Notice of Violation of IEEE Publication Principles

"Bearing Damage Detection of the Induction Motors using Current Analysis"

by Irahis Rodriguez and Roberto Alves
in the Proceedings of the IEEE/PES Transmission and Distribution Conference and Exposition,
2006

After careful and considered review of the content and authorship of this paper by a duly
constituted expert committee, this paper has been found to be in violation of IEEE’s Publication
Principles.

This paper contains substantial duplication of original text from the paper cited below. The
original text was copied without attribution (including appropriate references to the original
author(s) and/or paper title) and without permission.

Due to the nature of this violation, reasonable effort should be made to remove all past
references to this paper, and future references should be made to the following article:

"Models for Bearing Damage Detection in Induction Motors Using Stator Current
Monitoring"
by Martin Bloedt, Pierre Granjon, Bertrand Raison, and Gilles Rostating,
in the International Symposium on Industrial Electronics, 2004 (ISIE’04) Ajaccio, France

Bearing Damage Detection of the Induction
Motors using Current Analysis
Irahis Rodríguez, and Roberto Alves, Member, IEEE
Index Terms− Bearing faults, Current Signature Analysis,
Diagnostics, Motor, Frequency analysis
I. NOMENCLATURE
IM = Induction machine or induction motor
CSA = Current signature Analysis
MCSA = Motor current signature analysis.
MMF = Magnetomotive force
II. INTRODUCTION
A SYMETRIC operation of the Induction Machines
produces asymmetric flux, unbalance current, losses
increase and pulsating torque. Final results can be
efficiency reduction and excessive temperature increase, which
could lead to an early isolation failure on the machine. Thus
the detection of incipient failures is very important to increase
the time life of the electrical machines. IM can operate with
asymmetries, such as:
• Stator windings failures like, inter-turns short
circuit.
• Broken rotor bars and end ring faults.
This work was supported in part by Universidad de Carabobo, Bárbula
Edo. Carabobo, Venezuela.
I. Rodríguez is with Universidad de Carabobo, Naguanagua, Estado
Carabobo, irodrigu@uc.edu.ve, Valencia – Venezuela.
R. Alves is with Universidad Simón Bolívar, Sartanejas, Estado Miranda,
robert@usb.ve, Caracas - Venezuela
1-4244-0288-3/06/$20.00 ©2006 IEEE
1
• Mechanical failures, like bearing damage, motor
shaft failures or air gap eccentricities.
Actually bearings are one major cause of failures in rotating
machines [1]. Their detection is possible by monitoring noise,
vibration and temperature. The implementation of these
measuring systems could be expensive and probably only
proves to be economical and practical in the case of large
motors or critical applications [7]. During the past twenty
years, there has been a substantial amount of research into the
development of new condition monitoring techniques for
induction motors. One successful technique is MCSA [3].
A failure or not in the electric motor can be established by
detecting of some variations in input current (stator current),
torque and flux. Detection of input current is relative simple,
for this reason it was chosen, like diagnostic variable, in this
work.
A brief description of the power equipment and data
acquisition systems used is follow in the next section.
III. ANALYSIS CURRENT SPECTRUM CONSIDERATIONS
A. Bearing Fault Types
This paper considers rolling-element bearings with a
geometry shown in Fig. 1. The bearing consists mainly of the
outer and inner raceway, the balls and the cage which assures
equidistance between the balls.
Fig. 1. Geometry of a rolling-element bearing
The number of balls is defined as Nb. Usually damage
bearing is produced by dust and externals particles
contamination. When these particles go into bearing, they
damage the outer and inner raceway of the bearing.
Nevertheless, a method of distinguishing the different bearing
fault effects on the machine was to classify the faults according
 2

to the affected element: angular position of the minimal airgap length moves with
respect to the stator reference frame as the rotor turns at the
• Outer raceway defect
angular frequency wr (Fig. 3).
• Inner raceway defect
• Ball defect
B. Characteristic Frequencies
With each type of bearing fault, a characteristic frequency fc
can be associated. This frequency is equivalent to the
periodicity by which an anomaly appears due to the existence
of the fault. The characteristic frequencies are functions of the
bearing geometry and the mechanical rotor frequency fr. A Fig. 2. Radial rotor movement due to an outer bearing raceway defect
detailed calculation of these frequencies can de found in [2].
Their expressions for the three considered fault types are given Between two contacts with the defect, the defect itself has
by: moved by an angle described by:
Outer raceway:
ωr
⎛ D ⎞ ∴θi = ωr ∴ t = (5)
N fi
f o = b f r ⎜ 1 − b cos β ⎟ (1)
2 ⎝ Dc ⎠ Hence (5) becomes:
Inner raceway:
⎡ k =+∞
⎛ k ⎞⎤
g (θ , t ) = g o ⎢1 − ei ∑ cos (θ + k ∴θi ) δ ⎜ t − ⎟ ⎥ (6)
N b ⎛ Db ⎞ fi ⎠⎦
fi = f r ⎜1 + cos β ⎟ (2) ⎣ k =−∞ ⎝
2 ⎝ Dc ⎠
Where ei is the relative degree of eccentricity introduced by
Ball: the inner race defect.

Dc ⎛ Db ⎞
2
fb = f r ⎜1 − 2 cos 2 β ⎟ (3)
Db ⎝ Dc ⎠

IV. RADIAL ROTOR MOVEMENT

Bearing Fault produce variations in the air gap length, and
this variations produce changes in the magnetic field,
determining changes in induced voltage and finally in the
stator current. So, the first step in the theoretical analysis is the
determination of the air gap length g as a function of time t and Fig. 3. Radial rotor movement due to an inner bearing raceway defect.
angular position θ in the stator reference frame [4].
Then, airgap length becomes:
A. Airgap Length Variations
Outer Raceway Defect: When there is no contact between a ⎡ k =+∞
⎛ k ⎞⎤
ball and the defect, the rotor is perfectly centered. In this case
g (θ , t ) = g o ⎢1 − ei cos (θ + ωr t ) ∑ δ ⎜ t − ⎟ ⎥ (7)
⎣ k =−∞ ⎝ fi ⎠ ⎦
the airgap length g is supposed to take the constant value go,
neglecting rotor and stator slotting effects. On the contrary Ball Defect: In this case, the defect location moves in a
every t = k/fo (with k integer), the contact between a ball and similar as the inner raceway fault.
the defect leads to a small movement of the rotor center in the
stator reference frame (see Fig.2). In this case the airgap length 1 ⎛ Db ⎞
wcage wr ⎜⎜1 − cos β ⎟⎟ (8)
can be approximated by go(1-eo cosθ), where eo is the relative 2 ⎝ Dc ⎠
degree of eccentricity. These considerations lead to the
following expression for the airgap length: By analogy with (7), the expression of airgap length in
presence of a ball defect becomes:
⎡ k = +∞
⎛ k ⎞⎤
g o (θ , t ) = g o ⎢1 − eo cos θ ∑ δ ⎜⎜ t − ⎟⎟ ⎥ (4) ⎡ k = +∞
⎛ k ⎞⎤
⎣⎢ k = −∞ ⎝ f o ⎠ ⎥⎦ g b (θ , t ) = g o ⎢1 − eb cos(θ + wcage t ) ∑ δ ⎜⎜ t − ⎟⎟ ⎥ (9)
⎣⎢ k = −∞ ⎝ f b ⎠ ⎥⎦
Inner Raceway Defect. In this case, the situation is slightly
different fron the outer race defect. The fault occurs at the Where eb is the relative degree of eccentricity introduced by
instants t = k/fi. as the defect is located on the inner race, the the ball defect.
 3

Generalization: In order to simplify the following B (θ , t ) = Bo cos ( pθ − ωs t + ϕ )

considerations, (4), (7) and (9) can be combined in a ∞ (17)
generalized expression for the airgap length g in presence of a + B1 ∑ cos ⎡⎣( p + 1) θ ± kωc t − kωs t + ϕ ⎤⎦
bearing fault: k =0

⎡ Equation (17) clearly shows the influence of the rotor

k =+∞
⎛ k ⎞⎤
g (θ , t ) = g o ⎢1 − e cos (θ + ψ (t ) ) ∑ δ ⎜ t − ⎟ ⎥ (10) displacement caused by the bearing fault on the flux density:
⎣ k =−∞ ⎝ fc ⎠⎦ In addition to the fundamental sine wave (term Bo), a multitude
of fault-related sine waves appear in the airgap. These
Where fc is the characteristic bearing fault frequency given
supplementary waves have p ± 1 pole pairs and a frequency
by (1), (2) or (3), and ψ(t) is defined as follows:
content ffc given by
⎧ 0: for outer raceway defect
⎪ 1 ⎛ dψ ( t ) ⎞
ψ (t ) ⎨ ωr : for inner raceway defect (11) f fc = ⎜± ± kωc − ωs ⎟ (18)
2π ⎝ dt ⎠
⎪ ω : for fault in the squirrel cage
⎩ cage
D. Stator Current
B. Airgap Permeance The additional flux density components according to (17)
The airgap permeance Λ is proportional to the inverse of are equivalent to an additional magnetic flux Φ(θ,t). By
the airgap length g and is defined as follows: considering the realization of the winding and the geometry of
the machine, the additional flux in each stator phase can be
Λ=μ g (12)
obtained. With stator voltages imposed, the time varying flux
Where μ is the magnetic permeability of the airgap. In the causes additional components in the machine´s stator current
case of a bearing fault, the permeance becomes with (10): according to the stator voltage equation for the phase m:

μ / go dφm ( t )
Λ(θ , t ) = (13) Vm ( t ) = Rs I m ( t ) + (19)
⎡ +∞
⎛ k ⎞⎤ dt
⎢1 − e cos (θ +ψ ( t ) ) ∑ δ ⎜ t − ⎟ ⎥
⎣ k =−∞ ⎝ fc ⎠⎦ The frequency content of the flux in each phase is supposed
be equal to the frequency content of the airgap field according
In order to simplify this expression, the series of Dirac to (18). Under the hypothesis of imposed stator voltages, the
generalized functions is expressed as a complex Fourier series stator current in each phase is given by the derivative of the
development: corresponding flux. This leads to the following expression for
the stator current Im(t):
+∞
⎛ k ⎞ +∞

∑ δ ⎜t − f ⎟ = fc ∑ e
− j 2 kfc t
∞
k =−∞ ⎝ c ⎠ k =−∞
I m ( t ) = ∑ I k cos ⎡⎣ ±ψ ( t ) ± kωs t − ωs t + ϕm ⎤⎦ (20)
k =0
+∞
= fc + 2 f c ∑ cos ( 2π kf t )
k =−∞
c
(14) It becomes thus obvious, that the radial rotor movement due
to the bearing fault results in additional frequencies in the
stator current. For the three fault types, these frequencies are
So, the expression of the airgap permeance wave becomes:
obtained from (11) and (20):
μ
Λ (θ , t ) = 1 + e f c cos ⎣⎡θ + ψ ( t ) ⎦⎤ Outer race defect:
go (15)
+∞
ff pe = f s ± k fe k = 1, 2,3 (21)
+e f c ∑ cos ⎡⎣θ +ψ ( t ) ± kω t ⎤⎦
k =−∞
c
Inner race defect:

C. Airgap Flux Density ff pi = f s ± f r ± k fi k = 1, 2,3 (22)

The flux density in the airgap is determined by multiplying
the MMF with the permeance wave. The MMF Fr that is Ball defect:
obtained combining the rotor and stator fundamental is f f balls = f s ± f jaulas ± k f b k = 1, 2,3 (23)
assumed:

Fr (θ , t ) = F cos ( pθ − ωs t + ϕ )
Multiplication of (15) and (16) leads to thr expression of
the flux density distribution Br(θ, t):
(16)
V. MONITORING OF THE INPUT CURRENT
Instantaneous current monitoring implementation has been
shown in Fig. 4. It shows the induction motor, supply source,
load motor (DC generator), and the resistive load. To the left,
 4

the data acquisition system can be observed, that is basically a
computer with an acquisition card. The acquisition card
converts stator current and voltage analog signals to digital
quantities [5].
• Virtual Instrument 1: To Acquire and to save current
and velocity;
• Virtual Instrument 2: To read current and velocity;
• Virtual Instrument 3: To read and to compare
currents with healthy bearing and damage bearing.

Fig. 4. Power Equipment.

Fig. 6. Data Acquisition System

The circuit has been shown in Fig. 5, and the equipment

VI. EXPERIMENTAL RESULTS
characteristics are:
• M: IM, 2 HP, 220 V, 1680 rpm, Y, PF 0.76, 60 Hz; A. Outer Raceway Defect
• G: DC Generator 2.2 kW, 230 V, 1800 RPM; A test machine has been equipped with a faulty bearing
• Load: Resistance Box TB 40, 3.3 kW; carrying an outer raceway defect. The current spectrum for
motor with healthy bearing and the current spectrum for motor
• R: Resistance Measurement includes seven shunts
with damaged bearing are shown in Fig. 7 and Fig. 8.
resistances of 120 ohms, ¼ watts each;
HEALTHY BEARING
• PT: Voltage Transformer 200/10.
• CT: Current Transf. YEW, 10/1, 60 Hz, class 1;

Fig. 7. Current Spectrum for healthy bearing (The stator current is 96.87% of
the nominal current).
Fig. 5. Power Circuit
DAMAGED BEARING
Input current, whose frequency spectrum is going to be
used to detect bearing failure, was obtained by dividing the
voltage in the resistance box connected in Current Transformer
1 Ampere Side, into the value of the resistance shunt.
Data acquisition system has the following elements: Signals
conditioning equipment, Data acquisition card, and software to
drive the Data acquisition card. The acquisition system is
detailed in Fig. 6. The data acquisition target is AT-MIO-
16DE-10 of National Instruments, with sixteen single ended or
eight differential inputs. Fig. 8. Current Spectrum for damaged bearing, outer raceway defect (The
stator current is 96.87% of the nominal current).
The software LabVIEW was chosen to show the input
current. LabVIEW programs are called virtual instruments, or
In the case of damaged bearing, a peak appears for a value
VIs, because their appearance and operation imitate physical
of frequency determined by the (21) [6]. The frequencies of
instruments, such as Oscilloscopes and Multimeters. Lab
failure are:
VIEW contains a comprehensive set of tools for acquiring,
analyzing, and storing data. Three virtual instruments were • Failure frequency for healthy bearing
designed in this work: o FFPE = 160.8 Hz

• Failure frequency for damaged bearing with outer
raceway defect
o FFPE = 160.8 Hz
In the case of inner raceway defect, there isn´t peak for
failure frequency, and there is peak for failure frequency
(Fig.8). The failure frequencies are:
• Failure frequency for healthy bearing
o FFPE =184.15 Hz
• Failure frequency for damaged bearing with inner
raceway defect
o FFPE = 184.6 Hz
HEALTHY BEARING
DAMAGED BEARING
Fig. 9. Current Spectrum for healthy bearing and damaged bearing, inner
raceway defect (The stator current is 96.87% of the nominal current).
VII. CONCLUSIONS
• This paper has investigated the detection of rolling-
element bearing faults in induction machines by
monitoring the stator current. A fault model has been
analyzed which considers fault-related airgap length
variations.
• Complete expressions for the frequency content of
the stator current are obtained for the three types of
fault.
• The experimental study has demonstrated that
supervising these frequencies of failures it is possible
to improve the discovery of damaged bearings.
VIII. REFERENCES
[1] Monagas, M. and Mago, M., “Fallas más comunes en los motores de
inducción de empresas del sector industrial del Estado Carabobo”, Work
of Ascent of Category, Carabobo University, Bárbula, Venezuela,
January 2004.
5
[2] B. Li, M. Chow, Y. Tipsuwan, y J. Hung, “Neural-network base 1 motor
rolling bearing fault diagnosis”, IEEE Trans. Ind. Electron, vol. 47 no.
5, pp. 1060-1069, Oct. 2000.
[3] R. Schoen, T. Habetler, F. Kamran, y R Bartheld, “Motor bearing
damage detection using stator current monitoring”, IEEE Trans Ind.
Applicat., vol.31, no. 6, pp. 1274-1279, Nov./Di. 1995.
[4] Martin Blödt, Pierre Granjon, Bertrand Raison, Gilles Rostaing,
“Models for Bearing Damage Detection in Induction Motors using
stator current monitoring”, IEEE Trans Ind. Applicat.
[5] M. E. H. Benbouzid y G. B. Kliman, GAT stator current processing-
based technique to use for induction motor rotor faults diagnosis?, IEEE
Trans. Energy Conversion, vol. 18, no. 2, p.p. 238-244, June 2003.
[6] B. Yazici y G. B. Kliman, “An adaptative statistical time-frequency
method for detection of broken bars and bearing faults in motors using
stator current”, IEEE Trans Ind. Applicat., vol. 35, no. 2, pp. 442-452,
Mar./Abr. 1999
[7] S. Nandi y H. A. Toliyat, “Condition monitoring and fault diagnosis of
electrical machine – a review” Proc. IEEE_IAS Annual Meeting
Conference’99, vol. 1, Phoenix, AZ, Oct. 1999, pp. 197-204.
IX. BIOGRAPHIES
Irahis D. Rodríguez was born in in Valencia,
Carabobo State, Venezuela, on August 24, 1959.
She graduated from the Universidad de Carabobo.
She received the M.S. degree in electrical
engineering in 1993 from the Simon Bolivar
University, Sartenejas, Venezuela. From 1985 she
has been a faculty member at the University of
Carabobo, Bárbula, Venezuela, where she is now
Professor of Electric Machines and Power Systems.
She has authored or co-authored over publications
in the areas of power electronics and electric drives systems.
Roberto Alves was born in Artigas, Uruguay on
April 10, 1961 (robert@usb.ve -
http://prof.usb.ve/robert/). He received the B.Sc.
degree in Electrical Engineering from Simon
Bolivar University (USB), Caracas, Venezuela, in
1986, the M.Sc. from the same institution in 1990
and the Ph.D from the Polytechnic University from
Catalonia (UPC) in 1997. At this time, He is Full
Professor Electrical Engineering of the USB, since
1987. He has directed several studies and
investigation projects with the industry and has participated for more than 20
years in studies of electrical engineering in the following areas: power
quality, harmonic in electric systems, generators, control of electrical
machines and energy conversion.