Motor Current Signature Analysis R K Gupta

Predictive Maintenance of Induction Motors
With MCSA & TSA Technology
R K Gupta
Principal Consulting Electrical Engineer (Power System)
Reliability Engineer
BE Electrical
Protection, Automation, PLC & SCADA, Level 3 MCSA, Colorado
USA

What is MCSA & TSA ?
MCSA: stand for Motor Current Signature Analysis. The basis of MCSA is that the
stator current contains current components which is directly linked to rotating
magnetic flux components caused by electrical or mechanicals faults. These
harmonics current components caused by fault can be used for early failure detection.
TORQUE SIGNATURE: The acquisition of torque ripple and the analysis of the torque
spectrum is a relatively new addition in the realm of condition based maintenance
technology. By capturing the real components of the current and deriving the flux
generated from voltage a signal equal to air gap torque is developed. Mechanical
information buried in a complex modulated current signal can now be easily accessed
and interpreted for electrical and mechanical conditions.

Amazing Fact
World’s 30% - 40% produced electrical energy is consumed by electrical
motors.

Dynamic Motor Testing
 A motor is only one part of a “Machine System” that includes: Power in, the Motor
and the Load.
 Dynamic testing diagnoses issues for all three links.

 Gives clarity to the Root Cause Analysis
 Power Quality, Motor, Load issues
 Helps to find
 Mechanical issues including: Bearing problems, Rotor Bar issues, Cavitation's
and Eccentricities
 Electrical issues including: Oscillations, Transients and Harmonics (Distortion)

Benefits of Dynamic testing
 Helps separate Mechanical from Electrical issues
 Provides “early warning” information
 Provides “support” for other technologies
 Can be performed more often than static testing
 It is safe, fast and non-intrusive

Defining Dynamic Electric Motor
Monitoring
 WHAT IS IT?: The ability to find Power Condition, Motor and Load Related issues
while the motor is running under normal conditions.
 HOW?: We measure, analyze and trend the currents and voltages of the motor-
load system (typically at the MCC).

Maintenance Program Types
 Corrective – Run to failure
 Preventative – Pre-planned periodic motor replacement or recondition regardless
of condition, run time or history.
 Predictive – A complete program that includes static and dynamic motor testing,
etc. All diagnosed and trended.

Cost Vs. Benefits
 Corrective $17 - $18 Per Hp
 Preventive $11 - $13 Per Hp
 Predictive $ 7 - $ 9 Per Hp
 1983 EPRI Study of Maintenance Programs

Rotor
8%
Bearing
44%
Other
22%
Stator
26%
Bearing
41%
Other
14% Rotor
9%
Stator
36%
Motor Failure Areas
IEEE (early 1990’s) EPRI (Mid 1990’s)

What are we really after?
 Reduce unscheduled downtime by predicting imminent motor failures and
identifying problem areas
 Determine root-cause of problem
 Ultimate goal: Save time & money

The Questions:
 Is the motor operating properly?
 Is the load operating properly?
 How is the motor reacting to the load?
 What is the condition of the power?

To Find the Answers:
 Observe voltages & currents supplying motor.
 Obtain motor speed.
 Obtain motor torque.
 Analyze data.
 Trend routinely.

Motor
MCC
Load
1. frequency 2. speed
3. Torque4. Power
5. Voltage
6. Current
Chain of Events

Here is how this thing works: The MCC tells the motor: ‘I am sending you 50Hz.’
Motor replies ‘Ok. Load, I would like to turn you at this speed ‘(1500rpm, say). Load
says: ‘Ok, you want to do that, then give me this much torque’. Motor responds ‘Fine’.
Turns around and says to the MCC: ‘I need this much power’ (remember, mechanical
output power of the motor is speed times torque. Both of those parameters are set
right now). MCC says ‘Well, I know that I am giving you this particular voltage here’.
Then the motor says ‘All right, granted that you offer me this voltage, then give me this
amount and type of current’.
In short: If you find anything wrong with the torque, it m ust be caused by the load. If
you find anything wrong with the voltage, it m ust be caused by the MCC. If you find
anything wrong with the current, then it is the play of MCC, motor and load together.
Action & Reaction

Note
 Torque related problems are caused by the load
 Power related problems come from “upstream” and effect the whole buss.
 If you find anything wrong with the current then it is the play of the MCC, motor and
load together.

Motor
MCC
Load
Breaker
Safety First!
NFPA 70
Yourcompany’s requirements
Explorer
Yourindustry’s requirements
“Best Practices”
Safety

VoltageProbesVoltageProbes
CurrentCTsCurrentCTs
Connection

VoltageProbesVoltageProbes
CurrentCTsCurrentCTs
Connection for LT/LV Motors

Motor
Load
CTs
Breaker
Explorer
PTs
 No physical connection to motor required.
 No transducers required to be installed on the motor
 Permanent installation available
 All values derived from voltage and current measurements only
Connection for HV/HT Motors

Motor
Load
CTs
Breaker
PTs
EP
Explorer First Energy
RC Pump
1 of 700+ EPs at
Entergy
Connection with EP

Speed – accuracy is critical

Accuracy of Speed Critical:
Consider a 4 pole motor:
Synchronous speed = 1800RPM
Full Load Speed = 1780 RPM
 1% Error in speed measurement = 18 RPM!
 Must be able to resolve shaft speed to the 1RPM level or better: 0.05% (0.05% =
1/1800*100%)
 Special signal processing techniques required: DFLL (Digital Frequency Locked
Loop) – basically a form of correlation.
 FFT will not work due to very long observation times (60 seconds or more) to get
enough frequency resolution to resolve 1 RPM.

Current Signature: FFT vs. DFLL
Amplitude: 20dB
Amplitude: 60dB
Resolution: 0.13Hz Resolution: 0.005Hz
FFT DFLL

Motor Speed
 Air gap changes as rotor rotates within the stator core because of excentricities,
misalignments of rotor and stator, out of round rotors, etc.
 Current in stator changes (slightly) as a result of the changing gap distance.
 Change in current shows up as a peak in the current spectrum.

Torque Derivation
 Use “dq0” theory – also called two axis theory.
 Theory exists since 1929
 R.H. Park, “Two-Reaction Theory of Synchronous Machines – Generalized Method of
Analysis. Part 1”, AIEEE Transactions, Vol. 48, July 1929, pp. 717-717.
 Used in VFD’s since the 1980s
 E.C. Lee, “Review of Variable Speed Drive Technology”, www.powertecmotors.com/avsde4.pdf.
 T.A. Lipo, A.B. Plunkett, “A Novel Approach to Induction Motor Transfer Functions,” IEEE
Transactions on Power Apparatus and Systems. Vol. PAS 93 pp. 1420-1419, 1979.
 A. B. Plunkett, “A Current Controlled PWN Transistor Inverter Drive,” IEEE/IAS 1979
Annual Meeting, pp 785-792.
 Well documented in motor control texts.
 P.C. Krause, O. Wasynczuk, S.D Sudhoff, “Analysis of Electric Machinery,” IEEE Press
NY, ISBN 0-7803-1101-9, 1995.

Torque Shows What the Load is
Doing
 Variations in torque indicates a problem
with the smooth operation of the motor
and or load.
 Intermittent load variations -> T(t)
 Repetitive load variations -> T(freq)
 To the right: Fan with a flapping belt
causing excessive bearing wear on both
motor and fan pillow blocks.

 Power Quality
 Voltage level, voltage unbalance, harmonics distortion, total distortion, power, harmonics
 Machine Performance
 Payback period, effective service factor, load, operating condition, efficiency
 Current
 Current level, current unbalance
 Spectrum
 Rotor bar, V/I spectra, demodulated spectra
 Connections
 Waveforms, ABC/SYM components, phasors
 VFD details
 Torque and speed vs. time, frequency and voltage vs. time.
Overall Check

Summary page – After each test

Power Quality
 Power level
 Voltage unbalance
 Harmonic distortion
 Total distortion
 Power
 Harmonics

Voltage Level
Percent
Under
Voltage
Present
test
Trending
Log

Low Voltage
High Voltage
Over Currents (Over Heat)
Low Power Factor
Iron Saturation
Ultimately Higher Losses
Incoming Power - Voltage Issues

Over/Under Voltage
 Motors are designed to operate with +/- 10% of rated voltage
 Ideally, voltage supply deviation should be less than +/- 2%
 When operating over/under voltage a motors performance, efficiency, and power
factor change

Over/Under Voltage
 Voltage deviations usually caused by
 Poorly performing or improperly adjusted transformers
 Undersized conductors
 Poor connections
 Low power factor sources in the distribution system

Acceptable System Voltage Ranges
Nominal System
Voltage
Allowable Limits % Allowable Voltage
Range
120V(L-N) +/- 5% 114V- 126V
240V(L-L) +/- 5% 228V- 252V
480V(L-L) +/- 5% 456V- 504V

Comparison of voltage level to average
winding temp & motor efficiency*
Voltage -10%
(414V)
Normal
(460V)
+10%
(506V)
HPFull
Load Temp Eff Temp Eff Temp Eff
10 66 90.0 56 91.4 55 91.5
20 84 90.4 70 91.8 67 92.1
50 84 91.9 69 93.1 62 93.6
100 82 94.2 72 94.8 69 94.9
200 90 94.9 77 95.5 74 95.7

6.6kV 1000 hp Pump Motor at a
Nuclear Power Plant
 Motor being overdriven. Operating at 7.7kV
 Plant engineering thought there should be no problem.
 Currents are below nameplate – so its ok, right?
 Wrong! Saturating the iron in the stator caused overheating
 Motor has burned up 3 times - $150,000 to repair.
 Thermal output of plant reduced to 50% during 3-4 weeks per repair.
 Lost $10,000,000 in electrical sales during each repair.

Voltage Unbalance
Percent
Of
Unbalance
Trending
Log

Voltage Unbalances
 When a voltage unbalance reaches 5 %, the phase currents can differ by as much
as 40 %.
Unbal = 100 Vmaxdev-V
V
Where:
Unbalance = Voltage unbalance in %
Vmaxdev = Line to line phase voltage deviating most
from mean of 3 phases
V = RMS voltage, mean line to line of 3 phases

Effects of Voltage Unbalance on
Motor Losses

Voltage Distortion
Percent of
Distortion
Trending
Log

Voltage distortion: Torque Spectra
NormalPeaksNormalPeaks
60Hz60Hz
120Hz120Hz
360Hz360Hz
Harmful distortion
Clear – no distortion

Voltage Distortion – Case study

Harmonics Can Damage Motors
 Shotton Paper (Flintshire, UK, Shaw Bowden, shaw.bowden@upm-kymmene.com)
 Large inverter fed motor on the same bus as a smaller induction motor.
 Harmonics from the VFD destroying smaller 660V, 550hp motor.
 Voltage THD = 11%, 9%, 8% (three phases)
 Voltage Imbalance = 3.6%
 NEMA de-rating = 60%
 Running at Effective Service Factor = 1.29
 End result, motor running very hot.
 Time to get out the Baker Advanced Winding Analyzer to check for insulation
damage.

Power
TotalKwUsed
PowerFactor
AverageVoltage
AverageCurrent
DistortioninVoltage
DistortioninCurrent
VoltageUnbalance
CurrentUnbalance

Machine performance
 Effective Service Factor
 Load
 Operating Condition
 Efficiency
 Payback Period

Test station 300 hp 3570 rpm
Eff. s.f. =
% NEMA derating
% Load

Effective Service Factor
“MotorHealthy”
Load92%
PowerQualityGood

Effective Service Factor
Dataindicatesthatthis
motorisprobablynot
overheating.
Loadisover100%but
powerqualityisgood
MotorOK

Load History
Percent
Load
Trending
Log

Temperature(C)
FullLoad 1.15% 1.25%
49 64 77
56 75 91
75 102 128
64 80 94
69 89 106
Horsepower
10
20
50
100
200
* Courtesy U S Motors
Motor Performance
Percent Load vs Temperature

Operating Condition
Trending
Log
Percent
Change

Operating Condition
 Goal: To compare machine’s operation with the past.
 Method:
 If past data exists at present torque & frequency level, then
Is the voltage level similar to previous?
Is the current level similar to previous?
Is the phase angle similar to previous?
 Successes
 Flagged very early rotor bar deterioration
 Flagged high resistance connection on terminal box
 Issue
 Undiagnosed flags DO happen

Efficiency vs Load Level
Percent
Load
Percent
Efficiency

Efficiency
PercentLoadvsLossesinKwPercentLoadvsLossesinKw PercentLoadvsEfficiencyPercentLoadvsEfficiency

Money Savings/Efficiency Matters!
Pin
Pout
=η
Input Power
Efficiency
Output Power = Shaft Power
Input Power = Electrical Power

Fundamental Relationships on
Efficiency:
Example: 100hp motor at 90% efficiency,
with perfect voltage condition.
a) What is the Input Power and the Output Power?
Efficiency = 100% - Losses[%]
Pout = 100hp
Pin = 100hp / 0.9 = 111hp
Efficiency = Output Power / Input Power
Ploss = 111hp – 100hp
Ploss = 11hp = rated losses
b) What are the losses?

Fundamental Relationships on
Efficiency:
Example: 100hp motor at 90% efficiency,
with 2% Voltage unbalance.
c) What is the Input Power and the Output
Power?
Pout = 100hp
Ploss = 11hp * 1.1 = 12.1hp
Pin = Pout + Ploss = 112.1hp

Variability Study of Identical
Motors
Motor HP No. of Test Efficiency
Max. Min. Ave.
Average
Losses
Variation
Losses (1)
200 10 96.6 96.0 96.3 5.722 +7.9%
75 6 93.9 92.8 93.2 4.06 +6.5%
40 7 91.0 90.0 90.6 3.085 +7.9%
(1) MaximumPercent increase in Total Loss fromthe Average Loss

Payback Period
HoursperDay DaysperWeek
CostperKwH

Payback Period
At97.3%Load
95.6%Efficiency
AnnualCostto
operatethismotorat
thisloadlevel
$92,927.84

Current
 Over current
 Current unbalance

Over Current
Percent
Of
Current
Trending
Log

Current Unbalance
Percent
Of
Unbalance
Trending
Log

Current Issues
 Over Current
 Increased heat causing faster degradation in the insulation system
 Current Unbalance
 Can be caused by voltage unbalance
 Can indicate internal motor problems – or
 Connection problems

Spectrum
 Rotor Bar
 V/I Spectrum
 Demodulated Spectrum
 Harmonics

Symptoms
 Effect of Broken Bars
 Requires constant torque level or it will
pulsate
 Next one breaks sooner
 Current increases
 Temperature increases
 Insulation life shortens
 Typically non-immediate death
 Symptoms of Broken/Degradation
Rotor Bars
 High Rotor Impedance
 Increase in Rotor Current
 Heat/Temperature

Signature of a
Good Rotor Bar
Known good rotor bar

Broken Rotor Bars
Vertical Line
IdentifiesRotor Bar
Frequency

Rotor Bar
MCSA detects current components induced in the stator winding due to rotating flux
components caused by problems such as broken rotor bars or high air gap eccentricity. It
is true that broken rotor bars will result in a change to the vibration spectrum, but vibration
is traditionally sensed at the bearings and for each motor there is a different mechanical
stiffness between the electromagnetic forces caused by broken bars or air gap eccentricity
and the position where the vibration is sensed. Consequently the vibration from
electromagnetic forces is a second order effect compared to current components directly
induced from the specific rotating flux waves. This is not the case with MCSA. With
respect to detecting air gap eccentricity problems a similar reasoning applies. It has also
been shown that the current spectrum changes due to mechanical misalignment and due
to abnormal rotor drive dynamics.

DETECTION OF BROKEN BARS
 Primary Causes:
 Direct on-line (DOL)
 The starting duty cycles which the rotor cage winding was not designed to withstand.
This causes high thermal and mechanical stresses since the starting currents are
typically of the order of five times the full-load operating current.
 Pulsating Mechanical Load
 Pulsating Load arise when driving reciprocating compressors, ESP, Agitating
Pumps etc which can subject the rotor cage to high mechanical stresses.

DETECTION OF BROKEN BARS
Broken rotor bars result in current components being induced in the stator winding at
frequencies given by Equation
fsb = f1(1 ± 2s) Hz

CURRENT SIGNATURE DUE TO
BROKEN BARS
LOWE SIDEBAND
f1 =SUPPLY FREQUENCY (50 Hz)
Hz
dB
UPPER SIDEBAND

SIGNATURE PATTERN DUE TO BROKEN BARS
TWICE SLIP FREQUENCY SIDEBANDS
fsb = f1±2sf1
-2sf1 +2sf1
f1
dB
Hz
IDEALISED Current Spectrum – Due to Broken Rotor Bars

Case Study
 Vedanta Aluminum Extraction Plant,
India
 Motor Details
 ID Fan
 V=6600
 I=40 A
 HP=475.87
 Speed=992

Air Gap Eccentricity
 Air Gap Eccentricity consists of two types :
 Static
 Dynamic
Both air gap exists simultaneously in a real SCIM due to manufacturing
tolerances & installation procedure ( in large motors).

Static eccentricity where the position of minimum radial air gap length is fixed in space.
Causes of this include:
Static Eccentricity
 Manufacturing tolerances on each component part.
 Stator core ovality.
 Out of tolerance spigots in the end frames that house the bearings.
 Incorrect installation of large HV motors.

 Dynamic eccentricity where the minimum air gap revolves with the rotor and is a
function of space and time.
 Causes include –
 Non-concentric rotor and shaft.
 Thermal bow of the rotor.
 Combination of high air gap eccentricity, consequential unbalanced
 magnetic pull (UMP), and a flexible rotor.
 Severe bearing wear.
 Rotor core lamination movement independent of the shaft.
Dynamic Eccentricity

Dynamic Air Gap Eccentricity
Minimum air gap revolves with the rotor and it is the function
of space & time.

Diagnosis of Air Gap
Motor Current Signature Analysis (MCSA) is used to identify the characteristic current
signature pattern which is indicative of abnormal levels of air gap eccentricity. The
current signature pattern is identified via the following equation
fec = f1(R(1-s)/p ± nws) ± f1(1-s)/p
Consists of two parts:
frs = normal rotor slot passing frequencies = f1(R(1-s)/p ±
nws)
fr = rotor speed frequency = f1(1-s)/p = Nr (rpm)/60

High Air Gap Eccentricity Normal Air Gap Eccentricity
Air Gap Eccentricity Graph

Case Study
Vedanta Aluminum,
India
Motor Details
Load : Pump
V : 6600
I : 83.17
Speed : 2991
HP : 1072.39
Presence of Dynamic eccentricity

V/I Spectrum
 Voltage
 Current
 Demodulated Spectra

Voltage Spectra
120Hz
180Hz
300Hz

Current Spectra
120Hz
180Hz 300Hz

Demodulated Spectrum
 Analyzes the dynamic behaviour of Motor Load System
 It is the demodulated spectrum of Torque Signature
 With Channel Control, the demodulated spectrum can be changed from torque to
current or voltage of any one phase

Torque Domain
 Torque Ripple Displays the instantaneous torque requirements of the load
 Torque Spectra
Two tests are available in torque domain:

Torque Ripple
Normal
butHigh
Rated
Torque
 Torque Ripple shows the ripple
vs rated torque

Torque Ripple
Cavitations Normal
AutoFactorySanitaryPump
Toronto,Canada

Case Study
Vedanta Aluminum, India
Motor Details
Function: Coal crusher , Gear
Attached
V=6600
I=34.70
Speed=981
HP=375.34

Torque Spectrum
Torque Spectrum Analysis helps to find issue related dynamic behaviour of motor
load system.
 Bearing Issues ( Motor DE, Load End)
 Analyze Load health (Fan, Gear, Pump, Impeller etc)
 Shorted Turns
 Cavitations, Unbalance, Misalignment, Foundation Looseness etc

Case Study Bearing Failure
 5hp, 4 pole, Baldor Motor
 24” Axial Fan
 Instrumented with
 Cognitive Systems CV395B Spectrum Analyzer
 Bentley Nevada ADRE 208P (Automated Diagnostics
for Rotating Equipment)
 SWANTECH stress wave analysis system
 Air flow meters
 Accelerometers
 Laser tachometers
 Current meters
 Thermocouples
 Baker Explorer

 SKF 6503 Deep Groove Bearing
 .062 deep, .062 wide grove in outer race.
 Vibration should show at the ball pass
frequency in formula below:(n=#balls,
f=rot freq, Bd=ball dia, Pd= path dia,
beta=contact angle) BPFO=107Hz:
)1(
2
)( βCos
Pd
Bd
f
n
BPFORaceOuter −=

3-D Demodulation
)1(
2
)( βCos
Pd
Bd
f
n
BPFORaceOuter −=

Vibration Spectrum
Known Fault
harm. * BPFO ± 2 * RPM

Known Good Bearing Known Outer Race Defect
Torque Spectrum
Electrical Frequencies Removed Electrical Frequencies RemovedMarking 1 * BPFOAdding Electrical Harmonic Sidebands
BPFO Controlled Lab Test
… how it works

BPFO
120Hz –
BPFO
240Hz –
BPFO
120Hz +
BPFO

VFD Details
 Torque and Speed vs. Time
 Frequency and Voltage vs. Time

VFD: Typical issues
 Voltage spikes
 Non-sinusoidal sine wave
 Shaft voltages
 Bearing failures
 Voltage distortion
 Volts/Hertz ratio

Variable Frequency
VFD:
Conveyor Drive.
60 hp, 1200 rpm
Frequency
Voltage
Torque
Speed

VFD: What is going on?
Conveyor Drive. 60 hp 1200 rpm
• Oscillating
• Flawed control loop design
• Over rated
• Breaking
• Generator

Transient Analysis
Captures voltage, current and torque vs time at startup

Voltage vs. time Torque vs. time
Current vs. time
Transient Analysis - Normal start-up

Transient Analysis-Case Study
 Using start-up transient analysis to adjust soft-starts and locate problem areas.
 In this case the customer had six new soft-starts installed; five were performing as
expected but one was not.
 Result: New soft-start had to be replaced.

Transient analysis: Common signatures

 3 Identical Vertical Pumps Feeding a Manifold
 4160V, 1250hp, 254rpm
 Pumps designed to pull 32,000Nm from motor.
 Motor designed to provide 35,000Nm for full load.
 One Pump pulling only 28,000Nm from motor.
Cavitations
Case Study

Findings
 Problem found in the pump – not the motor:
 First pump would only run at 28,000Nm or 75% of capacity
 Pump showed fluctuating torque
 The utility wanted higher cooling capacity for summer months - pump
pulled for repair
 Repair cost $180,000

Pump # 1 Typical - Pump # 2 & 3
Torque Signature

The Saga Continues
 The pump was repaired & put back on line.
 Power output of the entire 730MW plant was available…..until
 One of the other pumps sheared its shaft.

Second Pump Failed
 It has been speculated that the first pump’s turbulent water flow in the manifold overstressed
the shaft of the second pump. The pump may have have been destined for failure regardless
of the first pump’s performance.
 Torque ripple was higher on 2nd and 3rd pump when 1st pump was running at 75% with
loose end bell. Didn’t think too much of this at the time.
 Plant now down to 2/3rd of it’s thermal output because 1/3rd of cooling capacity was lost.
 Repair of the second pump took 5 weeks
 During the 5 weeks, the first pump was able to provide 100% of its rated pumping capacity.
Had it still been at 75% capacity, $4,000,000 in power sales would have been lost due to low
cooling capacity. (E. Wiedenbrug, “Instantaneous Torque as Predictive Maintenance Tool
for Variable Frequency Drives and Line Operated Motors,”)

Induced Imbalance
 7.6 grams introduced at 2” from center of fan – 0.54ounce-inches of
imbalance.

Imbalance Results
 Torque spectra showed a significant difference between the balanced and
unbalanced signal at the 1 x shaft speed. Torque signal very usable in finding
imbalances.
 Accelerometers mounted on the outside of the duct showed a 6.7% increase in
amplitude at 1x shaft speed. Vibration does change, but it would be tough to be
definitive.

Loose Fan Belt Case Study
Mechanical Problem

 Using “Torque Ripple” to locate mechanical issues.
 In this case, an air handler was having bearing failures and alignment problems.
 The torque ripple identified the very cyclic beat of a loose belt.

Air Handling Fan
 125hp, 480V, 1780RPM motor driving an air
handling fan – 6 ft between pulleys.
 Torque varying 25% +
 Can’t keep motor and fan aligned.
 Replace fan pillow block bearings every 35-
45 days.
 Replace motor bearings every 6 months.
 Use torque deviation to determine when to
realign.
 No longer necessary to replace bearings so
often.

Motor Current Signature Analysis R K Gupta

More Related Content

What's hot (20)

Similar to Motor Current Signature Analysis R K Gupta (20)

More from RajuGupta88 (6)

Recently uploaded (20)

Motor Current Signature Analysis R K Gupta

Editor's Notes