See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/262674501

Condition Monitoring of Electrical Machines

Article · January 2012

CITATIONS READS
9 2,746

2 authors:

Toomas Vaimann Ants Kallaste

Tallinn University of Technology Tallinn University of Technology
100 PUBLICATIONS   275 CITATIONS    89 PUBLICATIONS   231 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

MOBTP13 "Deployment of Digitalization in Operation and Condition Monitoring of Energy Conversion Systems" View project

Research and Development of Exterior-Rotor Permanent Magnet Synchronous Machines for Light Aerial Vehicles View project

All content following this page was uploaded by Toomas Vaimann on 28 May 2014.

The user has requested enhancement of the downloaded file.

 Condition Monitoring of Electrical Machines
Toomas Vaimann, Ants Kallaste
Tallinn University of Technology
[email protected], [email protected]

Abstract- All machines, no matter how well they are designed, and pharmaceutical industries, marine propulsion, metal and
calculated or produced have the tendency to fail at some point of mining industry, transportation, etc.
their existence. If the condition of the machines in use is not
monitored in any way and the maintenance is forgotten, the II. MAIN FAULTS IN ELECTRICAL MACHINES
resulting failures may pose a large economic and safety risk.
Even the small and at first not important faults can end in Electrical machines are critical components in many
catastrophic measures. This paper describes the main faults that commercially available equipment and industrial processes.
occur in electrical machines and the diagnostic methods that can Furthermore, they are often used in critical duty drives where
be used to detect these faults. sudden failures can cause safety risks and large economic
expenses. Different failures can occur in electrical machines,
some of which are listed below.
I. INTRODUCTION
A. Rotor Bar Faults
In nowadays world, the number of electrical machines and One of the most common rotor faults in induction motors is
their responsible tasks is continuously growing. This means the breaking of the rotor bars. Fig. 1. shows broken rotor bars
that their reliability has to be a priority for all the producers due to heavy duty operation.
and users of electrical machines.
There are a vast number of different faults that can pose
danger for the reliability of electrical machines. Most of those
faults start as a small declination from the normal operating
conditions. The problem with those faults is that if they are
not treated at an early stage, they can grow to catastrophic
measures. One possible solution for the early detection of
faults in electrical machines is condition monitoring.
Condition monitoring of electrical machines can be used
for different reasons. These include [1]:
1) Preventing catastrophic failures and significant
damage of the machines;
2) Avoiding loss of life, environmental harm and Fig. 1. Broken rotor bars due to heavy duty operation. [2]
economic losses; The main reasons for such faults is poor manufacturing,
3) Stopping unscheduled outages; such as defective casting and poor jointing. Another common
reason is over current e.g. due to jam condition of the
4) Optimization of machine performance; rotor [3], but there can be various reasons that will lead to
5) Reducing repair time and spare parts inventory; cracking or broken rotor bars [4]:

6) Lengthening of the maintenance cycle; 1) Thermal stress due to over-load, non-uniform heat
distribution, hot spot and arc;
7) Reducing price and raw material consumption;
2) Magnetic stresses due to electromagnetic forces, magnetic
8) Increasing product quality. asymmetry forces, noises and electromagnetic
In other words, the reasons can be safety (1, 2), production vibrations;
assurance (3, 4), predictive maintenance (5, 6) and quality 3) Residual stress from the fabrication process;
control (7, 8) [1].
4) Dynamic stress due to rotor axial torque and centrifugal
Possible features to be monitored in the diagnostic
forces;
purposes can include vibration, noise, heat, power
consumption, displacement, rotation speed, different electric 5) Circumferential stress due to wearing and pollution of
parameters (such as voltage, current, frequency, resistance, rotor material by chemical materials and humidity;
etc.). The list is long and growing. 6) Mechanical stress due to mechanical fatigue
This is also one of the reasons why condition monitoring of different parts, bearing damage, loosened laminations
can be used in very different industries. These industries etc.
include power generation, oil and gas industry, petrochemical
industry, pipelines, refineries, waste water treatment, food
 When a rotor bar is cracked or broken, resistance in this bar bearing at commissioning or following a repair, and its level
rises. This stops the current flow in broken bar and additional usually does not change over time [8]
current starts flowing in the bars next to the broken bars, Dynamic eccentricity is a condition where the center of the
which means that thermal stress in those bars rises and they rotor is not at the center of rotation, and the position of the
are likely to break when this fault is not dealt with. If such minimum radial air-gap rotates with the rotor [10]. This can
rotor is not fixed, the fault propagates until the whole rotor be produced by worn bearings, a bent shaft, asymmetric
cage is damaged. thermal expansion of the rotor, or by high level of static
eccentricity [8-9].
B. Bearing Faults
Eccentricity causes unbalanced magnetic pull which results
Damage of bearings is the most common cause of failures
in vibration, acoustic noise, bearing wear, and/or rotor
in squirrel-cage induction motors. Fig. 2 shows a typical
deflection. This increases the risk of stator-rotor rub, which
bearing fault of an induction motor. Two types of bearing
can cause serious damage in the motor, stator or rotor core,
faults are usually distinguished. The first are single point
and/or insulation. [10]
defects and the second is generalized roughness [5].
D. Stator Faults
Winding turn faults are one of the most common problems
that arise in stator. Fig. 3. shows such a fault. Other common
group of stator faults is the interlaminar short circuits.

Fig. 2. Flaking of the bearing. [3]

In case of single point defects, the characteristic spectral

components in vibration signal can be predicted for inner
ring, outer ring, rolling element and cage fault. These
frequencies can also appear in stator current around the Fig. 3. Short-circuited stator windings.
fundamental harmonic [6]. Although they are usually clearly Stator winding failure usually starts with a short circuit
visible using vibration analysis, in case of stator current it is between the adjacent turns in the stator windings [11].
difficult to observe them due to their low amplitude and noise Winding insulation damages may cause faults which in turn
disturbance [7]. produce high currents and winding overheating. This
In contrary to single point defects, generalized roughness overheating can quickly result in severe faults between
does not produce characteristic frequency, but rather specific windings of different phases or between winding and ground,
frequency bands. Therefore, methods for diagnosing producing then permanent and irreversible damages both in
generalized roughness problems are usually based on windings and stator core. [12]
removing the non-bearing faults components from diagnostic Early detection of such winding faults is needed to prevent
signal and utilizing the residuum or on seeking and utilizing serious damage for the motor. The use of variable-speed
the frequency bands with high probability of presence of drives increases these problems due to the high rates of
bearing faults components. Since the generalized roughness is voltage changes produced by inverter switching [13].
the most often type of bearing fault these methods are hardly
investigated [5]. III. DIAGNOSTIC METHODS FOR ELECTRICAL MACHINES

C. Air-gap Eccentricity In developed countries today there are more than 3 kW of

Air-gap eccentricity can be introduced due to electric motors per person and most of it is from induction
manufacturing imperfections or during operation and the motors [14]. During the past few decades there has been a
inherent level of static or dynamic eccentricity is typically continually increasing interest and investigation into fault
within 10% of the air-gap [8-9]. detection, condition monitoring and diagnosis of electrical
Static eccentricity is a condition where the position of the machines. As this interest has grown, the literature has also
minimum radial air-gap is fixed [10]. It can be caused by grown [15]. Many different techniques can be found for the
stator core ovality, or incorrect positioning of stator core or fault diagnostics and condition monitoring of induction
motors. An incomplete, but quite common list of most used The same transformation can be made using voltage instead
methods for induction motor rotor diagnostics via stator of current [21]. The usage of voltage can prove to be a betted
signals is brought below. method, because voltage is not influenced by torque. Both
open loop and closed loop motor diagnostics is possible with
A. Fast Fourier Transformation
using stator voltage for the reference pattern of this method.
Fourier analysis is very useful for many applications where
Figs. 4 and 5 show the stator voltage patterns of healthy and
the signals are stationary, as in diagnostic faults of electrical
faulty rotors
machines [16].
Its purpose is to monitor a single-phase stator current. This
is accomplished by removing the 50 Hz excitation component
through low-pass filtering and sampling the resulting signal.
Single-phase current is sensed by a current transformer and
sent to a 50 Hz notch filter where the fundamental component
is reduced. Analog signal is then amplified and low-pass
filtered. Filtering removes the undesirable high-frequency
components that produce aliasing of the sampled signal while
the amplification maximizes the use of the analog-to-digital
converter input range. Analog-to-digital converter samples
the filtered current signal at a predetermined sampling rate
that is an integer multiple of 50 Hz. This is continued over a
sampling period that is sufficient to achieve the required fast
Fourier transform. [17]
Fast Fourier transform is however not always the best
solution for condition monitoring of an electric motor. As the
properties vary with the time-varying normal operating
conditions of the motor, it can be difficult to differentiate Fig. 4. Stator voltage Clarke vector pattern of a healthy rotor. [21]
fault conditions from the normal operating conditions of the
motor using only fast Fourier analysis.
B. Wavelet Analysis
To overcome the problems, described in the last paragraph,
wavelet analysis can be used. Wavelet is a time frequency
analysis tool originated from seismic signal analysis, which
uses narrow windows for high frequency component [18].
Continuous wavelet transforms (CWTs) have constant
frequency to bandwidth ratio analysis and therefore, CWTs
provide powerful multi-resolution in time-frequency analysis
for characterizing the transitory features of non-stationary
signals [19]. Wavelet analysis can be used for localized
analysis in the time-frequency or time scale domain, which
makes it a powerful tool for condition monitoring and fault
diagnosis [16].
C. Park (Clarke) Transformation
Fig. 5. Stator voltage Clarke vector pattern of a faulty rotor. [21]
A two-dimensional representation can be used for describ-
ing three-phase induction motor phenomena. This transfor-
D. Finite Element Method
mation can be made using Park transformation (or Clarke
The finite element method, which is well established for
transformation)
induction motors modeling, could be used to provide an
iα = ia accurate evaluation of the magnetic field distribution inside
 , (1) the motor [17]. To assess the magnetic saturation effect on
iβ = 2 3 ( ib + ic ) faults detection, the time-stepping finite-element method is
where ia, ib, and ic are phase currents, iα and iβ are alpha and recommended [22].
beta components of the current. Faults in the rotor can be simulated using the finite element
Its representation is a circular pattern centered at the origin method and this would give an opportunity to estimate the
of the coordinates. This is a very simple reference figure, magnetic field anomalies, based on the motor simulation.
which allows the detection of an abnormal condition due to IV. CONCLUSION
any fault of the machine by observing the deviations of the
acquired picture from the reference pattern [20]. Using of different condition monitoring devices and
techniques is becoming increasingly important as the number
 of different machines performing very responsible tasks is [8] J. Cameron, W. Thomson and A. Dow, “Vibration and current
monitoring for detecting airgap eccentricity in large induction motors,”
continuously growing. Different faults, even the smallest Proceedings of IEE – Pt. B, May 1986, vol. 133, no. 3, pp. 155-163.
ones, in machines and drives have to be detected in a very [9] S. Nandi, S. Ahmed and H. Toliyat, “Detection of rotor slot and other
early stage. If those faults are not dealt with the earliest as eccentricity related harmonics in a three phase induction motor with
different rotor cages,” IEEE Transactions on Energy Conversion,
possible, propagation of the fault can lead to catastrophic September 2001, vol. 16, no. 3, pp. 253-260.
outcomes. [10] J. Hong, S. Lee, C. Kral and A. Haumer, “Detection of Airgap
In general it is possible to decide on the motor state using Eccentricity for Permanent Magnet Synchronous Motors based on the
d-axis Inductance,” Proceedings of the 8th IEEE Symposium on
the described condition monitoring and diagnostic methods. Diagnostics for Electrical Machines, Power Electronics and Drives,
Choosing of a certain method or a set of methods and also the September 5-8, 2011, Bologna (Italy), ID-96.
parameters, which will be used for diagnostics, depends [11] S. Cheng, Y. Zhang, S. Nawrocki, T. Habetler and M. Salman, “Using
Only the DC Current Information to Detect Stator Turn Faults in
highly on the properties and peculiarities of the machine and Automotive Claw-pole Alternators,” Proceedings of the 8th IEEE
the setup which has to be monitored. Symposium on Diagnostics for Electrical Machines, Power Electronics
Condition monitoring and diagnostic of electrical machines and Drives, September 5-8, 2011, Bologna (Italy), ID-139.
[12] G. Bossio, C. De Angelo, P. de la Barrera, J. Solsona, G. Garcıa and M.
plays a huge role in prolonging the lifetime of the machines. Valla, “Stator Winding Fault Detection in Induction Motor Drives
Using Signal Injection,” Proceedings of the 8th IEEE Symposium on
ACKNOWLEDGMENT Diagnostics for Electrical Machines, Power Electronics and Drives,
September 5-8, 2011, Bologna (Italy), ID-254.
This paper was supported by European Social Fund’s [13] A. Bonnett, “Analysis of the impact of pulse-width modulated inverter
Doctoral Studies and Internationalisation Programme DoRa. voltage waveforms on AC induction motors,” IEEE Transactions on
This research work has been supported by Estonian Industry Applications, 1996, vol. 32, no. 2, pp. 386–392.
[14] I. Boldea and S. Nasar, The Induction Machine Handbook. Boca Raton,
Archimedes Foundation (project „Doctoral School of Energy FL: CRC Press, 2002.
and Geotechnology II“). [15] M. Benbouzid, “Bibliography on Induction Motors Faults Detection
and Diagnosis,” IEEE Transactions on Energy Conversion, December
1999, vol. 14, no. 4, pp. 1065–1074.
REFERENCES [16] M. Benbouzid and G. Kliman, “What Stator Current Processing-Based
Technique to Use for Induction Motor Rotor Faults Diagnosis,” IEEE
[1] L. Menozzi, “Technologies for Machine Condition Monitoring,” Transactions on Energy Conversion, vol. 18, no. 2, June 2003, pp. 238–
SDEMPED 2011 Tutorials, 8th IEEE Symposium on Diagnostics for 244.
Electrical Machines, Power Electronics and Drives, September 5-8, [17] M. Benbouzid, “A Review of Induction Motors Signature Analysis as a
2011, Bologna (Italy), pp. 51–88. Medium for Faults Detection,” Proceedings of the 1998 International
[2] R. Fišer and H. Lavric, “On-line detection and diagnostics of induction Conference of the IEEE Industrial Electronics Society, Aachen
motor rotor faults using spectral analyses of stator current,” (Germany), vol. 4, pp. 1908–1913.
Proceedings of the 5th International Symposium on Topical Problems [18] Z. Cao, H. Chen, G. He and E. Ritchie, “Rotor Fault Diagnosis of
in the Field of Electrical and Power Engineering, Doctoral School of Induction Motor Based on Wavelet Reconstruction,” Electrical
Energy and Geotechnology, January 14-19, Kuressaare, Estonia, 2008, Machines and Systems, ICEMS 2001, Proceedings of the Fifth
pp. 7-11. International Conference, vol. 1, 2001, pp. 374-377.
[3] T. Lindh, On the condition monitoring of induction machines. [19] P. Tse, W. Yang and H. Tam, “Machine Fault Diagnosis through an
Lappeenranta: Lappeenranta University of Technology, 2003. 148 p. Effective Exact Wavelet Analysis,” Journal of Sound and Vibration,
[4] J. Faiz, B. Ebrahimi and M. Sharifian, “Time Stepping Finite Element vol. 277, 2004, pp. 1005-1024.
Analysis of Broken Bars Fault in a Three-Phase Squirrel-Cage [20] L. Szabo, E. Kovacs, F. Toth, G. Fekete and J. Dobai, “Rotor Faults
Induction Motor,” Progress in Electromagnetic Research, PIER 68, Detection Method for Squirrel Cage Induction Machines Based on the
2007, pp. 53–70. Park’s Vector Approach,” Annals of the University of Oradea, Fascicle
[5] O. Vitek, M. Janda, V. Hajek and P. Bauer, “Detection of Eccentricity of Electrotechnics, Section of Computer Science and Control Systems,
and Bearings Fault Using Stray Flux Monitoring,” Proceedings of the 2007.
8th IEEE Symposium on Diagnostics for Electrical Machines, Power [21] T. Vaimann, A. Kallaste and A. Kilk, “Sensorless Detection of
Electronics and Drives, September 5-8, 2011, Bologna (Italy), ID-218. Induction Motor Rotor Faults Using the Clarke Vector Approach,”
[6] R. Schoen, T. Habetler, F. Kamran and R. Bartfield, “Motor bearing Scientific Journal of Riga Technical University, vol. 29, 2011, pp. 43-
damage detection using stator current monitoring,” IEEE Transactions 48.
on Industry Applications, November/December 1995, vol. 31, no. 6, pp. [22] C. Kim, Y. Jung, S. Yoon and D. Im, “The Fault Diagnosis of Rotor
1274 – 1279. Bars in Squirrel Cage Induction Motors by Time Stepping Finite
[7] F. Immovilli, A. Bellini, R. Rubini and C. Tassoni, “Diagnosis of Element Method,” IEEE Transactions on Magnetics, vol. 33, March
Bearing Faults in Induction Machines by Vibration or Current Signals: 1997, pp. 2131-2134.
A Critical Comparison,” IEEE Transactions on Industry Applications,
July/August 2010, vol. 46, no. 4, pp. 1350 – 1359.

View publication stats