voltages. It appears in the two
detect multiple kinds of faults at the same time without Fig. 5 (a) and (b), sidebands at frequencies (1+2k)fs. It is
requiring access to the motor. However, the main clear seen that the magnitudes of these sidebands rose
drawback of this approach, for low slip IM operating at no when the drop in the supply voltage is increased. The
load condition it may then be difficult to detect the increasing in the magnitude of frequencies created by the
frequencies (1±2.k.s).f due to its low amplitude. Thus, the unbalanced supply voltage is clearer at the supply
sensitivity of the technique depends on the external load. frequency third harmonic (150 Hz).
150
-50
0 50 100 150 200 250 300 350 400 450 500
(a) frequency (Hz)
150
Stator current (dB)
fs
100
3fs 5fs
7fs 9fs
50
-50
0 50 100 150 200 250 300 350 400 450 500
(b) frequency (Hz)
Fig. 5. Stator current spectrum for motor operating under unbalanced supply voltage: (a) Partial unbalanced (30 %
drop in one supply voltage phase); (b) Full unbalanced (cut of one supply voltage phase).
4. CONCLUSION in induction motors," Eng. Appl. of Artificial Intelligence, vol. 19,
pp. 169-177, 2006. [Online]. Available:
A review of the most frequent IM failures and its http://www.sciencedirect.com
detection via the MCSA is presented. Initially, the [6] G. Didier, and H. Razik "Sur la détection d’un défaut au rotor des
phenomenon of asymmetrical rotor is presented and all moteurs asynchrones," 3EI magazine, no. 27, pp. 53-62, 2001.
the characteristic frequencies occurring in the stator [7] S. Poyhonen, P. Jover and H. Yotyniemi, "Independent component
current due to cracked or broken IM bars are explained. analysis of vibrations for fault diagnosis of an induction motor," in
Similarly, the phenomena of unbalanced stator voltage Proc. 2003 IASTED Int. Conf. on Circuits, Signals, and Systems,
and load variation, in healthy motor and in the case of pp. 203-208.
broken rotor bars, are studied. Experiment and simulation [8] M. Fenger, M. Susnik, P. Laderoute and W. T. Thomson,
results showed the efficiency of the proposed method. It is "Development of a tool to detect faults in induction motors via
clearly seen that symptoms of both unbalanced stator current signature analysis," in Proc. Iris Rotating Machine Conf.,
Santa Monica, CA, Jun. 2003.
voltage supply (2.k.fs) and load variation effect (fs±k.fr)
appear, simultaneously, in the stator current spectrum with [9] M. Messaoudi, "Diagnostic des défauts du moteur et actionneur à
induction," Master ATI Dissertation, Dept. Elect. Aut., Nat. Eng.
the indicator components of broken rotor bars (1±2.k.s)fs.
Sch., Gabes, Tunisia, 2006.
More experimental results can be extracted from the
experimental setup to investigate the other studied cases. [10] H. Razik, "Le contenu spectral du courant absorbé par la machine
asynchrone en cas de défaillance, un état de l’art," 3EI magazine,
REFERENCES vol. 29, pp. 48-52, Jun. 2002.
[1] C. M. Riley, B. K. Lin, T. G. Habetler, and R. R. Schoen, "A [11] R. R. Schoen, "An unsupervised, on-line system for induction
method for sensorless on-line vibration monitoring of induction motor fault detection using stator current monitoring," IEEE Trans.
machines," IEEE Trans. Ind. Appli., vol. 34, no. 6, pp. 1240-1245, Ind. Appl., vol. 31, no. 6, pp. 1281-1286, 1995.
1998. [12] W. T. Thomson, "On-Line MCSA to Diagnose Shorted Turns in
[2] H. Çalis, and A. Çakir, "Rotor bar fault diagnosis in three phase Low Voltage Stator Windings of 3-Phase Induction Motors Prior to
induction motors by monitoring fluctuations of motor current zero Failure," in Proc. 2001 IEEE-IEMDC’01 Conf., pp. 1-8.
crossing instants," Elec. Power Sys. Research, pp. 1-8, 2006. [13] G. G. Acosta, C. J. Verucchi, and E. R. Gelso, "A current
[Online]. Available: http://www.sciencedirect.com monitoring system for diagnosing electrical failures in induction
[3] M. E. H. Benbouzid, "A Review of induction motors signature motors," Mechanical Systems and Signal Processing, vol. 20, pp.
analysis as a medium for faults detection," IEEE Trans. Ind. 953-965, 2006. [Online]. Available: http://www.sciencedirect.com
Electron., vol. 47, no. 5, pp. 984-993, 2000. [14] R. R. Schoen, T. G. Habetler, F. Kamran, and R. G. Barthled,
[4] G. Didier, E. Ternisien, O. Caspary, and H. Razik, "A new "Motor bearing damage detection using stator current monitoring,"
approach to detect broken rotor bars in induction machines by IEEE Trans. Ind. Appli., vol. 31, no. 6, pp 1274-1279, 1995.
current spectrum analysis," Mechanical Systems and Signal [15] D. R. Rankin, "The industrial application of phase current analysis
Processing, pp. 1-16, 2006. [Online]. Available: to detect rotor winding faults in squirrel cage induction motors,"
http://www.sciencedirect.com
Power Eng. J., vol. 23, no. 2, 1995.
[5] R. Casimir, E. Boutleux, G. Clerc, and A. Yahoui, "The use of
features selection and nearest neighbours rule for faults diagnostic