Condition monitoring of motors

8/7/2019 Condition monitoring of motors

1/10

ON-LINE CONDITION MONITORING OF MOTORSUSING ELECTRICAL SIGNATURE ANALYSIS*

Aditya Korde B. E. (Elec.) - Diagnostic Technologies India Pvt. Ltd.

PREMISEMotor current acts as an excellent transducer for detecting faults in the motor.Spectrum analysis of the motors current & voltage signals can hencedetect various faults without disturbing its operation. Typical faultsdetectable by this are:

Rotor bar damage Misalignment/ unbalance Foundation looseness Static eccentricity

Dynamic eccentricity Core damage Loose wedges Interturn shorts Defective bearings

HISTORYBack in the early 70s, the US Nuclear RegulatoryCommission felt the need to check the condition ofmotors located inside the nuclear reactors usingnon-intrusive techniques. Research for discoveringa new technology was initiated by Oak RidgeNational Labs & subsequently licensed toFramatome ANP, which is the largest erector ofnuclear plants world-wide. It was found that themotor current signal was always modulated by any

fault condition inside the motor. Further (& continuing) research has led to newtechniques for conditioning the current & voltage signals in order to analyzethese signals & determine the nature of the fault.

INTRODUCTIONElectrical signature analysis is the procedure of acquiring the motor current &voltage signals, performing signal conditioning and analyzing the derived signalsto identify the various faults. The three phase signals are collected either directly(for a LT motor) or through a CT (for a HT motor). Thus, motors can be testedfrom the control panel, enabling easy testing of remote, inaccessible orhazardous area motors. A FFT (Fast Fourier Transform) analyzer is required forconverting the signals from the time domain to the frequency domain.

* Presented at the Recent Advances in Condition-Based Plant Maintenanceseminar organized by the Indian Institute of Plant Engineers on 17th & 18th May2002 at NITIE, Mumbai.


2/10

THEORYA motor current signal is ideally a perfect sinusoidal wave at 50 Hz. Pictorially;we can represent the current in terms of time as well as frequency (see Fig. 1).Here, the first picture shows the current vs. time while the second shows thecurrent vs. frequency.

The amplitude of the peak in frequency is equal to the RMS amplitude of the sinewave. As this is a theoretical situation with no harmonics, we see only one peakin the frequency spectrum. The conversion of the current from time to thefrequency domain is achieved using an algorithm called the Fast FourierTransform (FFT).

Fig. 1 A perfect 50 Hz signal in both time & frequency domains

During actual operation, many harmonics will be present in the motor signal, thusan actual signal will show many peaks including line frequency and its harmonics(see Fig. 2). This is known as the motors current signature. Analyzing theseharmonics after amplification and signal conditioning will enable identification ofthe various motor faults.

Voltage

Current

Hz

ig. 2 Typical low & high frequency spectra of a good motorF


3/10

C

Fig. 3 Typical spectra of a motor with severe rotor bar da

ertain harmonics come in on the supply & these are of little consequence.wever, harmonics are also generated due to various electrical & mechanical

ults. All faults cause a change in the internal flux distribution, thus generatingd cannot be

ency (PPF = motor slip xo. of poles) will show up as a sideband to the line frequency,

i. e. we will see peaks at FL PPF

health. Empirical

research has dB indicates an excellent rotor ar condition. (A dB scale is used for the Y-axis in order to resolve the PPF

ease in the amplitude of thePF peaks in the current spectrum.

the exact damage level can be assessed inach case (see Table 1).

oltage

urrent

zmage (Essar Steel)

Hofathe harmonics. Note that these are intermediate harmonics andetected by standard harmonic analyzers. As fault generated harmonics

appear only in the current spectrum (but not in voltage), superimposition ofcurrent & voltage spectra can easily identify them.

ROTOR BAR DAMAGEIn the current signature, the motor pole passing frequn

The difference in amplitude between the line frequency peak and the polepassing frequency sidebands is an indication of the rotor bar

shown that a difference of over 60bpeaks clearly. This is very difficult on a linear scale.)

As the rotor bars start degrading (i. e. high resistance joints are present or acrack starts developing), the rotor impedance rises. Due to this, the currentdrawn at the PPF frequency rises, leading to an incrPA difference of about 48 dB would indicate the presence of high resistance jointswhereas a difference of about 35 dB would indicate multiple broken bars. Mostcases lie somewhere in between ande

V

C

H


4/10

M

Fig. 5 RMS Demodulated spectrum of a misaligned motor (Indal, Hirakud)

ISALIGNMENT/ UNBALANCEo identify these problems, it is required to perform another signal conditioning of

motor current signal. The process is known as RMS Demodulation, which is

zFig. 4 RMS Demodulated spectrum of a properly aligned motor (Indal, Hirakud)

Here, the motor running speed will show up as a peak. Based on its amplitude,we can judge whether any misalignment or unbalance i resent. In the

de barely visible. In case the motor was to be misaligned or mechanically

z

Tthecarried out in order to eliminate the line frequency.

H

s pdemodulated signature of a healthy motor, the motor running speed peak woulbunbalanced, high peaks will show up at the motor running speed & its harmonics.

H


5/10

F

Fig. 7 Same motor after the foundation was tightened (RCF, Thal)

OUNDATION LOOSENESS

z

Fig. 6 RMS Demodulated spectrum of a motor with loose foundation (RCF, Thal)

Uneven foundations or loose foundation bolts will result in foundation looseness.his can by identified by looking at the RMS Demodulated spectrum. The

z

H

Tlooseness will show up as high peaks at half the running speed of the motor.

H


6/10

S

Fig. 8 Typical spectra of a motor with static eccentricity

Static eccentricity is the phenomenontypically caused due to soft foot in the foundation, cocked

TATIC ECCENTRICITY

oltage

z

(MSEB, Bhusaval)

of uneven stator-rotor air-gap,bearing or an

eakst the principal rotor bar passing frequency appearing as sidebands to the line

f the motorFL = line frequency

ine frequencyR CENTRE STATOR

Normal Static Dynamicmotor eccentricity eccentricity

V

Current

H

improperly adjusted air-gap for plain bearings. It will show up as high pafrequency and its harmonics.

Static eccentricity = RB x RS nFLwhere RB = no. of rotor bars

RS = running speed o

n = odd harmonics of the lROTO


7/10

Fig. 9 Typical spectra of a motor with dynamic eccentricity

STATOR MECHANICAL FAULTS*These faults are typically loose wedges, laminations or core damageLoose wedges damage the coil insulation mechanically aconducting varnish on the coil sides, leading to corona. The cthen start degrading the motor insulation.

Core damage results in shorted stator laminationscurrents and heating the core locally. Thisand destroy it over time.

DYNAMIC ECCENTRICITYDynamic eccentricity is the phenomenon of a variable stator-rotor air gap,typically caused due to worn out bearing housings or end covers. This is asignificant problem as it can damage the bearing & housing soon and caneventually lead to the rotor rubbing with the stator.

Close trending is recommended on detection of this condition. It will show up ashigh peaks at principal rotor bar passing frequency and its harmonics along withthe running speed sidebands around these.

Dynamic eccentricity = RB x RS nFL RS

Current

z

(Godrej, Ankleshwar)

.

n e therona discharges

, causing localized eddywill eat away at the motor insulation

Detection of these faults & distinguishing between them is proprietary

Voltage

H

d also erodo

*information and fault frequency calculations are hence not given over here.


8/10

Fig. 10 Typical spectra of a motor with core damage (Essar Steel)

INTERTURN SHORTS*Interturn shorts lead to excessive heating in the stator coil and also cimbalance. These will cause localized and uneven heateventually resulting in a ground fault. The current specshorts as well as interturn insulation degradation (in severe cases

Voltage

z

urrenting, reduced output and

trum can pick up interturn)

Current

H

.

Current

Hz

Fig. 11 Typical spectra of motor with degraded interturn insulation (IPCL,Baroda)

Voltage


9/10

BEARING PROBLEMS*

of unique defect frequencies (as specified by the

ce of high peaks at these(in terms

s r cage). The

T

Current

Hz

Fig. 12 Typical spectrum of a motor with a defective bearing (IPCL, Baroda)

All bearings have a setbearing manufacturer), which allow identification of bearing problems. When thecurrent signature of a motor is examined, the presenbearing defect frequencies can identify and pinpoint the problemof whether the damage is in the inner race, outer race, balls, rollerdegree of degradation can be assessed based on the amplitudes of these peaks.

ROTOR BAR DAMAGE SEVERITY LEVEL CHARSeverity

levelFL /FP

(dB)

FL / FP(Ratio)

FP / FL(Ratio %)

Rotor ConditionAssessment

1 >60 >1000


10/10

ADDITIONAL APPLICATIONSThe motor current is modulated by any form of vibration, which causes pulses inthe torque & results in harmonics. Hence, current spectrum analysis can also besed to detect problems in the driven loads. Typical detectable problems include:

Fan bla

lt n

Gea ft unbalance,

Load bearing problems, etc.

The chn is el r or assessing ion ofmo ized s ), o be

Variations in stem taper,

Wo r ea

Stem nut wear,

Degraded worm gear & valve stem lubrication,

Obs ns va r Motor pinion disengagement, etc

hile this paper has mostly covered faults in induction motors, the

ariable frequency drivesC motors, signature analysis can detect

aut , which has greatly simplified the technology. Thisexpert system is known as the EMPATH, which has been developed by

Fra now available in India through our organizationDia Pvt. Ltd.

ON

ill berev t aintenance in the new millennium.

u

de damage,

Be Gear tooth damage,

loose ess,

r sha

te ology extrem y popula in USA f the condit tor valve (MOVs in which f llowing problems can detected:

rm gea tooth w r,

tructio in the lve seat a ea,.

Wtechnology is equally applicable for generators, v(VFDs) & DC motors. For Dcommutator & armature faults, along with problems in the firing circuitry.

Most of the above analyses are now available commercially using anomated expert system

matome ANP, USA & isgnostic Technologies India

C CLUSIONThis is a highly versatile and proven technology for condition monitoring and faultanalysis of motors. It solves the biggest hurdle of any Plant Manager, which is toobtain a shutdown for testing his machines. We believe that it w

olu ionizing Condition Based M

Condition monitoring of motors

Documents

Transcript of Condition monitoring of motors