Vibration Analysis Level - 1 Final

Analysis Level I Course
INTRODUCTION OF CVCM
• PAEC has established Centre for Vibration & Machine Condition

Monitoring (CVCM) at DNPER with the following objectives:
• Universal application of fault detection techniques in NPPs, nuclear reactors,
industrial plants for prevention of critical equipment damage.
• Fault correction with precision balancing, laser alignment, vibration analysis.
• Collection of baseline signatures of major equipment for trend monitoring
equipment performance.
• Shock & Vibration Testing/Measurements on Structures.
• Develop reactor surveillance and condition monitoring systems.
• Organize certification, CM training, workshops and conferences.
Vibration Analysis Level-I
Typical Machinery Problems that Can Be Found Using
Vibration Analysis
• Unbalance
• Misalignment
• Mechanical looseness
• Structural problems
• Bent shaft
• Bearing faults
Typical Machinery Problems that Can Be Found Using Vibration
Analysis
• Gear faults
• Belt problems
• Lubrication problems
• Electrical motor faults
• Cavitation and turbulence
• others
Overview
of
Condition Monitoring
(Maintenance Philosophies)
Definition of Maintenance
• The act of causing to continue

(Webster)
• Keeping equipment in repair

(Oxford)
Maintenance
Reactive Maintenance Planned Maintenance
– Often called ‘Breakdown Maintenance’ – Also known as ‘Shutdown
and has the concept ‘fix it when it Maintenance’. This is based upon
breaks’.
•
‘Timed Intervals’ between
This is probably the most common type of
maintenance in industry today but can be maintenance.
the most costly, especially on critical • Can be very effective if maintenance
machines. and resources are aimed at the
• Maintenance costs are usually higher due machines that need it the most.
to the catastrophic failure that occurs.
• However it can be very difficult to
distinguish which machines actually
need maintenance.
Predictive Maintenance Proactive Maintenance
– Also known as ‘Condition Based – Often referred to as ‘Root Cause
Maintenance’. Analysis’.
Right Balance of• This philosophy works hand in hand
• This approach uses non-intrusive
technologies to determine the actualEach with Predictive Maintenance,
condition of a machine and its rate of eliminating the source of the fault to
failure. try to prevent it from re-occurring.
• This can be very effective in
extending machine life with big
financial savings if implemented
properly.
Today’s Industrial Demand
• It should be unacceptable to deliver
– less performance for more money
– same performance for more money
• It could be acceptable to deliver

– same performance for less money
– more performance for the same money
– more performance for more money
• The desire is More Performance for Less Money!!!!

Predictive Maintenance Objectives
• To confirm good-condition machines
• To detect developing problems
• To determine the nature and severity of the problem
• To schedule repairs that can best fit with production and

maintenance needs
Predictive Maintenance Techniques
• Vibration measurement
• Electrical testing
• Motor current analysis
• Oil Analysis
• Thickness testing
• Visual inspection
• And many more…

Predictive Maintenance Basic Facts
• Every mechanical or electrical faults on a machine has a

distinct vibration behavior.
• Any change in the vibration signature indicates changes in the

dynamic operating condition of the machine
Predictive Maintenance Mechanism (VA)
• Establish a database of all the machines that need to be

monitored
• Establish a data collection route that best optimize the data

collection time
• Download route into the data collector
• Collect data
• Upload collected data into the database

Predictive Maintenance Mechanism
• Run exception reports to detect the problematic machines
• Analyze only the machines in the exception reports
• Generate repair work to be performed
• Again collect data on the machine on which work is being done.

Predictive Maintenance
Rules
Start +
Experi
NO
YES
Create Regular Compare Fault Fault
Ref. Meas. limits Diagnostics correction
Input
m/c
specs
Create
New Ref. & Limits
Overview
of
Condition Monitoring
CBM Techniques
What is CBM & Why ?
• To try and maximise the plants production and increase the
mean time between outages many industries are moved
towards a ‘Condition Based Maintenance’ approach.
• Condition Based Maintenance or CBM is an activity that

attempts to predict and trend component failure non-
intrusively given the end user valuable advanced warning of
the problem at hand.
• Maximising asset reliability is of the utmost importance in

today’s global economy.
CBM Overview
• Most machine faults generate some kind of signature that is
unique to the particular fault developing. By using the correct
technology to detect these signatures we can not only tell
that a fault is developing, but distinguish what the fault type
is.
• There are several technologies available to help determine
the condition of the machine being monitored and the type of
fault developing and these are:
– Vibration Analysis
– Tribology
– Sonics
– Thermography
Diagnosing a machine is just like a person…
Vibration: Oil: Thermography: Motor Current:

The ‘pulse’ of the The ‘life blood’ of ‘Taking its The ‘brain waves’
machine the machine temperature’ of the machine
Total
Picture
Overview of Technologies
• Vibration Analysis
– Used to Detect, Analyse and Confirm plant machinery problems. This
can be done in three ways:
• On-line (CSI 2600 & CSI6500) for automated and continuous monitoring
and protection of critical plant items
• Portable (2130 /2140 Analyser) Route based data collection and analysis
• Wireless used for remote monitoring of moving or inaccessible equipment
• Tribology
– Is the analysis of ‘interactive surfaces in relative motion’.
• Lubricants are analysed on-site using the 5300 mini-lab series.
• The results are plotted in a simple to understand tri-vector plot showing the
‘Chemistry’ ‘Contamination’ and ‘Wear’ of each lubricant, this allowing the lubricant
to be changed on condition rather than on a time-based interval.
Wear
Contamination Chemistry
Tribology and Oil Analysis
• The practical application of Tribology within

mechanical and maintenance engineering is
referred to as Industrial Oil Analysis.
Benefits of Oil Analysis
• Reduce maintenance costs

• Reduce unexpected downtime
– Indication of component failure
– Identify type of damage (chemical, abrasion,
fatigue, or other), and
– Locations of the damage
• Increase equipment availability
Potential Cost Saving from Oil Analysis
• Lubricant consolidation
• Extended oil change intervals
• Extended machine life
• Power consumption
Tribology Technology, a 3D Approach to
Industrial Fluid Analysis
Oil Chemistry
and
Viscosity Iron and Water
Tribology Parameters
• Particle counts
• Ferrous Index
• Large ferrous and non Ferrous indication
• Contaminant Index
• Estimated Water and free water droplet indication
• Chemical Index
• Viscosity at 40 degree C
• Percentage change in viscosity
• Sonics
– Through a process known as ‘heterodyning’ Ultrasonic sounds that are
non-audible to human ears are converted back down to a frequency that
is audible to human ears, allowing the operator to hear and recognise
faults developing within plant operating systems, such as:
• Mechanical – Bearings, Rubs, Gear Defects etc

• Electrical Defects
• Valve Operation
• Steam Trap Operation
• Leak Detection – Pressurised Systems and Vacuum Systems
Bearing Steam Trap Air Leak

• Thermography
– Thermal Imaging is used to locate potential problems by detecting
abnormal temperature fluctuations at a glance.
– This can be used in a wide array of circumstances but is most

commonly used in electrical control panels
Applications of Infrared Thermography
• Condition Monitoring
• Research and Development
• Quality control and process monitoring
• Non-destructive testing
• Security Systems
• And many others!
• Wear Particle Analysis
• Wear particle analysis is a machine condition analysis technology that is
applied to lubricated equipment.
• It provides an accurate insight an accurate condition of a machine’s
lubricated components by examining particles suspended in the lubricant.
• By trending the size, concentration , shape and composition of particles
contained in a systematically collected oil samples, abnormal wear related
conditions can be identified at an early stage.
• Spectroscopy only recognizes particles up to 6 microns and ignores the
larger abnormal wear particles. For this reason it is a good practice to have
Wear Particle analysis performed on oil samples
Motor Testing
• Motor Current Signature Analysis (MCSA) is a condition monitoring
technique used to diagnose problems in induction motors.
• Concept originates from early 1970s and was first proposed for use in
nuclear power plants for inaccessible motors and motors placed in
hazardous areas.
• It is rapidly gaining acceptance in industry today.
• MCSA can be used as predictive maintenance tool for detecting common

motor faults at early stage and as such prevent expensive catastrophic
failures, production outages and extend motor lifetime.
Motor Testing
• MCSA is method from wider field of Electrical Signature Analysis (ESA),
useful for analyzing not only electrical induction motors, but also
generators, power transformers as well as other electric equipment.
• Current spectra can be collected on each power leg at the panel using a
current clamp.
• One general comparison that can be made is the amperage draw from each
leg.
• These should be approximately the same for proper load balancing.
• An exceptionally high or low load indicates problems.
Introduction to Vibration Analysis
Why Vibrations ?
 Early Fault Detection
 Clear Signs of Problem
 Diagnose Variety of Problems
 Accurate Results
What Is Vibration?
Vibration is the pulsating motion of a body from

its original position of rest or neutral position
Upper Limit
Neutral Position
Lower Limit
Vibration in Every Day Life
Vibration is an every day phenomenon,

everyone experience in his daily life.
Useful Vibration
• Vibration is not always destructive,
but have some useful applications
and is produced intentionally
sometimes.
• Heavy drill machines
• Concrete compactors
• Mobile phones
• Massagers
• Component feeders
• Ultrasonic cleaning baths

General Description-Vibration
• There are many different parameters we can measure to help us determine
machinery health:
Voltage Current
Power Flow Temp
Pressure Torque Speed
Viscosity Density Emission
Particles Load
• None contains as much information as the vibration signature!!!
• Not only does it provide the severity of the problem but can also point to the
source of the problem
• ‘Vibration’ can be simply stated as ‘A response to some form of excitation’

– The ‘excitation’ is generally referred to as the ‘Forcing Function’
• Vibration is the motion of a body about a reference position caused by a force
General Description – Forcing Function
• When a forcing function is applied

to a shaft within a plain bearing
the free movement will cause the
shaft to vibrate within the bearing
– Here we are measuring actual shaft movement

Vibration from Mechanical Faults
Vibration Characteristics
• Amplitude How Much
• Frequency How Often
• Phase. When
Vibration Amplitude
Amplitude is the measure of magnitude of

vibration
• Peak Amplitude
• Peak-Peak Amplitude
• RMS Amplitude
47
General Description – Measuring Response
• You can also look at vibration as the amount of ‘Time’ it takes to complete a
particular cycle
– If we examine the motion of a forcing function on a fan blade ‘Heavy Spot’ over
a period of time a distinct signature will occur.
• This motion is called a

sine wave.
– The horizontal axis is
measuring Time
– The vertical axis is

measuring Amplitude
• This is known as a
‘Time Waveform’
– Amplitude versus Time
Time Waveforms
• Unfortunately there are multiple sources of forcing functions that can emit
from a machine or component.
– Thus resulting in the time waveform becoming complex in nature
A8 - Example 15
Ex15 -F2V Fan Outboard Vertical
0.4
• The plot shown on the right

Route Waveform
22-Aug-02 11:33:16
0.3
is a complex time PK = .1495
LOAD = 100.0
waveform. 0.2
1.0
EX 8
Los - Example 8
-P2V Pump Outboard Vertical
RPM = 832. (13.86 Hz)
PK(+) = .3263
Analyze
PK(-) =Spectrum
.3572
– Amplitude versus Time

0.1 15-Nov-95
CRESTF= 10:00:16
3.38
Acceleration in G-s
RMS = 1.27
0.0
0.8 LOAD = 100.0
• This is just one format -0.1

RPM = 737.
RPS = 12.28
(domain) for analysing

RMS Velocity in mm/Sec
0.6
vibration data. -0.2
• Data can also be analysed in -0.3

0.4
a ‘Spectrum’ – (Amplitude -0.4

Time: 120.44
Vs Frequency) through a 0.2
0 50 100 150 200
Time in mSecs
250 300 350 Ampl: -.07595
process known as the FFT

0
Freq: 736.86
0 6000 12000 18000 24000 30000 Ordr: 1.000
Frequency in CPM Spec: .245
Label: Looseness
Frequency Domain
• The frequency domain (Spectrum) plots the data as ‘Amplitude’ in the (Y)
axis and ‘Frequency’ in the (X) axis. This data is derived from the time
domain – mathematical manipulation of the time waveform.
• Recall the waveform and spectrum from the previous slide. If you tried to
determine all the frequencies from the waveform plot, you would need all
day just to analyse one point of data.
• As the FFT plots the frequencies from the waveform for you the analysis of
this data becomes easier and reduces the amount of time needed for analysis
of each point.
A8 - Example 15 Los - Example 8
Ex15 -F2V Fan Outboard Vertical
0.4 EX 8 -P2V Pump Outboard Vertical
Route Waveform 1.0
Analyze Spectrum
22-Aug-02 11:33:16
15-Nov-95 10:00:16
0.3
PK = .1495
RMS = 1.27
LOAD = 100.0
0.8 LOAD = 100.0
RPM = 832. (13.86 Hz)
0.2 RPM = 737.
PK(+) = .3263 RPS = 12.28
PK(-) = .3572
0.1 CRESTF= 3.38
Acceleration in G-s
0.6
0.0
-0.1 0.4
-0.2
0.2
-0.3
-0.4
Time: 120.44 0
0 50 100 150 200 250 300 350 Ampl: -.07595 Freq: 736.86
Time in mSecs 0 6000 12000 18000 24000 30000 Ordr: 1.000
Label: Looseness
Fast Fourier Transform – FFT Process
• When a problem starts to develop within a rotating component it will
generate a vibration signature. This signature should be captured in the time
waveform
– Distinguishing that signature can be very difficult when looking at a
time plot
• To understand the problem we need to understand the frequency
– ‘How often is it occurring?’
• The ‘FFT’ is a process that determines the frequency of a signal from a time
waveform.
• The FFT is named after an 18th century mathematician named ‘Jean
Baptise Joseph Fourier’. He established:
– ‘Any periodic signal can be represented as a series of sines and
cosines’.
– Meaning if you take a time waveform and mathematically calculate the
vibration frequency, it can be converted to a more familiar format
Vibration Spectrum
• Raw vibration signal is rarely sinusoidal and forms a complex

waveform.
• Total vibration signal is actually made up of smaller waveforms,
having varying frequencies
• Each of these individual waveform add up algebraically to one
another to generate total waveform.
• Vibration spectrum transforms the complex vibration signal in
time-domain in clearly defined spectral lines in frequency
domain & this is done by FFT.
How the Vibration Spectrum is Created
Amplitude
Tim
e
Amplitude
Amplitude
y
e nc
e qu
Tim Fr
e
Fast Fourier Transform (FFT) Spectrum
• Fourier proved that any real world complex waveform can be

separated into simple sinusoidal waveform components
• As the sine waves are separated from combined waveform
they are converted into vertical peaks which have an
amplitude & a position along frequency axis.
• Frequency domain presentation of time waveform is called a
vibration spectrum or FFT spectrum.
Amplitude
1x 2X 3X Frequency
54
Fast Fourier Transform (FFT) spectrum (Contd.)
[m/s²] Time(H ) - Input

Working : Input : Input : F F T A nalyzer
40
-40
0 100m 200m 300m 400m

[s]
Raw Signal (Amplitude vs. Time)

[m/s² ] A utospectrum(H ) - Input
Working : Input : Input : F F T A nalyzer
4
0
0 400 800 1.2k 1.6k 2k
[H z]
FFT (Amplitude vs. Frequency)

55
Overall Vibration
Summation of all vibrations present in the spectrum
n
 Ai
2
1 2 2 2
RSS  i 1
 A1  A2  A3  ...
N BF 1.5
OARSS = Digital Overall Level of Vibration Spectrum

RSS = Root Sum Square Level (energy summation) in Spectrum.
n =Number of FFT lines of Resolution
Ai =Amplitude of each FFT line.
NBF =Noise Band Width 1.5 for Hanning Window
56
Measuring Amplitude and Frequency
• You can measure amplitude from a time waveform as shown:
0 to Peak
+
RMS
Average
Amplitude
Time ‘t’
-
Peak to Peak
• The period ‘t’ is the time required for one revolution of the shaft in this
illustration, which equals one cycle of the waveform
– During this period, the amplitude of the waveform reaches a positive (+) peak,
returns to rest, and reaches a negative (-) peak before returning to rest
• Peak (Pk) – Amplitude measured from the ‘at rest’ position (0) to the
highest value (0 to Peak)
• Peak to Peak (Pk-Pk) – Amplitude measured from the peak positive (+)
value to the peak negative (-) value
• RMS (Root Mean Square) – obtained by averaging the square of the signal
level over a period of time and then taking the square root result
• Average (Avg) – Amplitude value that averages the peak values of the
waveform
0 to Peak
• You can calculate the different RMS
amplitudes when one of the values are Average
known:
Amplitude
– RMS = 0.707 times the peak value Time ‘t’
+
– Avg = 0.637 times the peak value
– Pk-Pk = 2 times the peak value
-
Peak to
Peak
• Severity of a vibration problem can be determined by the amplitude of
the vibration.
• We can measure amplitude in one of three ways
1. Displacement – measures the distance the shaft moves in relation to a
reference point.
2. Velocity – measures the displacement of the shaft in relation to time
3. Acceleration – measures the change in velocity in relation to time
• The most common industrial applications are:

1. Displacement - Microns - Peak to Peak value
2. Velocity - mm/sec - RMS
3. Acceleration - G-s - Peak value
– G-s = 1 x force of gravity (G-force)
Amplitude Relationships
• The three types of amplitude measurements used to display data are
directly related to each other
– Changing from one amplitude unit to the next alters the way in which the data is
displayed
• Velocity is the default unit

for standard data collection
0.35
1405
Ex15
A8 - Example 15
-F1H Fan Inboard Horizontal techniques
Route Spectrum
– High and low frequency
Low frequencies require very 22-Aug-02 11:30:50
events can be seen

0.30
120
4
Forlittle
normal
forceoperating
to move anspeed
object
OVERALL= 3.45 V-DG
RMS==.3909
PK
P-P 3.44
104.98
0.25
100 ranges, velocity data
LOAD = 100.0
RPM = 831. (13.85 Hz)
 Displacement measures
low frequency events
AccelerationininMicrons
provides the best indication of

G-s
0.20
3
80
Increasing
machine the frequency that
condition
ignoring high frequencies
Displacement
the objects move with the – Relative shaft motion

0.15
60
2
same velocity, the force
 Acceleration accentuates
P-PPK
needed to move it increases,

the high frequencies
0.10
40
thereby reducing the distance

1
0.05
20
it can travel
ignoring the low frequencies
– Good for early bearing
00
detection (Whenever there
0 20000
Frequency in CPM
40000 60000
is Metal to Metal Impacting
Label: Large Fan Unit - Easy involve)
Vibration Amplitude (Contd.)
Upper Limit
Peak Peak to Peak
Mean Position
RMS
Lower Limit
Multiply
Number → PEAK –
PEAK RMS
BY PEAK
To Obtain ↓
PEAK-PEAK 1.000 2.000 2.828
PEAK 0.500 1.000 1.414
RMS 0.354 0.707 1.000
61
Displacement, Velocity & Acceleration Conversion
Formulas
METRIC UNITS
V = DF/19,100 D = 19,100V/F
V = 93,640 A/F A = VF/93,640
A = DF2/1,790,000,000 D = 1,790,000,000A/F2
Where,
D = Peak-To-Peak Displacement (Mils Pk-Pk)
V = Peak Velocity (In/Sec-Pk)
A = Peak Acceleration (G’s-Pk)
F = Frequency (CPM)
62
When To Use Displacement, Acceleration & Velocity
Dis
p l ac
em on
en t r at i
c ele
Ac
LOG AMPLITUDE
Velocity
1 10 100 1000 10000

LOG FREQUENCY (Hz)
Comparison between Displacement, Velocity and Acceleration
63
Vibration Nomograph
64
Frequency Units
• Frequency refers to how often something occurs:
– How often a shaft rotates?
– How often a rolling element hits a defected race?
• There are three ways to express frequency:
1. CPM – Cycles Per Minute
– 1CPM = 1RPM
2. Hz – Cycles Per Second
– CPM / 60
3. Orders – Multiples of Turning Speed
– Frequency/Turning Speed
• Consider a motor has a rotational speed of 1485RPM, in terms of

frequency this equates to:
– 1485 CPM (1rpm = 1cpm)
– 24.75 Hz (1485/60) (minutes to seconds)
– 1 Orders (1 x revolution of the shaft)
Frequency Units
• Shown below is a table showing the
relationship between all three frequency
units with reference to the turning speed
Motor Turning Speed = 1500RPM

CPM 1500 2250 3000 6000 12000
Hz 25 37.5 50 100 200
Orders 1 1.5 2 4 8
Frequency Domain
• The vibration analyst can divide the frequency domain data

into three major areas of interest
1. Synchronous Equal to Ts or Harmonics of Ts
2. Sub synchronous < 1 x Ts
3. Non synchronous > 1 x Ts but not an integer
• Note ‘Ts’ is the turning speed or rotational frequency (RPM)
of the shaft at the position where you make the
measurement
• Each defect that can materialise in the frequency domain
can be categorised into one of three types of energy listed
above
– Knowing the type of energy within the data can help the analyst
quickly eliminate 2/3rd of the fault types
Harmonic Orders
• Harmonics are cursors that are exact multiples of the primary frequency
– They are used to locate other frequencies related to the primary cursor
Los - Example 3
EX3 -P2V Pump Outboard Vertical
1.0
Analyze Spectrum
15-Nov-95 10:00:16
Here the primary cursor is at 1 RMS = 1.27

LOAD = 100.0
0.8
Order (1xTs). All the other cursors RPM = 737.
RPS = 12.28
are harmonics (exact multiples of
R M S Velocity in mm/Sec
0.6 the primary cursor)
0.4
• Therefore: 0.2
– When the primary cursors is located on 1Order all the harmonics will be
synchronous 0
Freq: 736.86
– Harmonic cursors can be used to show non-synchronous and sub-
0 6000 12000 18000
Frequency in CPM
24000 30000 Ordr: 1.000
Spec: .245
synchronous harmonics depending upon the energy of the primary

frequency
Energy in the Spectrum
C1 - Example 4
E4 -MOH MOTOR OUTBOARD HORIZONTAL
0.5
Route Spectrum
09-Feb-00 12:41:33
OVRALL= .5785 V-DG

0.4 RMS = .5716
LOAD = 100.0
RPM = 2937.
RPS = 48.95
0.3
0.2
0.1
0
Freq: 2.937
0 20 40 60 80 100 120 140 160 Ordr: 1.000
Frequency in kCPM Spec: .01038
Synchronous Energy
Los - Example 8
EX 8 -P2V Pump Outboard Vertical
1.0
Analyze Spectrum
15-Nov-95 10:00:16
• Synchronous energy - related to RMS = 1.27

LOAD = 100.0
0.8
turning speed. RPM = 737.
RPS = 12.28

• We can see from the 0.6
spectrum that the first peak

is at 1 Orders (which means it 0.4
is 1 x turning speed)
0.2
• All the other peaks are
harmonics off, which means
0
they are related to the first peak 0 6000 12000 18000 24000 30000
Freq: 736.86
Ordr: 1.000
Label: Looseness
Examples of synchronous energy:

1) Imbalance 2) Misalignment 3) Gearmesh
Non-Synchronous Energy
BF - Example 5
E5 -R4A ROLL BRG. #4 - AXIAL
2.0
Route Spectrum
12-Jul-96 17:16:42
• Non-synchronous energy - 1.8
OVRALL= 2.63 V-DG
1.6 RMS = 2.69
not related to turning speed LOAD = 100.0
MPM = 3225.
1.4 RPM = 380.

1.2
• We can see from the 1.0
spectrum that the first 0.8
peak is at 10.24 Orders. 0.6

This is not related to 0.4
turning speed. 0.2
0
Freq: 3888.9
0 6000 12000 18000 24000 30000 Ordr: 10.24
Label: Outer Race Defect
Priority: 1
• Examples of non-synchronous energy:

• Bearings Multiples of belt frequency Other Machine Speeds
Sub-Synchronous Energy
• Sub-synchronous energy - Less
than turning speed
• The spectrum shows the first

impacting peak below 1 Order.
This is sub-synchronous energy
• Examples of sub-synchronous
energy are:
• Belt Frequencies
• Other Machine Speeds
• Cage Frequencies
Energy in a Spectrum
 Synchronous
– N x RPM where N is an integer
 Sub-synchronous
– <1 x RPM
 Non-synchronous
– F x RPM where F is >1x RPM but not integer
Causes of Sub Synchronous Energy
• Frequencies that show
below the rotational
frequency (Less than 1
Order) are sub synchronous.
– Another component
– Cage frequencies
– Primary belt frequency
– Oil whirl (plain bearings)
Causes of Synchronous Energy
• Frequencies that are equal
too or a direct multiple of
running speed are
Synchronous
• Possible causes of
Synchronous energy are: Los - Example 8
–
EX 8 -P2V Pump Outboard Vertical
Imbalance
1.0
Analyze Spectrum
15-Nov-95 10:00:16
RMS = 1.27
–
0.8 LOAD = 100.0
Misalignment RPM = 737.

RPS = 12.28

–
0.6
Looseness
–
0.4
Vane pass frequency

–
0.2
Gears etc
0
Freq: 736.86
0 6000 12000 18000 24000 30000 Ordr: 1.000
Label: Looseness
Causes of Non Synchronous Energy
• Frequencies above (but not
integer multiples of) turning
speed are non synchronous.
• Possible causes of non
synchronous energy are:
– Another component
– Antifriction bearings BF - Example 5
–
2.0
Electrical 1.8
Route Spectrum
12-Jul-96 17:16:42
OVRALL= 2.63 V-DG
–
1.6 RMS = 2.69
System resonances 1.4

LOAD = 100.0
MPM = 3225.

RPM = 380.
– Multiples of belt frequency

1.2
1.0
0.8
0.6
0.4
0.2
0
Freq: 3888.9
0 6000 12000 18000 24000 30000 Ordr: 10.24
Priority: 1
Lines of Resolution
(LOR)
Lines of Resolution
• Lines of Resolution (LOR) determine how clear the peaks(data) are
defined within our spectrum.
• The more lines we have over the same F-max (Maximum frequency
scale). The more accurate our data will be
• Example.
– The diagram below shows data that has been collected using 400 LOR.
Notice how the top of the peaks are capped. When the LOR are
increased the data becomes more accurate.
Lines of Resolution
L2 - TA 16
TA16 -M1H Motor Outboard Horizontal
0.20
0.5
Analyze Spectrum
13-Mar-01 09:14:16
09:13:53
PK = .3852
.7078 • The spectrum shown displays
0.16
0.4 LOAD = 100.0
RPM = 1497.
1496.
RPS = 24.95
24.94
data at 800 L.O.R with an
Fmax of 1600 Hz
PK Acceleration in G-s
PK Acceleration in G-s
0.12
0.3
0.08
0.2  The second spectrum
displays the same data but
0.04
0.1 with 3200 L.O.R over the
same Fmax
0
0 400 800 1200 1600
Frequency in Hz
Lines of Resolution
• The range of LOR settings that we can choose from on the analyzer
starts at 100 Lines and go up to 12800 Lines.
• The average number of LOR is around 1600 Lines for a typical

motor/pump set up
To change the LOR settings we need to alter our parameter set.

This is done in the Database Setup program
Remember. If you double your lines of resolution you double your

data collection time.
Selecting Number of Averages
• To remove data from transient sources more than one data sample is
usually collected for predictive maintenance.
• Increasing number of averages will also increase data collection time
• Time= (Lines of Resolution x No. of Averages)/Frequency Range
• General rule is to take 2-4 averages for each spectrum in a routine
checkup
• For trouble shooting purposes the number of averages may increased
to 8-10.
• Also for machines which have higher frequencies, 4-8 averages should
be taken.
Spectral Summary
BF - Example 5
2.0
Route Spectrum
1.8 Energy Types Harmonics
12-Jul-96 17:16:42
Displacement
OVRALL= 2.63 V-DG

1.6 Synchronous Multiples of Primary RMS = 2.69
LOAD = 100.0
Non Synchronous Frequency MPM = 3225.
1.4 RPM = 380.
Velocity in mm/Sec
Sub Synchronous
Amplitudes
1.2
Resolution
Velocity
1.0
Clarity of the spectral
0.8 data
RMS
Acceleration,
0.6
0.4
0.2
0
Freq: 3888.9
0 6000 12000 18000 24000 30000 Ordr: 10.24
FrequencyFrequency in CPM Spec: .748
Priority: 1
Hz (CPS), CPM, Orders
Phase
• The position of a vibrating part at a given instant with
reference to a fixed point or another vibrating part
• It shows a picture of how different members of a machine are

vibrating relative to each other
PHASE IS A POWERFUL TOOL IN

VIBRATION DIAGNOSTICS
83
Phase (Contd.)
Weights vibrating 180 degrees out-of-phase. Weights vibrating 90 degrees out-of-phase.
Weights vibrating 0 degrees out-of-phase or "in-phase."

Phase Relationship as Used With Machinery Vibration
1 2
PHASE 180o BETWEEN POINTS 1 & 2

1 2
PHASE 0o BETWEEN POINTS 1 & 2

Significance of Phase
Problems that cause vibration at a frequency of 1xRPM,
including:
• Unbalance
• Bent shafts
• Misalignment of couplings, bearings and pulleys
• Looseness
• Distortion from soft feet and piping strains
• Resonance
• Reciprocating forces
• Eccentric pulleys and gears
86
(Data Collection)
Transducers and Mounting Techniques
• Although there are many different types of transducers
available, the most common type used for day to day data
collection are Accelerometers.
• These transducers provide an electrical charge proportional to
acceleration by stressing piezoelectric crystals typically
100mV/g sensors are used.
Transducers for Vibration Detection
• Accelerometers; Measures Acceleration
• Velocity Transducers; Measures Velocity
• Non contact Eddy Current Displacement

Probes; Measures Displacement
Eddy Current Displacement Probes
Advantages Disadvantages
• Measures displacement • Value of displacement is small at high
• Non contacting frequencies
• Variation in geometric irregularities &
• No moving parts
magnetic properties of shaft give
• No wear incorrect results
• Permanent installation
Velocity Pickup
Advantages Disadvantages
• Measure Velocity • Moving parts prone to wear
• Self Generating • Large size
• Low Impedance • Sensitive to orientation
• High lower limiting frequency (> app. 10
hz)
Piezoelectric Accelerometer
Advantages
• Rugged Disadvantages
• High impedance output
• Very large dynamic range
• No true DC response
• Wide frequency range
• Compact, often low weight
• High stability
• Can be mounted with any
orientation
Importance of Accelerometer Mounting
Four common methods of accelerometer

mounting are:
• Stud Mounting
• Adhesive Mounting
• Magnetic Mounting
• Extension Probe
Four Different Methods of Accelerometer Mounting
Hand-held Mount
Magnet Mount
Stud Mount
Adhesive Mount
94
Handle the Accelerometer Carefully
A drop on hard floor or hit against machine part

may result in change in sensitivity or even
severe damage to the accelerometer
95
Data Quality
• Whether it is your job to collect the data and/or analyse the
data it is important to understand that the technologies will
not give you the answer to a machines problem unless you
have collected meaningful, quality data
• There are certain considerations that must be taken prior to any

data being collected, these are:
– A good understanding of the internal make up of the machine, in order
to understand the best transmission path for data collection - bearing
locations, load zones etc.
– Ensure data is collected in a repeatable manner so we can compare two
or more readings to each other - trending purposes
– Variable speed machines - it is very important to collect data with the
correct running speed enter into the analyser
Transmission Path
• Damaged caused to a machine component will cause a certain
amount of vibration/sound or heat to propagate away from the
initial impact.
– It is the effect of the impact/force that we are trying to detect
• In many cases the further you are away from the initial event the
weaker the signal will become, resulting in the data appearing to be
lower in value.
– In more extreme cases the impact can be lost amongst other machine noise
by the time it has reached your transducer, resulting in no detection of a
machine problem.
• Usually the best place to acquire data from a machine, is at the bearings.
– This is because the bearings are the only part of the machine that connect the
internal rotating components to the stationary components (Casing)
Repeatable Data
• Collect data in the same manner each time.
– This consistency will allow you to trend the machinery condition and
properly judge the progression of faults
• In order to aid with repeatable data the analyser requests for

data to be collected in certain locations on the machine.
– These are called ‘Measurement Points’
 A measurement point is determined by three characters

and a description.
 Each character refers to a particular place on the
machine being monitored
– E.g. M1H is a typical measurement point
Measurement Points
• A measurement point is defined as three alpha numeric digits
along with their respective definition
– Orientation and location on each component
• The image on the right is
taken from the screen of
the 2130 analyser during
a collection ‘route’
• The measurement ‘point
identifier’ can be seen in
the top right while the
‘point description’ is
shown just below
Measurement Points
• The first letter of the ‘Point Identifier’ refers to the type of
machine being monitored
– M = Motor P = Pump F = Fan
• The second character represented by a number indicates the
location on the machine
– Inboard (Drive End) or Outboard (Non Drive End)
• The third letter refers to the orientation of the sensor or the
type of processing being done by the analyser
– H = Horizontal V = Vertical P = Peakvue Change in DSP of Analyser
Measurement Points
• The following example shows how the numbering system
changes as you cross from one component to the next
 M1H – Motor Outboard Horizontal
 M1P – Motor Outboard Horizontal Peakvue 2
1
1
2
 P1H – Pump Inboard Horizontal
 P1P – Pump Inboard Horizontal Peakvue
• Notice how the ‘1’ is not always the ‘Outboard’

– This changes when the next component is required for data collection
• The numbering system starts from 1 again
Creation of Route
A1 - Recirculation Pump #5
RCP#5 -M2H MOTOR INBOARD BRG. - HORIZONTAL
0.025 Trend Display
FAULT 36-65xTS
PK In/Sec
0.020
0.015
0.010
ALERT
0.005
 • 0
0 The
Eachmachines
technologydue
100 to beand
is stored
200 monitored are defined
analysed from
300 within the
a single software
400 platform,
500
software.
Days: 11-Aug-95 To 11-Dec-96
0.05
allowing the analyst to: Route Spectrum
11-Dec-96 17:33:57
PK In/Sec
0.04
OVERALL= .0604 V-DG
As much information as possible about the machines being monitored

0.03
0.02
0.01
–– Store all data and information in one database PK = .0605
LOAD = 100.0
is preferred when building the database.

RPM = 3593. (59.89 Hz)
0
0
– 40
Easily cross reference data for conformation of analysis
80 120
Frequency in kCPM
160 200 240
–
1.0 Route Waveform
Collaborate all data into one single report.
Acc in G-s
0.5 11-Dec-96 17:33:57

0 RMS = .4233
-0.5 PK(+/-) = 1.13/1.22
CRESTF= 2.89
-1.0
-1.5
0 1 2 3 4 5 6 7
Revolution Number
Types of Vibration Frequency Analyzers
• Analog Swept Filter Analyzer
• Digital FFT Analyzer
• Real Time Analyzer

Swept Filter Analyzer
• These are older instruments used for vibration analysis not commonly
used in these days.
• In these analyzers the frequency is tunable for range of interest.
• Selector switches are provided for selecting the desired parameter of
vibration amplitude, full-scale amplitude range and filter "bandwidth"
characteristics.
• A hard copy print out of the spectrum can be obtained directly from the
analyzer
• Not able to store data and producing better resolved frequency spectra
• Not have the compatibility with predictive maintenance software.
• Outdated these days and replaced by FFT Analyzers
Swept Filter Analyzer
FFT Data Collectors
• Widely used for current predictive maintenance programs

• Have the compatibility to store large amount of data
• Have the ability to transform time waveform data to FFT spectrum
• Compatibility with most predictive maintenance soft wares
• Available in single channel to four channel.
• Frequency range lies normally between 1 – 25kHz
• Very handy & light in weight
• Many of the advance FFT analyzers have the ability to perform
onsite analysis using special tests
FFT Data Collector
Real Time Analyzer
• Most powerful diagnostic tool for advanced diagnostic techniques

• The real time display updates very quickly.
• Have the ability to record and play them back over and over again.
• Also able to catch very short duration transient events (less than 20s)
• Have multi channel capabilities
• Excellent in performing special tests like natural frequency test, run
up/ coast down test, transient capture etc
• Complex in nature and require additional training for use
• Bulky and not very portable.
Real Time Analyzer
Good Vibration Analyzer
• Portability  Ease of Use

• Typical Frequency Range  Time Waveform Storage Capability
• Data Measurement Format  Frequency Spectra Storage
• Typical Display Type Capability
• Typical Transducer Type  Predictive Maintenance Software
• Phototach Capability Compatibility
• Multi Channel Availability  Natural Frequency Testing
• Spectral Live Display Capability
 Relative Costs
110
Vibration Analysis & Diagnostics
111
Vibration Analysis
Following steps are involved while analysis or solving a vibration

problem:
• Define the Problem
• Determine Machine History
• Determine Machine Details
• Visual Inspections
• Obtain Tri axial Data on Each Bearing
• Identifying the Most Common Machinery Problems
112
Define The Problem
• Establish base line data for future analysis needs
• Identify the cause of excessive vibration
• Identify the cause of significant vibration increase
• Identify the cause of frequent component failure
• Identify the cause of structural failures
• Identify the source of noise problem
• Identify why a machine fails to meet an established performance

standard
113
Determine Machine History
• When did the problem starts?
• Were any changes made to the machine

before the problem occurred?
114
Determine Machine Details
• RPM of each machine component
• Types of bearing
• Number of fan blades
• Number of impeller vanes
• Number of gear teeth
• Type of coupling
• Machine critical speeds
• Background vibration sources

115
Visual Inspection
• Loose or missing mounting bolts
• Cracks in the base, foundation or structural

welds
• Leaking seals
• Worn or broken parts
• Wear, corrosion or build-up of deposits on

116
rotating elements such as fans etc
Obtain Tri-Axial Data at Each Bearing
• Obtain the data on both drive end and non drive end of each of the
machine component in all the three directions i.e.
– Horizontal
– Vertical
– Axial
• Define frequency and amplitude parameters
• Identify the problem component based on frequency and amplitude
• Reducing the list of possible problems based on frequency
• Consult time waveform to confirm the problem
117
Fault Diagnostics
• Each type of machine fault or defect reveals a specific

vibration characteristic in the spectrum and time waveform
domain that distinguish that fault from another.
• Simply by gaining a basic knowledge of these patterns and
applying a few rules of thumb we can start to analyse
machine vibration and prevent machine failure.
• This section concentrates the characteristics / patterns and
rules that apply to diagnose machine faults such as:
– Imbalance Misalignment Looseness Gears
– Bearings (Peakvue) Belts Electrical
– Resonance
Imbalance
Imbalance
• Imbalance (Unbalance) occurs when the centre of mass differs

from the centre of rotation.
• If the centre of mass changes on the rotor due to a heavy spot
or some other influence then a centrifugal force is produced.
This results in the centre of rotation being offset from the
centre of mass causing the vibration to increase at the
rotational frequency.
Imbalance (Types)
Imbalance
• Causes of Imbalance
– Improper Assembly
– Material build up / dirt
– Wear to components
– Broken or missing parts
All of the above conditions will result in an unbalanced state
• Diagnostic Rules for Imbalance

– Periodic non-impacting sinusoidal waveform
– Spectral peak at 1xTs (1 Order)
– Very little axial vibration
– Similar amplitudes between horizontal and vertical plains
– Synchronous fault type
– Amplitudes will increase with speed
– Very low harmonics of 1xTs
Imbalance Spectral Data
• The spectrum shown represents a simple unbalance state
– Single peak at 1xTs (1 Order)
– Little indication of harmonics
IF - Example 2
Ex2 -F1H Fan Inboard Horizontal
6
Route Spectrum
16-Sep-99 08:36:29
5
• What should the OVRALL= 4.58 V-DG
RMS = 4.56
LOAD = 100.0
4
waveform show? RPM = 3000.
RPS = 50.00
0
Freq: 3000.0
0 20000 40000 60000 80000 Ordr: 1.000
Frequency in CPM Spec: 4.539
Imbalance Waveform Data
• Despite the waveform being displayed in Acceleration
– Default unit for route based waveform data
• There is still a predominant sinusoidal waveform pattern
– 1 x Revolution sine wave
IF - Example 2
Ex2 -F1H Fan Inboard Horizontal
1.0
Waveform Display
02-Feb-00 15:13:51
0.8
PK = .5289
0.6 LOAD = 100.0
RPM = 2985.
RPS = 49.76
0.4
PK(+) = .8332
PK(-) = .8893
0.2
CRESTF= 2.38
Acceleration in G-s
-0.0
-0.2
-0.4
-0.6
-0.8
• Changing the units to velocity would reduce the amount of high

-1.0
frequency noise residing on the waveform

0 0.5 1.0 1.5 2.0 2.5 3.0
Revolution Number
3.5 4.0 4.5 5.0
Imbalance Trend Data
• The trend data is a good way of determining if there has been
a change in condition, as this plots amplitude against time
(where time is in days)
• Here the 1xTs parameter is being trended
– Vibration has been steady at 3mm/sec for a period of time
– A sudden change instate should alert the analyst to a fault developing
E02N - JB1420C CONDY RECOVERY PUMP
JB1420C -M1H Motor Outboard Horizontal
14
Trend Display
of
1xTS
12
-- Baseline --
Value: 3.063
10
Date: 07-Apr-00
8
FAULT
0
Date: 21-May-01
0 100 200 300 400 500 Time: 14:24:29
Days: 07-Apr-00 To 21-May-01 Ampl: 11.21
Imbalance Problem - Practical
• The following fan unit has an imbalance present on the rotor.
– 1xTs Peak in the Spectrum
– 1xTs Peak in the Waveform
Imbalance
IF - Example 2
Ex2 - F1H Fan Inboard Horizontal
6
Route Spectrum
16- Sep- 99 08: 36: 29
R M S V e lo c it y in m m /S e c
5 OVRALL= 4.58 V- DG
RM S = 4.56
LOAD = 100.0
RPM = 3000.
RPS = 50.00
4
0
Freq: 3000.0
0 20000 40000 60000 80000 Ordr: 1.000
IF - Example 2
Ex2 -F 1H Fan Inboard Horizontal
1.0
Waveform Display
02-Feb-00 15: 13: 51
0.8
PK = .5289
0.6 LOAD = 100.0
A c c e le r a t io n in G - s
RPM = 2985.
RPS = 49.76
0.4
PK(+ ) = .8332
PK(-) = .8893
0.2
CR ESTF = 2.38
-0.0
-0.2
-0.4
-0.6
-0.8
-1.0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Revolution Number
• What would happen to the data if the following occurred to

the fan?
Imbalance Case Study 1
• Background
• The following data is taken from a Recirculation Fan designed to circulate
the hot air through an Oven to aid with drying the process. The oven is
vertically mounted and the product comes into the oven from the top and
exits at the bottom. There is one Recirculation Fan and one Extract Fan.
Loss of function from either fan results in the oven being taken offline.
Bm/c - TOP RECIRC FAN
TRF B m/c -F2H Fan Outboard Horizontal
6
ROUTE SPECTRUM
08-Nov-04 14:16:45
• The spectral plots shows 5 OVERALL= 4.70 V-DG

RMS = 4.70
a dominant 1xTs peak (1 LOAD = 100.0

RPM = 1246.
RPS = 20.77
4
Order) with very little
other vibration present 3
0
Freq: 1246.3
0 20000 40000 60000 80000 100000 Ordr: 1.000
• The waveform from this data shown on the following page
represents a sinusoidal waveform clearly shown once per
revolution of the shaft – here the waveform is displayed in
velocity.
8
• All indications point

Route Waveform
08-Nov-04 14:16:45
RMS = 4.66
towards an imbalance 4 LOAD = 100.0

RPM = 1246.
RPS = 20.77
problem. The 0
PK(+) = 7.03
PK(-) = 7.40
Velocity in mm/Sec
CRESTF= 1.59
amplitudes should be
checked in both radial -4
directions to confirm -8
this problem
-12
0 1 2 3 4 5
Revolution Number
• The plot shown indicates a multi spectral plot showing all the
radial directions.
TRF B m/c - Multiple Points (08-Nov-04)
Max Amp
4.27
Plot
Scale
TRF B m/c -F2V
5
• It is clear that the TRF B m/c -F2H
amplitudes are
common to both 0 TRF B m/c -F1V
bearing – Inboard
and Outboard in all 0 8000 16000
TRF B m/c -F1H
24000
radial plains Frequency in CPM

• The trend data for the 1xTs parameter has been steady for a
considerable amount of time. The last two readings has shown
a significant increase in amplitude

6
Trend Display
of
1xTS
5
-- Baseline --
Value: .428
Date: 24-Nov-00
• The fan was 3
recommended to be 2
FAULT
cleaned at the next 1

ALERT
available opportunity
and for it to be re- 0
0 300 600 900
Days: 24-Nov-00 To 08-Nov-04
1200 1500
Date: 08-Nov-04
Time: 14:16:38
Ampl: 4.688
tested afterwards
Imbalance Summary
• Diagnostic Rules for Imbalance

– Periodic non-impacting sinusoidal waveform
– Spectral peak at 1xTs (1 Order)
– Very little axial vibration
– Similar amplitudes between horizontal and
vertical plains
– Amplitudes will increase with speed
– Very low harmonics of 1xTs
Misalignment
Misalignment
• When two mating shafts do not share the same collinear axis
then misalignment is induced.
• Misalignment is one of the primary reasons for premature

machine failure. The forces that are exerted on the machine
and its components when in a misaligned state are greatly
increased from normal operating conditions
Misalignment
• Misalignment can be broken into three basic categories, these
are:
• Angular – Where the shaft centrelines cross producing a 1xTs peak axially
 Offset – Where the shaft centrelines are parallel but they do not meet
producing a radial 2xTs peak
 More commonly seen – A combination of the above

Misalignment
Misalignment
• Another common problem
associated with alignment is
‘bearing misalignment’.
• Bearing misalignment occurs
when the bearings are not
mounted in the same plain
possibly due to:
– one or more of the bearings
being cocked in the housing
– The machine itself distorts due to
thermal growth or soft foot
conditions
– Misalignment at the drive causes
shaft bending.
Misalignment
• Diagnostic Rules for Misalignment
– High axial levels of vibration at 1xTs
– High radial levels of vibration at 2xTs
– Repeatable period sine waveform showing 1 or 2 clear peaks per
revolution (Most likely “M” or “W” shape)
– Data can usually be seen across the coupling
• Diagnostic Rules for Bearing Misalignment

– High levels of vibration at 1xTs and 2xTs
– Repeatable periodic sine waveform showing 1 or 2 clear peaks per
revolution
– Data usually shown either the driver or driven component
Offset Misalignment Spectral Data
• The spectral data shown represents a simple misalignment
plot.
– The primary cursor denotes the 1xTs peak while the harmonic cursors
indicate a larger 2xTs peak. This type of data is common to that of
Offset Misalignment
ST.1 - Raw Water Pump
P029 -M2H
7
Route Spectrum
15-FEB-93 11:04:18
6
OVRALL= 6.50 V-DG
RMS = 6.47
LOAD = 100.0
5 RPM = 2976.
RPS = 49.61
0
Freq: 2925.0
0 10000 20000 30000 40000 50000 Ordr: .983
Angular Misalignment Spectral Data
• The spectral data below represents a simple misalignment
plot.
– The primary cursor denotes the 1xTs peak while the data was taken
in the axial direction. This type of data is common to that of Angular
Misalignment
B29 - PUMP NO 3
3601PUM003-M2A Motor Inboard Axial
8
Route Spectrum
04-Aug-04 08:49:05
7
OVERALL= 6.33 V-DG
RMS = 6.31
LOAD = 100.0
6
RPM = 1071. (17.84 Hz)
0
Freq: 1.071
0 30 60 90 120 Ordr: 1.000
Frequency in kCPM Spec: 5.966
Offset Misalignment Waveform Data
• The waveform above is showing two clear peaks per
revolution of the shaft. This type of waveform resembling an
‘M’ or ‘W’ shape is common to offset misalignment.
– Data shown in velocity
ST.1 - Raw Water Pump

P029 -M2H
40
Waveform Display
26-MAR-93 13:32:52
30
RMS = 17.00
LOAD = 100.0
RPM = 2996.
20 RPS = 49.93
PK(+) = 30.66
Velocity in mm/Sec
PK(-) = 26.81
10
CRESTF= 1.82
-10
-20
-30
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Revolution Number
Misalignment Waveform
• The waveform data shown above is predominantly showing
one sinusoidal waveform per revolution of the shaft.
– Here the data is shown Acceleration
B29 - PUMP NO 3
3601PUM003-M2A Motor Inboard Axial
0.8
Route Waveform
04-Aug-04 08:49:05
0.6
PK = .2596
LOAD = 100.0
RPM = 1071. (17.84 Hz)
0.4
PK(+) = .6277
PK(-) = .5683
0.2 CRESTF= 3.42
Acceleration in G-s
0.0
-0.2
-0.4
-0.6
-0.8
Rev : .680
0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 Ampl: -.306
Revolution Number
Case Study 3 – Kiln Main Motor Gearbox
• Introduction
• The Kiln drive gearbox motor had been replaced during a
planned plant shutdown.
• During the start up of the plant after the shutdown it was
noted that the motor and gearbox were excessively noisy.
Vibration data was taken during the run up of the plant to
determine the cause of the problem.
Main Motor
Kiln Gearbox
• The spectral plot shown above is the data taken from the drive
end of the motor. Here there is a dominant 2xTs peak.
04 - Kiln Drive
0804 -M2H Motor Inboard Horizontal
2.4
Route Spectrum
 In addition to the misalignment the excessive forces 2.1
29-Mar-01 11:33:43
being applied to the machine were causing excessive

OVRALL= 2.47 V-DG
RMS = 2.46
LOAD = 100.0
loading on the gears.

1.8
RPM = 1418.
RPS = 23.64
04 - Kiln Drive
1.5 0804 -G2A Shaft 01 Outboard Axial

5
Max Amp
5.98 4
3
1.2 2
1
A m p lit u d e - M ixed U n its
0
29-Mar-01
0.9 09:40: 20
After 29-Mar-01
09:40: 09
0.6 Shutdown 26-Mar-01

12:11: 12
0.3 23-Jan-01
15:02: 00
0 Before 25-Oct-00
09:04: 17
Freq: 1418.3
0 20000 Shutdown 40000 08-Aug-00
60000 Ordr: 1.000
Frequency in CPM 14:06: 56 Spec: 1.346
0 60 120 180 240 300
Frequency in kCPM
• During data collection it was also observed that the grouting
around the front feet of the motor had begun to crack as a
result of the excessive force being applied to the motor base
and feet due to the misalignment.
• Conclusion
– It was confirmed the engineers that replaced the motor during the
shutdown and assumed as the motor was a like for like swap, as long as
they kept the shims in the correct place then alignment was not
necessary.
– Corrective action was required and production was stopped so the
motor could be re-aligned and the mountings re-secured.
Misalignment Summary
• Diagnostic Rules for Misalignment

– Periodic non-impacting sinusoidal waveform with
1 or 2 clear peaks per revolution (Most Likely “M”
or “W” shape)
– Spectral peak at 1xTs and 2xTs
– Axial vibration at 1xTs
– Data can be seen across the coupling or across the
component itself
Looseness
How would looseness ?
Looseness
• Looseness can be broken down into two main categories,
Structural and Component
 Structural looseness occurs when there is free
movement within the machines support structure causing

excessive vibration. This can be a result of:
– Loose support bolts to the components feet and supports
– Cracked welds
– Deterioration of the base itself.
 Component looseness generally occurs when there is
excessive clearance to the components within the
machine, such as:
– Excessive clearance between the shaft and bearings
– Excessive clearance between the shaft and an impeller etc.
Looseness
• Diagnostic Rules for Looseness
– Multiple harmonics of the 1xTs peak - Structural
– Multiple Harmonics of the component that is loose - Component
– Number of harmonics will increase as the looseness progresses
– Random, non-periodic waveform - Structural
– Waveform shows predominant impacts - Component
– Raised noise level around the 1xTs + harmonics
– Half harmonics may also be present
– Can be present in all Directions
Looseness Spectral Data (Structural)
• The spectral plot shown is demonstrating Looseness.
– The 1xTs peak has been highlighted by the primary cursor and the
relevant harmonics have been displayed.
– Multiple harmonics of 1xTs are shown up to around 10 orders of 1xTs.
40 - Kiln Main Drive

M4441 -G2H Shaft 01 Outboard Horizontal
4.0
Route Spectrum
06-Nov-02 11:02:11
3.5
OVERALL= 5.22 V-DG
RMS = 5.22
LOAD = 100.0
3.0
RPM = 635. (10.58 Hz)
2.5
2.0
1.5
1.0
0.5
0
Freq: 10.58
0 200 400 600 800 1000 Ordr: 1.000
Frequency in Hz Spec: 3.088
Looseness Spectral Data (Component)
• The spectral plot shown is demonstrating rotational
Looseness.
– The primary cursor is on 5xTs peak
• The 5 Order peak is vane pass frequency (5 vanes on the impeller)
– Multiple harmonics of 5xTs are shown indicating the impeller has
come loose. Ex 9
L1 - Example 9
-P2A Pump Outboard Axial
1.5
Route Spectrum*
17-Aug-01 08:52:02
1.2
 The raised noise level around
OVERALL= 6.62 V-DG
RMS = 6.13
the vane pass frequency is

LOAD = 100.0
RPM = 2974. (49.57 Hz)
common to a pumping
0.9
problem known as Cavitation

0.6
– This would be the likely cause
of the impeller problem
0.3
0
Freq: 14.88
0 40 80 120 160 200 240 Ordr: 5.002
Frequency in kCPM Spec: .742
Label: Centrifugal Pump - Medium
Looseness Waveform Data
• Here the waveform is demonstrating a lot of energy and
appears to be more random and non-periodic.
– Displaying the waveform in velocity may help to show the random
non-periodic pattern.
1.2
Route Waveform
06-Nov-02 11:02:11
0.8 RMS = .3174

LOAD = 100.0
RPM = 635. (10.58 Hz)
PK(+) = .9797
0.4 PK(-) = .9874
CRESTF= 3.11
Acceleration in G-s
0.0
-0.4
-0.8
-1.2
0 50 100 150 200 250 300 350 400
Time in mSecs
Looseness Trend Data
• Here the trend plot is showing the parameter labelled as the 3-
15xTs. This is measuring the amount of energy from 3 orders to
15 orders, which is where the harmonics of looseness will
appear.

8
Trend Display
of
7 3-15xTS
-- Baseline --
6
Value: .837
Date: 28-Feb-02
FAULT
4
ALERT
3
0
0 10 20 30 40 50
Days: 28-Feb-02 To 16-Apr-02
Case Study 4 – Reciprocator Fan
• Introduction
• Data had been collected on the following fan for several
months as part of a routine periodic vibration routine. During a
routine visit to the machine it was observed that there was a
lot of low frequency activity showing around the bearing on
the inboard of the fan (F1H)
• The multiple plots shown above indicate the change over time
from the data taken on F1H.
– It is quite apparent that the data shown here is indicating multiple
harmonics of the 1xTs frequency (the rise energy as you move further
away from the 1xTs).
– This type of data is common to that of a looseness problem.
40 - Precip Fan
M2237 -F1H Fan Inboard Horizontal
Max Amp 2.4

2.74 2.0
1.6
1.2
0.8
0.4
0
29-Oct-02
11:00:02
18-Sep-02
09:13:26
29-Aug-02
15:30:18
29-Oct-02
11:00:02
RPM= 998.9
22-Aug-02
11:14:48
Freq: 16.65
0 300 600 900 1200 Ordr: 1.000
Frequency in Hz Sp 4: 2.811
• The waveform data taken for this particular point is not
showing a random type of waveform pattern which you
would expect from Structural looseness, but there is a more
a repeatable (timed interval) pattern.
40 - Precip Fan
3
Analyze Waveform
18-Sep-02 09:24:16
2 RMS = .3747
LOAD = 100.0
RPM = 998. (16.63 Hz)
1
PK(+) = 2.36
PK(-) = 2.83
CRESTF= 7.55
Acceleration in G-s
-1
-2
-3
-4
0 100 200 300 400 500 600 700 800
Time in mSecs
• This type of waveform would more be indicating Component
looseness and may indicate a problem with a loose bearing.
• Conclusion
• It was recommended that the bearing should be inspected at
the next available opportunity.
– Upon inspection it was found that the bearing was a ‘Taper-Lock’
bearing and the taper lock was loose, thus resulting in excessive
clearance between the bearing and the rotor.
Looseness Summary
• Diagnostic Rules for Looseness

– Multiple harmonics of the 1xTs peak
– Number of harmonics will increase as the looseness
progresses
– Random, non-periodic waveform – Structural
– Waveform shows predominant impacts - Component
– Raised noise level around the 1xTs + harmonics
– Half harmonics may also be present
– Can be present in all Directions
Fault Diagnostics
Gears, Bearings, Peakvue, Electrical, Belts,
Resonance
Gear Defects
• There are many different types of gears and gear combinations
available for various speed and power requirements.
• Regardless of gear type they all produce the same basic
vibration patterns and characteristics when a defect is present
• The following topic will discuss the

basic characteristics for the following
types of gears:
– Spur Gears
– Helical Gears
– Bevel Gears
Spur Gears
• Spur Gears are most commonly thought of when diagnosing
gears. The teeth are cut parallel to the shaft. These gears are
good at power transmission and speed changes but are noisier
than other gear types.
• Spur Gear Advantages

– High efficiency
– Low heat generation
• Spur Gear Disadvantages
– Can be very noisy
Helical Gears
• Helical Gears have teeth cut at an angle to the shaft. These
gears are much quieter than spur gears but due to the angular
nature of the gear meshing, axial thrust and therefore axial
vibration is higher than those of spur gears
– Sometimes to counter act the axial thrust these gears can be double up
and are known as ‘Double Helical’ or ‘Wishbone Gears’
• Helical Gear Advantages
– Quiet Operation
• Helical Gear Disadvantages
– Less power transmission efficiency
and greater heat generation than
spur gears
– Axial loading on bearings
Bevel Gears
• Bevel Gears are used to transmit power and speed to an
output shaft perpendicular to the drive shaft. These gears use
a bevel design to transmit the power better.
– These gears are most commonly seen on right angle gearboxes (where
the input shaft is at 90 degrees to the output shaft)
• Bevel Gear Advantages

– Converts the direction of power transmission
• Bevel Gear Disadvantages
– Less efficient
– Higher heat generation
Gear Analysis
• Vibration analysis of gears can provide a wealth of information about the
mechanical health of the gears. This section discusses the basic frequencies
that may be present within a gearbox.
• Gear Mesh Frequency Spectral Data

• The gear mesh frequency (GMF) refers to the frequency at which to mating
gears interact with each other and is the most commonly discussed gear
frequency.
• However, GMF by itself is not a defect frequency. The GMF should always
be present in the spectral data regardless of gear condition. What is
important is the amplitude as this may vary depending upon gear condition
or loading of the gear.
Gears
• Two mating gears will generate a frequency known as the
GMF and will show in the spectral data regardless of gear
condition.

M4441 -G1V Shaft 01 Inboard Vertical
1.2
Route Spectrum*
08-Jun-02 23:11:51
OVERALL= 2.22 V-DG

RMS = 2.14
LOAD = 100.0
0.9
RPM = 1548. (25.80 Hz)

0.6
0.3
0
Freq: 386.98
0 200 400 600 800 1000 Ordr: 15.00
Frequency in Hz Spec: .864
Calculating GMF – Single Reduction
• Single Reduction Gear Train
– The GMF is simply defined as the number of teeth on a gear multiplied
by its turning speed
GMF = (#teeth) x (Turning speed)
• Example:
– Consider the following gear train,
INPUT Input = 1490RPM
Gear 1= 44 Teeth
OUTPUT
Gear 2= 71 Teeth
GMF = #teeth x turning speed
GMF = 44teeth x 1490 RPM
GMF = 65560 CPM or 65560/60 = 1092.6 Hz

Calculating GMF – Multi Reduction
• Calculating the GMF for gearboxes that have multiple trains use
the following.
GMF = (#teeth) x (Turning speed)
Gear Ratio = (#teeth in) / (#teeth out)
Speed out = (Speed in) x (Gear Ratio)
• Example:
– Consider the following gear train:
Input = 1490RPM
INPUT
Gear 1 = 15 teeth
Gear 2 = 21 teeth
Gear 3 = 19 teeth
Gear 4 = 54 teeth
OUTPUT
Calculating GMF – Multi Reduction
Input = 1490RPM
INPUT
Gear 1 = 15 teeth
Gear 2 = 21 teeth
Gear 3 = 19 teeth
Gear 4 = 54 teeth
OUTPUT
Gear Ratio 1 = 15 teeth / 21 teeth = 0.714

Speed Out = 1490 RPM x 0.714 = 1064.28 RPM
Gear Ratio 2 = 19 teeth / 54 teeth = 0.351

Speed Out = 1064.28 RPM x 0.351 = 374.47 RPM
GMF 1 = 1490 RPM x 15 teeth = 22350 CPM

GMF 2 = 1064.28 RPM x 19 teeth = 20221.32 CPM
GMF Calculation Exercise
• Using the formulas on P153 from the manual calculate:
– Speeds of all shafts
– All GMF from the following gearbox arrangement
Input = 1000 RPM

INPUT
Gear 1 = 10 teeth
Gear 2 = 40 teeth
Gear 3 = 10 teeth
Gear 4 = 20 teeth
OUTPUT
• Gear Ratio 1 = 10/40 = 0.25

• Shaft 2 speed = 1000 x 0.25 = 250 RPM
• Gear Ratio 2 = 10/20 = 0.5
• Shaft 3 Speed = 250 x 0.5 = 125 RPM
• GMF 1 = 1000 x 10 = 10000 CPM
• GMF 2 = 250 x 10 = 2500 CPM
Gears – Sideband Frequencies
• Sidebands are the most common indication that a gear is
defected.
– Sidebands are equally spaced frequencies in the spectral data that
materialise either side of the main GMF peak.
– The sideband frequency spacing is equal to either the turning speed of
the input gear or the turning speed of the output gear.
• Sidebands show in the data when either the

gear is worn, loose or eccentric.
– The speed of the shaft with the bad gear on it will
produce the most dominant sidebands in the
spectral data.
Gears
• The spectral data shows GMF with sideband data.
– The sidebands are equally spaced at intervals of 310 CPM. This is
indicating the gear that rotates at 310 RPM is the one that is worn or
damaged.
FPP - SAND MILLS (OLD)A
X401A -G3A Shaft 02 Inboard Axial
1.0
Route Spectrum
07-Nov-02 09:11:53
(SST-Corrected)
0.8 OVERALL= 2.18 V-DG

RMS = 2.17
GMF LOAD = 100.0
RMS Ve locity in m m /Sec
RPM = 310. (5.17 Hz)
0.6
0.4
Sidebands
0.2
0
Freq: 18363.
0 8000 16000 24000 Ordr: 59.23
Frequencyin CPM Spec: .564
Dfrq: 310.82
Gears – Waveform Data
• Gears can produce different types of waveforms, the one
shown below is indicating gear wear.
– As the defective teeth come into mesh the noise generated increases
showing an increase in amplitude in the vibration data
FPP - SAND MILLS (OLD)A
X401A -G3A Shaft 02 Inboard Axial
1.5
Route Waveform
07-Nov-02 09:11:53
1.2
PK = .4580
0.9 LOAD = 100.0
RPM = 311. (5.19 Hz)
0.6 PK(+) = 1.27

Ac c e le ra tio n in G -s
PK(-) = 1.13
CRESTF= 3.91
0.3
-0.3
-0.6
-0.9
-1.2
-1.5
0 1 2 3 4 5 6
Revolution Number
Case Study 5 –Gearbox
• The following case study is from a motor gearbox unit that
drives a roller.
– Product (Fibre) is fed along the top of the roll while being washed
through a series of baths.
– There are several of these Wash Nip Rollers in a continuous stream,
failure of any one of them results in lost production
• Data is collected on a
fortnightly basis as part of
a routine data collection
route
• The spectral data shown below is taken from the motor in the
axial direction
– (As the motor is mounted directly into the gearbox the first helical gear
is mounted on the end of the motor shaft).
L1NG - WASH LINE NIP UNIT 3
3-32J03 -MIA MOTOR INBOARD AXIAL
0.6
Route Spectrum
• The GMF is highlighted by

EEEEE EEEEE 01-Aug-04 10:21:41
0.5 OVERALL= 1.08 V-DG

RMS = 1.07
the primary cursor at 49 LOAD = 100.0
RPM = 1175. (19.58 Hz)
Orders
0.4 >REN Wash Nip

E=Gm(1>2)-S1
• The fault frequency data
0.3
(dotted lines) indicate the
sideband data showing 0.2
gear wear on the first gear

in the gear train 0.1
0
Ordr: 49.00
0 10 20 30 40 50 60 70 80 Freq: 57551.
Frequency in Orders Spec: .275
Dord: .00649
• The waveform data is showing a distinct pattern commonly
associated with gears.
• The amplitude increases In noise as the damaged teeth come
into mesh
– Producing over 2G-s of force in both the positive and negative direction
• The gears were inspected due to the critical nature of the
asset. It was found the gear to be severely damaged.
• A new gearbox was fitted and new data was taken showing the
difference between the good and bad gear
Bearing Defects
Rolling Element
Plain Bearings
Peakvue
Rolling Element Bearings
• Rolling element bearings have specific bearing failure modes
that can be observed in the spectral and waveform data.
• Bearing frequencies differ from most other frequencies

present within the spectral data because unless the bearing
has a defect there will be no frequency peaks in the data
relating to the bearing. Only if the bearing has a defect will
frequencies show in the spectral data.
 There are four main fundamental bearing defect
frequencies these are:

Outer Race
Inner Race
How Bearing Faults Generate Vibration
How Bearing Faults Generate Vibration
• Bearing defect frequencies are calculated based upon the geometry of the
bearing these calculations may include:
– Number of rolling elements
– Pitch Circle Diameter
– Rolling element diameter
– Contact angle
• Defined within Machinery Health Manager there are over 100000 predefined bearing
stored in the CSI bearing warehouse
BEARINGS in CSI Warehouse:

c:\RBMsuite\SysData\CSI_CMP.WH
****************************************************
BRG ID Bearing Type #B/R FTF BSF BPFO BPFI
12143 RHP 6218 11 0.418 2.967 4.598 6.402
24421 SKF 6313E 8 0.376 1.894 3.009 4.991
25372 SKF I-26313 19 0.433 3.568 8.219 10.781
• Characteristics of Bearing Defects
– High frequency raised noise level (Hump of energy)
– Non-Synchronous harmonic peaks (Both low and high
frequency)
– Time waveform will show a lot of noise/impacting
– Early stages of bearing wear may show better if viewed in
acceleration in the frequency domain
– Fundamental bearing defect frequency (First calculable
frequency) may not be present in the spectral data
Failure Mode 1
• The early stages of bearing defects produce low
amplitudes of vibration at higher frequencies
– (Appears on the right hand side of the spectrum).
• These are normally humps of energy or peaks
that are harmonics to the fundamental
frequency.
– (The fundamental frequency should not be visible
at this stage).
Failure Mode 2
• Distinct harmonics of Non-Synchronous peaks
appear.
– (These should appear lower down the scale of the
spectrum – towards the left / middle of the plot)
• Sidebands may appear around these frequencies
usually equating to turning speed.
– (The fault frequencies may not match exactly with
the peaks in the spectrum due to the fact that the
bearing geometry will have changed).
Failure Mode 3
• The fundamental frequency normally appears at this stage
– (First calculable frequency of the bearing – towards the left-
hand side of the spectral plot). This is classed as advanced
stages of bearing wear.
• Sidebands may be visible that equate to other bearing
frequencies – BSF, FTF etc).
Failure Mode 4
• The bearing degrades so much that the spectrum

becomes a mass of noise. At this point the bearing
will fail at any point (If it last this long – most fail
around Mode 3).
Rolling Element Bearings - BPFI
• Typical data showing a defected inner race
– Fundamental frequency showing
– Harmonics low and high frequency + sidebands
Rolling Element Bearings - BPFO
• Data showing a defect related to the BPFO
– The fundamental frequency is showing
– Harmonics from low to high frequency
Rolling Element Bearings - BSF
• Bearing defect showing the BSF – Rolling elements
– Sidebands around the BSF = FTF
Rolling Element Bearings - FTF
• The FTF is the only bearing frequency that is sub-synchronous
– May not detect then with conventional vibration data
– FTF defect at 0.4 orders shown in Peakvue
• Bearing
FTF & BSF
BPFI & BPFO
Rolling Element Bearings - Waveform
• As a bearing becomes defected then the amount of
noise/force generated as the rolling elements impact the
defective area increases.
– This can show significant G-levels in the time waveform. This value is
trended in the software as the Peak-Peak value
• This data is taken from a

pump with a damaged
bearing
– The force levels are
reaching 40G-s
Case Study 6 – Bearing Defect
• The spectral plot below is showing the data from the
inboard vertical direction of the motor.
– The primary cursor is indicating the fundamental defect
BPFO frequency + harmonics.
– The frequency range of the harmonics covers both low
and high frequency ranges suggesting the bearing is more
advanced stages of failure.
• The time waveform is showing significant impacting levels
reaching in excess of +/- 8G-s of force.
– This level of impacting is higher than would be suspected for a motor of
this type.
• The repetitive impacting

pattern shown above is
common to antifriction
bearing defects.
– In this instance the impacting
is representing the rolling
elements striking a defect on
the race.
• The trend plot above is showing the increase in amplitude of
the Peak-Peak parameter.
– The peak-peak parameter is measuring the amount of energy in the
time waveform from the Peak+ to the Peak-
• Conclusion
• The motor was reported as having a bearing
defect to the engineering group. As the
fundamental defect frequency was present
and the trend had shown sudden increases it
was recommended to change the bearing at
the next available opportunity.
Bearing Defects
Rolling Element
Plain Bearings
Peakvue
Plain Bearings
• Rotating elements are not used in sleeve (plain) bearings;
rather the shaft rides on a layer of lubricating oil inside the
bearing journal.
– Therefore the fundamental frequencies seen from antifriction bearings
do not apply to sleeve bearings.
• Since there is no contact between the

bearing and the shaft monitoring of
sleeve bearings for vibration analysis
usually requires the use of
displacement probes mounted 45
degrees either side of top dead centre.
Plain Bearings
• As there are no rotating components in the bearing that
produce high frequency noise (force) there is no need to
monitor a high frequency range. Usually 10 to 15 orders of
turning speed will be sufficient.
• Sleeve bearings have specific defects that contribute
towards bearing failure, these are:
– Excessive clearance
– Hydraulic instability (oil whirl)
Plain Bearings – Spectral Diagnostics
• Excessive Clearance
– When there is excessive clearance between the rotor and the bearing
then this will have an effect on the system vibration. When the
bearings have excessive clearance then a ‘looseness’ occurs.
• The spectral data shown below is indicating a sleeve bearing with

excessive clearance. Fu - Turbine Brg Thrust End

16 As the clearance increases then the
TBT -R1Y Radial 'Y' Direction
Route Spectrum*
harmonics of 1xTs will increase and can 27-Jul-04 14:08:21
go up to 10–15xTs. OVERALL= 2.93 V-DG

P-P = 22.71
12 – Like looseness the more harmonics
LOAD = 100.0
RPM = 941. (15.69 Hz)
there are the more severe the problem
P-P Displacement in Microns
will be.
– A good sleeve bearing will still show
8
a few harmonics as there is a small
clearance between the shaft and
bearing
4
0
Ordr: 1.000
0 3 6 9 12 Freq: 15.68
Frequency in Orders Spec: 7.494
Plain Bearings – Spectral Diagnostics
• Oil Whirl
– One of the major problems encountered with these types of bearings is
the possibility of hydraulic instability of the shaft within the bearing;
known as oil whirl or oil whip.
– Oil Whirl is a result of turbulent flow within the oil resulting in the oil
pushing the shaft around of centre.
Fu - Turbine Brg Thrust End • The dominant peak within the spectral
TBT -R1Y Radial 'Y' Direction
16
Route Spectrum* data will be typically at 0.4 orders.
27-Jul-04 14:08:21
(.40-.48)
Oil Whirl at 0.4 orders OVERALL= 2.93 V-DG
P-P = 22.71 – This defect is sub-synchronous data.
LOAD = 100.0
12
RPM= 941. (15.69 Hz) – When the amplitude of the oil whirl is
equal to or greater than the 1xTs peak a
P-P Displacement in Microns
problem exists
8
• In this instance oil whirl can be
corrected by:
– Properly loading the bearing
4
– Change the oil viscosity
– Change the oil pressure
0
Ordr: 1.000
0 3 6 9 12 Freq: 15.68
Frequency in Orders Spec: 7.494
Oil Whirl
Bearing Defects
Rolling Element
Plain Bearings
Peakvue
Peakvue Processing
• The detection of bearing and gear defects is one of the primary
expectations of a predictive maintenance program.
– As analysts we can spend a lot of time tying to determine these faults.
– Peakvue is a process that concentrates on these defects to help the analysts
determine potential faults developing
• Peakvue stands for the Peak Value and is a technique that detects high
frequency stress waves generated from metal to metal contact, such as:
– Bearing defects – Rotating elements striking a defect on the race
– Gear defects – Damaged teeth in mesh
– It is the detection of these high frequency stress waves that will aid with
analysis
Peakvue Processing - Filters
• In order to capture the stress wave signal the process requires
the use of a filter to remove all unwanted noise that can
dominate the data
1. Conventional Vibration 2. Peakvue filter removing

Signals that are filtered low frequency noise from
from the Peakvue Signal the stress wave data
· Imbalance · This is to prevent low
· Misalignment frequency noise
· Gears consuming the stress
· Bearings wave activity
· Resonance
3. High frequency stress wave

activity occurring in the 1000Hz -
20000Hz frequency range at a
rate governed by a low
frequency event
· Bearings
· Gears
Peakvue Processing - Filters
• There are two types of filters available
• Band Pass Filters
f
– The band pass filter removes all the data above and below the filter
corner values
• High Pass Filter f

– The high pass filter removes all data lower in frequency to that of the
filter selection allowing only the high frequency stress waves to pass
through
• After the filtering process what should remain is the high
frequency stress wave activity that is occurring at the rate of
the excitation – such as from a bearing.
Peakvue Processing – Spectral Data
• Shown below is a typical Peakvue spectrum with a defect
present
 Stress waves are showing • The filter used is shown in the top
clearly in the data at 4.6 Orders right hand corner
 Good Spectrum will

show only a noise
level
 Noise removed by
filter
Peakvue Processing – Waveform Data
• As stress waves are small in amplitude severity of the problem
can be judged using the time waveform
– Peak Value of force from the impact
• The waveform can resemble a spectrum as there is no
negative half to the data
B42 - ZONE 5 DF FAN 1
16/16EXT01-M2P Motor Inboard Horz Peakvue
0.8
0.7 Route Spectrum
N N N N N N N N N
RMS Acceleration in G-s
09-Jul-03 09:50:49
0.6 (PkVue-HP 1000 Hz)
OVERALL= 1.37 A-DG
For Peakvue analysis
0.5
0.4
RMS = 1.37 
LOAD = 100.0
0.3 RPM = 1342. (22.37 Hz)
0.2
>NTN 6217
N=BPFO -OB
 Use the Spectrum
0.1
0 – Diagnose the defect
0 200 400 600 800 1000
Frequency in Hz  Use the Waveform
8 Route Waveform
7
09-Jul-03 09:50:49
(PkVue-HP 1000 Hz)
– Determine the severity
Acceleration in G-s
6 RMS = 2.97
5 PK(+) = 8.35
4 CRESTF= 2.81
3
2
1
0
Freq: 1.250
0 4 8 12 16 20 24 28 32 36 Ordr: .05587
Revolution Number Spec: .01367
Label: Bearing Fault - BPFO NTN6217
Case Study 7 – Peakvue on Fan Bearing
• The following machine is a
pre-heater fan designed to
heat the product prior to it
entering a Kiln
– There is no standby for this
machine
– Failure results in stopped
production
• The following data was taken from the above fan unit.
– The problem bearing resided on the fan inboard bearing.
– Data was collected on a monthly basis. Both conventional vibration
data and Peakvue data were taken during the route collection.
• The data shown below is taken using conventional vibration
methods on the inboard bearing of the fan
– 1x peak is highlighted showing amplitudes of 4mm/sec
– Waveform is showing less than 1G of force both +/-
40 - Preheater Fan
5
Route Spectrum
4 29-Oct-02 11:19:26
OVERALL= 4.18 V-DG
RMSVelocity in mm/Sec
RMS = 4.18
3
LOAD = 100.0
RPM = 825. (13.75 Hz)
2
1 • There are indications of

0
bearing frequencies
0 10 20 30 40 50 60 70 80
Frequency in Orders showing high frequency
1.5
1.0
Route Waveform
29-Oct-02 11:19:26
– These may be missed due to
0.5
RMS = .3837 the amplitude of the 1x peak
PK(+/-) = 1.19/1.05
Acceleration in G-s
CRESTF= 3.11
0
-0.5
-1.0
-1.5
Ordr: 1.000
0 1 2 3 4 5 Freq: 13.75
Revolution Number Spec: 3.721
• The Peakvue data above is taken from the same point as the
previous data.
– This particular reading is using a 1000 Hz High Pass filter.
40 - Preheater Fan
M4425 -F1P Fan Inboard Horz Peakvue
0.7
Route Spectrum
0.6 F F F F
0.5
29-Oct-02 11:15:59
(PkVue-HP 1000 Hz) • Here the data is showing there
RMSAcceleration in G-s
OVERALL= 1.10 A-DG

0.4 RMS = 1.10 is stress wave activity at 8.176
LOAD = 100.0
0.3
RPM = 830. (13.84 Hz) orders.
0.2
0.1
>SKF 22240CC
F=BPFO -IO – This is not non-synchronous
0
data and the frequency
0 5 10 15 20 25 30 35 40
matches that of the BPFO for
Frequency in Orders the bearing.
8
7 Route Waveform • The waveform data is
29-Oct-02 11:15:59
6 (PkVue-HP 1000 Hz) measuring over 7 G-s of force
RMS = 3.31
Acceleration in G-s
5
4
PK(+) = 7.47 as oppose to the 1G from the
3
CRESTF= 2.25
DCoff = -3.08
previous data.
2
1
0
Ordr: 8.176
0 10 20 30 40 50 Freq: 113.14
Revolution Number Spec: .194
• Conclusion
• There is significant bearing damage relating the outer race of
the bearing.
• As the machine was critical to the process, the bearing was
changed on the next available opportunity that tied in with
process requirements.
Electrical Defects
Electrical Defects
• A motor can be simply broken down into two key components
– Rotor
– Stator • The stator is stationary
– Consists of wire wound in coils
and placed in slots of an iron
core.
– The stator produces a rotating
magnetic field.
 The rotor is not stationary

– Consists laminations with solid conductors called rotor bars
– A circular flow of current through these rotor bars causes
the rotor to become an electromagnet which will rotate in a
magnetic filed.
Electrical Defects – Spectral Data
• The most common electrical frequency that materialises in the
spectral data is the 2 x Line Frequency.
– For most industrial applications the line frequency used to supply
motors is 50Hz (Europe).
– Therefore the frequency of concern for most electrical faults would be
100Hz (2xLf [Lf=line frequency])
Ex7 - Example 7
Ex7 -M1H Motor Outboard Horizontal
0.6
Route Spectrum
08-Nov-00 14:27:35
0.5 OVERALL= .5613 V-DG

RMS = .5607
• The spectral plot is
LOAD = 100.0
0.4
RPM= 2967. (49.44 Hz)
showing a peak at 100Hz
(6000cpm)
0.3
– 2xLf
0.2
– This can be mistaken for
misalignment
0.1
0
Freq: 100.00
0 500 1000 1500 2000 Ordr: 2.023
Frequency in Hz Spec: .386
Electrical Defects – Waveform Data
• The waveform data from a 100Hz peak will show a sinusoidal
pattern like the waveform shown below
Ex7 - Example 7
• Again this type of pattern

Ex7 -M1H Motor Outboard Horizontal
1.5
Route Waveform
08-Nov-00 14:27:35
1.0 RMS = .5291

LOAD = 100.0
can be associated with
RPM= 2967. (49.44 Hz)
0.5
PK(+) = 1.50
misalignment.
PK(-) = 1.77
CRESTF= 3.31 – Usually misalignment would
Velocity in mm/Sec
0
produce higher force (Higher
-0.5 waveform levels) than those
from electrical defects due to
-1.0
the stress being applied to
-1.5 the machine
-2.0
0 1 2 3 4 5 6
Revolution Number
Electrical Defects - Causes
• Common fault types that can produce the 2xLf peak are as
follows:
• Dynamic Eccentricity – Usually Rotor Related
• Static Eccentricity – Usually Stator Related
• Loose Iron or Slot Defect – Rotor or Stator
• Open or Shorted Windings
• Insulation Breakdown or Imbalanced Phase
• Loose Connectors
Electrical Defects - Peakvue
• Peakvue data also shows electrical defects at the 2xLf peak.
– This may be due to the rotor or stator bowing; due to heat build up.
• The spectral plot below is indicating a 100Hz peak using
Peakvue with a 1000Hz filter.
Case Study – Electrical Defect
• The following case study was taken from a glass manufacturer.
The data was from the ‘Electric Front Wall Cooling Fan’.
– This fan unit is a critical fan to the process and has no standby unit.
– In this particular instance the motor failed shortly after the data was
collected.
• The Peakvue data taken on

the motor non-drive end is
showing a dominant 100Hz
peak.
– This frequency is at 2xLf and
is associated with electrical
problems
Case Study – Electrical Defect
• The multi-plot above shows the same measurement point
going back over the last 5 route readings.
– This particular plot is useful for determining rate of change.
– It is quite clear how this particular frequency suddenly appeared
• Conclusion
– As the motor failed shortly after
data collection no action was
taken to prevent failure.
– The investigation in the motor
showed one of the connectors had
come loose causing the motor to
burn out.
Belt Defects
V-Belts
Timing Belts
Belt Defects
• Belts are the most common low cost way to transmit power
from one shaft to another.
– Belt drives rely on friction between the belt and pulley to transmit
power between drive and driven shafts
• The ability of belt to transmit power depends upon
1. Belt Tension (tension on the belt holds it tightly against the sheave)
2. Friction between the belt and sheave
3. The arc of contact between the belt and sheave (Wrap)
4. The speed of the belt
• However, belts can be easily damaged by heat, oil and

grease and since belts slip with in the sheaves they can not
be used where exact speed changes are required (except for
timing belts)
Belt Defects
• Belt defects can be considered non-critical faults by many
maintenance groups due to the relative ease of replacement
requiring minimum downtime.
– But belt defects are a major contributor to the overall vibration of the
machine resulting in premature failure of other machine components.
Sources of belt drive defects

Poor Maintenance
Enviromental Factors
Poor Installation
Poor Design
Other Defects
Belt Defects – Belt Types
• There are many different types of belt drive systems. This
section covers the most commonly used types of belt in
industry today.
• V-Belts
• V-belts are the most common type of belts used. They are ‘V’ shaped in
cross-section, this allowing the belt to wedge against the side of the
sheave.
– This design allows the belt to be run faster than most other type of belt
applications with power transmission efficiencies as high as 95%
Belt Defects
• Timing Belts
• These are flat belts with equally spaced teeth that mesh
with notches on the pulley. Timing belts are different from
other belt drives as they do not induce any slip.
– Most commonly used where constant velocity and strict timing
application is required.
Belt Defects – Fault Characteristics
• Belt defects, such as cracks, broken or missing pieces, hard and
soft spots can generate vibration at the turning speed of the
belt (1xbelt) and harmonics
– Due to the length of the belt in relation to the pulleys (sheaves) the
1xbelt frequency is sub-synchronous and very often the 2xbelt
frequency may be sub-synchronous as well
• The predominant harmonic is typically the 2xBelt frequency
and can be seen in the radial plain in-line with the belts.
– Severity is judged by the number and amplitude of the harmonics
seen in the spectral data
Belt Defects – Fault Characteristics
• Just like two mating shafts, belt drive systems can also be
misaligned in both angular and offset directions.
– When misalignment is induced into a belt drive system then the life of
the belt is significantly reduced as well as the overall vibration of the
system increases.
Offset Misalignment
Angular Misalignment
• Pulley misalignment results in high axial vibration at the

shaft turning speed.
– If the belt is also defected then 1xbelt frequency and harmonics may
also show in the axial direction
Belt Defects – Calculations
• The fundamental belt frequency can be calculated using the
following equation:
Belt Freq. = (3.142 * Pulley Ts * Pulley PCD)
Belt (Length)
– Where:
• Ts = Turning Speed
• PCD = Pitch Circle Diameter
• Note: The PCD and belt length must be in the same units
• A timing will belt will also have a specific frequency related to

the number of teeth on the pulley
Timing Belt Freq. = (Pulley Ts) * (# Pulley Teeth)
Belt Defects – Calculation Example
• Belt Frequency Calculation
• Belt Frequency = (3.142 * 1480 * 300) / (2000)
• Belt Frequency = (1395048) / (2000)
• Belt Frequency = 697.524 CPM
– This is sub-synchronous to the 1xTs of the pulley
Motor RPM = 1480 RPM

Pulley Diameter = 300 mm
Belt Length = 2000mm
Belt Defects – Spectral Data
• The spectral data above is data taken of a motor from an Air
Handling Unit.
– The frequency highlighted by the primary cursor is showing the 1xTs of
the motor (1 Order)
• There are a lot of sub-
synchronous peaks showing in
this data.
– The first peak is the
1 x Belt Frequency fundamental frequency of the
showing with harmonics belt rotation.
– The second peak is the 2xbelt
Dominant 2 x Belt
Frequency
frequency suggesting there is
damage to the belt
– As the harmonics of the belt
increase in number they
surpass the 1xTs of the motor
and in this case the third
harmonic becomes non-
synchronous data.
Case Study 9 – Belt Defect
• The following data was taken on an Air Handling Unit. The Air
Handling Unit is a supply fan from shared services. This is a
stand alone unit with no stand by capability
BL31 - 559 AHU Supply
559S -M2H Motor Inboard Horizontal
•
0.5
J J J J J J J J J J
Route Spectrum*
22-Feb-05 13:53:33 The data shows the
OVERALL= 1.22 V-DG
motor turning speed
0.4 RMS = .7701
LOAD = 100.0
along with a sub-
RPM = 1272. (21.21 Hz) synchronous peak of
the belt frequency.
>Belt Freqs
J=Belt 1 Freq
0.3
• The primary cursor is
highlighting the 1xbelt
X - Motor speed
with several
x - Fan speed
0.2
harmonics.
• The 2xbelt is very
0.1 dominant suggesting
there is damage to the
belts.
0
Freq: 835.69
0 8000 4000 12000 16000 Ordr: .657
Label: Belt defect/worn belts & sheaves
Case Study 9 – Belt Defect
• As this is a critical machine it was recommended on the next
available opportunity that the belts needed to be checked for
damage and re-aligned.
• The machine was stopped and the belts were inspected based upon the
recommendation.
• Significant damage was found to several of the belts during this
inspection as well as worn pulleys. Both the belts and pulleys were
replaced and correctly aligned before re-starting the machine.
Resonance
Resonance
• Resonance is defined as:
An excitation of a natural frequency by a periodic forcing
function.
• All assets contain natural frequencies that vary depending

upon the stiffness and mass.
– Resonance can be considered to be a vibration amplifier, that takes the
force level of the periodic forcing function and amplifies it; which
significantly increases the movement of the asset.
If Vibration is a Fire, The Resonance is a Fuel
Example of Resonance
• The example shown represents the effect on amplitude of the forcing
function when in resonance.
– In plot 1 the 1xts is running below the natural frequency (Fn).
– Fn can be seen in plot 2.
– Plot 3 shows the increase in amplitude of the forcing function when run at the
natural frequency – this is resonance
Before Excitation
Frequency
Resonance Curve
Frequency
Amplified Signal
Frequency
Resonance
• There are two factors that determine the natural frequency of
an asset these are;
1. Mass – The heavier an object the lower the natural frequency
2. Stiffness – The more rigid a structure the higher the natural frequency
• Resonance is becoming more of a problem in industry in

recent years due to:
– Older equipment having to run faster to meet current production
demands (often above what it was designed for)
– Equipment is being built cheaper and lighter
• This is resulting in amplification of the forcing function
creating excessive machine movement resulting premature
machine failure.
Effects of Resonance
• The ODS data is showing a steel frame structure deflecting at
one corner in the vertical direction due to a resonant
condition.
Characteristics of Resonance
• Characteristics of Resonance
– Resonance is very directional in nature (Movement may be greater in
one plain than the other)
– Vastly different amplitudes of the forcing function from one direction to
the other (between Horizontal and Vertical – Rule of thumb ratio is 3:1
difference)
– Resonance is very speed sensitive (small changes in speed can show
large differences in amplitude of the forcing function)
– Resonance can occur at any frequency but most commonly associated
with the 1xTs
Resolving a Resonance
• There are a number of alterations to the system that can be
made to resolve a resonance condition.
– However if structural changes are to be made you need to be careful
you don’t excite another natural frequency once the change has been
made?
• Once you are sure you have a resonant condition it can be
corrected by one of the following methods:
– Change the Mass
– Change the Stiffness
– Remove the forcing function
– Dampen the structure
 Dampening is a method used to convert mechanical energy into
thermal energy. It does not remove the resonant condition only
controls the amount of movement.
Resonance – Spectral Data
• The spectrum is showing the 1xTs peak of the motor with amplitudes
reaching 19mm/sec.
– This is high for the 1xTs.
• Very often this type of data can be mistaken for Imbalance as this defect
can also produce a high 1xTs peak.
– However Imbalance is a centrifugal force and should show similar amplitudes in
both radial plains where as resonance is very directional.
40 - No 1 GCTCompressor
M4551 -M2H Motor Inboard Horizontal
27
• In order to help resolve this issue

Route Spectrum
13-Feb-03 10:14:46
24
OVERALL= 19.95 V-DG
21
RMS = 19.85
LOAD = 100.0
we need to check the amplitude
RPM = 1484. (24.73 Hz)
18 of the 1xTs 90 degrees to this
15 point (horizontal to vertical)

12 – This can easily be done by using
9 the ‘multi point plot’ in the
6 software
3
0
Freq: 24.72
0 500 1000 1500 2000 Ordr: 1.000
Frequencyin Hz Spec: 19.50
Resonance – Multi Plot
• The multi point plot allows the analyst to display several measurement
points on the same plot. Here we are showing all the radial points from the
motor.
– It is very clear that the amplitudes of the 1xTs peak are excessive in the
horizontal direction when compared to the vertical. This is a characteristic of a
resonant condition. 40 - No 1 GCTCompressor
M4551 - Multiple Points (13-Feb-03)
24
Max Amp 20
22.0
16
12
0
RMSVelocity in mm/Sec
M2V 10:15
M2H 10:14
M1V 10:14
Point= M2H
13-Feb-03
10:14:46
RPM= 1484.
M1H 10:14
Freq: 25.00
0 500 1000 1500 2000 Ordr: 1.011
Frequency in Hz Sp 3: 19.35
Case Study 10 – Resonance
• The following case study is taken from a motor and a
reciprocating compressor. The unit is mounted on a steel
frame which, in turn sits on spring mounts designed for
dampening
• Recently the motor had been replaced due to bearing
defect; however the new motor was smaller and lighter but
delivered the same power as the previous motor.
• When the compressor was put back into service it was
noted there was excessive vibration coming from the unit.
The unit was left to run like this for several months until the
vibration became to excessive.
• Data was taken across the unit using route based data
collection.
SL - Compressor
CP1 -M1H Motor Outboard Horizontal
60
Route Spectrum
02-Feb-04 15:09:54
50 OVERALL= 45.58 V-DG

PK = 45.32
LOAD = 100.0
RPM= 1490. (24.83 Hz)
40
PK Velocity in mm/Sec
30
20
10
0
Freq: 24.83
0 300 600 900 1200 1500 1800 Ordr: 1.000
Frequency in Hz Spec: 45.19
• The plot above is taken from the motor showing a 1xTs peak
in excess of 40mm/sec.
• This data is very high in amplitude.
• The data was then displayed in a multi plot format to show
how the amplitude was across the radial plains.
• Due to the vastly different amplitudes at the 1xTs frequency
the defect on this motor was Resonance.
SL - Compressor
CP1 - Multiple Points (02-Feb-04)
50
Max Amp
44.1
40
Amplitude differences between radial plains 30

20
10
0
M2H 15:26
M2V 15:26
0 4000 8000 12000 16000
Frequency in CPM
• Recommendation
• It was determined that the change in motor size may be the
cause of the resonance as the mass had been altered. A visual
inspection of the frame work also revealed that one of the
support beams had cracked along the weld – this altering the
stiffness of the structure. The support was welded and
strengthened and more data was acquired to determine if any
effect on the resonance had occurred.
• The spectra, shows the ‘Before’ and ‘After’ plot of the motor
inboard horizontal. It shows a significant drop in amplitude of
the 1xTs peak.
– By stiffening the structure the natural frequency had increased moving
it away from the 1xTs peak thus resulting in a significant drop in
amplitude. CP1
SL - Compressor
-M2H Motor Inboard Horizontal
50
Max Amp
44.1
40
30
20
10
0
07-May-04
10:08:05
02-Feb-04
15:26:38
0 1000 2000 3000 4000
Frequency in Hz
Summary of Faults
Resonance
Electrical
Imbalance
t
Misalignmen
Looseness Advanced Bearing Early Bearing Wear

Wear
Electrical Gearmesh Frequency
Lower Gearmesh
Severe Electrical Slot Pass Frequency
Frequency
Belt
Frequencies
Misalignment
Severe Looseness
c c c c c c c
Frequency
In Terms
Of RPM Most Likely Causes Other Possible Causes & Remarks
1 x RPM Unbalance 1) Eccentric journals, gears or pulleys

2) Misalignment or bent shaft - If high axial vibration
3) Bad Belts - If RPM of belt
4) Resonance
5) Reciprocating forces
6) Electrical problems
7) Looseness
8) Distortion - soft feet or piping strain
2 x RPM Mechanical 1) Misalignment - if high axial vibration
Looseness 2) Reciprocating forces
3) Resonance
4) Bad belts - if 2 x RPM of belt
3 x RPM Misalignment Usually a combination of misalignment and excessive axial
clearances (looseness).
Less than Oil Whirl (less 1) Bad drive belts
1 x RPM than 1/ 2 RPM 2) Background vibration
3) Sub-harmonic resonance
4) "Beat" Vibration
Synchronous Electrical Common electrical problems include broken rotor bars, eccentric
(A.C. Line Problems rotor unbalanced phases in poly-phase systems, unequal
Frequency) air gap.
2 x Synch. Torque Pulses Rare as a problem unless resonance is excited
Frequency
Many Times RPM Bad Gears Gear teeth times RPM of bad gear
(Harmonically Aerodynamic Forces Number of fan blades times RPM
Related Freq.) Hydraulic Forces Number of impeller vanes times RPM
Mechanical Looseness May occur at 2, 3, 4 and sometimes higher harmonics if
severe looseness
Reciprocating Forces
High Frequency Bad Anti-Friction 1) Bearing vibration may be unsteady - amplitude and frequency
(Not Harmonically Bearings 2) Cavitation, recirculation and flow turbulence cause random,
Related) high frequency vibration
3) Improper lubrication of journal bearings
(Friction excited vibration)
4) Rubbing
Vibration Fundamentals
How Much Vibration is Too Much ?
1. Use Absolute Vibration Levels

- Given by machine makers
- Published Vibration Severity Standards
eg. ISO 2372, VDI 2056, BS 4675
2. Use Relative Vibration Levels

ISO 10816-3
11 0.44
7.1 0.28
4,5 0.18
3,5 0.11
2,8 0.07
2,3 0.04
1.4 0.03
0,71 0.02
mm/s rms inch/s rms

rigid flexible rigid flexible rigid flexible rigid flexible Foundation
pumps > 15 kW medium sized machines large machines
radial, axial, mixed flow 15 kW < P  300 kW 300 kW < P < 50 MW Machine Type
integrated driver external driver motors motors
160 mm  H < 315 mm 315 mm  H
Group 4 Group 3 Group 2 Group 1 Group
A newly commissioned
B unrestricted long-term operation
C restricted long-term operation
D vibration causes damage
ISO 10816-3
140 5.51
113 4.45
90 3.54
71 2.80
56 2.20
45 1.77
36 1.42
28 1.10
22 0.87
18 0.71
11 0.43
µm rms mil rms

rigid flexible rigid flexible rigid flexible rigid flexible Foundation
pumps > 15 kW medium sized machines large machines
radial, axial, mixed flow 15 kW < P  300 kW 300 kW < P < 50 MW Machine Type
integrated driver external driver motors motors
160 mm  H < 315 mm 315 mm  H
Group 4 Group 3 Group 2 Group 1 Group
A newly commissioned
B unrestricted long-term operation
C restricted long-term operation
Vibration standards are guidelines
ISO2372 ( BS 4675 , VDI 2056 )
45
Not 28
Velocity mm/s RMS

2.5 times = 8dB
Not Not Permissible 18
Permissible
10 times = 20dB
Permissible 11.2
Just Tolerable 7.1
Just
Tolerable 4.5
Just Tolerable Allowable 2.8
Allowable 1.8
Allowable Good 1.12
Good Large Machines
15 kW<
Medium Machines
with rigid and heavy 1.71
foundations whose
Good <75kW
natural Frequency
0.45
Small <300 kW on special exceeds 0.28
Machines< 15 kW foundations machine speed
0.18
Group K Group M Group G
Setting Alarm Limits
Alarm criteria:
• The best alarm criteria will be based on prior data from the same machine,
tested under identical conditions.
• You can trend data
• You can compare to identical machines
• The goal here is to use a low cost, low tech method to see if overall vibration
levels are changing over time and then to take action based on an alarm.
• This type of monitoring may be done for a number of reasons, such as for
compliance with regulators.
Data collected:
• RMS overall levels
Alarm criteria:
• ISO or other standards
• Trending
ISO Alarm Limit
• The ISO 10816.1 General Machines severity chart has limits for four classes of
machines. The alarm limits are in Velocity and shows the values in both metric
and imperial units. Note that both are RMS values.
ISO 10816.1 Severity Chart: These are for overall values from 3 Hz to 1000 Hz.
Spectrum Alarm Limits
• The Overall value is a good number to trend. But it does not catch everything. It
does not let us know about small amplitude values that may indicate severe or
critical conditions.
• A better method is to have the system scan particular regions of the spectrum,
and compare against a level for that region.
• Which method should be used…Band, envelope, or statistical?
• Reference is still needed to compare against particularly when starting out.
• There are basically two ways to set a reference alarm level.
• The first is to utilize published alarm limits and set fixed alarm limits.
• The second is to start with existing vibration readings and perform a calculation
to derive the alarm limit.
• There are pros and cons to both approaches.
ISO 2372/10816.1 General Machinery Standard
• ISO limits are defined for overall limits. Because 1X vibration normally
dominates, it can be used for 1X limits.
• But limits are needed for the bearing frequencies, etc.
Band Alarms-1
• How fixed alarm limits and statistical calculations can be used to compute values
that can be applied to vibration limits.
• This is an area where the approach taken by the software package being used.
• In brief, there are "band alarms", "envelope (or mask) alarms", "expert systems"
and "artificial intelligence systems".
Six to twelve bands may be available for setting alarms. Bands are often defined by Analysis Parameter Sets.
Band Alarms-2
• Simply put, band alarms work on the principle that it is possible to consider
different portions of the spectrum, and different scalar measurements (overall
readings, bearing measurements, etc.) and apply different alarms to each band.
• For example, in a spectrum, one band could be created around the running speed
peak.
• The software will focus on the vibration level between 0.9X to 1.1X, for example.
• The software will look at the vibration in that frequency band and compute the
maximum level, an average level, the RMS level, or some other parameter and
see if it exceeds a limit.
• The same technique is also applied to other "scalar" data such as overall readings,
waveform pk-pk readings, crest factor calculations, bearings measurements and
other useful parameters.
Frequency Band Descriptions for Pumps
• One limitation of Band Alarms is that the bands are not necessarily sensitive enough.
• The RMS calculations are dominated by the highest peaks and harmonics and
sidebands may be ignored.
• Small peaks can grow next to big ones and the alarm will not be tripped.
• This is especially true of bearing frequencies.
Frequency Band Descriptions
Common settings include:

• Band 1: (sub-synchronous): 0.3 – 0.8x
• Band 2: (1x): 0.8x – 1.2x
• Band 3(synchronous) 1.2x – 3.5 x
• Band 4: 4x to N depending on the machine.
Envelope / Mask Alarms-1
• Envelope alarms (also known as mask alarms, and not to be confused with
envelope detection used in bearing analysis), take a different approach.
• Rather than breaking the spectrum into individual bands, an alarm limit is applied
to the entire spectrum.
• As seen in this example, it is not a single line across the spectrum; it is an

envelope that hugs the spectrum at all frequencies.
Envelope / Mask Alarms-2
• The benefit to this approach is that every frequency is covered, and it is

potentially more sensitive to peaks that can appear at unexpected frequencies.
• Whereas a single "band" may be used to cover a wide frequency range, from 1X
to 10X, for example, an envelope/mask may be computed to have up to 50
individual limits that follow the shape of the spectrum.
Envelope Alarm Weakness
• A weakness of the envelope alarms is demonstrated in above Figure.

• The alarm level envelope around the 6x vane pass was set based on the 6x peak.
• However, now a small peak at 6.2 orders has grown considerably.

• It will have a long way to grow before it reaches the alarm limit.
• If this peak is a bearing frequency, it may reach the final stages or even fail
before this peak reaches the alarm limit.
Thank You!!

Vibration Analysis Level - 1 Final

Uploaded by

Copyright:

Available Formats

Vibration Analysis Level - 1 Final

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Vibration Analysis Level - 1 Final

Uploaded by

Copyright:

Available Formats

What are the objectives of establishing the Centre for Vibration & Machine Condition Monitoring (CVCM) by PAEC?

What are the objectives of establishing the Centre for Vibration & Machine Condition Monitoring (CVCM) by PAEC?

What are some typical machinery problems that can be found using vibration analysis?

What are some typical machinery problems that can be found using vibration analysis?

Analysis Level I Course

• PAEC has established Centre for Vibration & Machine Condition

• The act of causing to continue

• Keeping equipment in repair

• It could be acceptable to deliver

• The desire is More Performance for Less Money!!!!

• To confirm good-condition machines

• To detect developing problems

• To determine the nature and severity of the problem

• To schedule repairs that can best fit with production and

• Motor current analysis

• And many more…

• Every mechanical or electrical faults on a machine has a

• Any change in the vibration signature indicates changes in the

• Establish a database of all the machines that need to be

• Establish a data collection route that best optimize the data

• Download route into the data collector

• Upload collected data into the database

• Run exception reports to detect the problematic machines

• Analyze only the machines in the exception reports

• Generate repair work to be performed

• Again collect data on the machine on which work is being done.

• Condition Based Maintenance or CBM is an activity that

• Maximising asset reliability is of the utmost importance in

Vibration: Oil: Thermography: Motor Current:

• The practical application of Tribology within

• Reduce maintenance costs

• Mechanical – Bearings, Rubs, Gear Defects etc

Bearing Steam Trap Air Leak

– This can be used in a wide array of circumstances but is most

• MCSA can be used as predictive maintenance tool for detecting common

 Early Fault Detection

 Clear Signs of Problem

 Diagnose Variety of Problems

Vibration is the pulsating motion of a body from

Vibration is an every day phenomenon,

• Heavy drill machines

• Ultrasonic cleaning baths

• ‘Vibration’ can be simply stated as ‘A response to some form of excitation’

• When a forcing function is applied

– Here we are measuring actual shaft movement

• Amplitude How Much

• Frequency How Often

Amplitude is the measure of magnitude of

• This motion is called a

– The vertical axis is

• The plot shown on the right

– Amplitude versus Time

• This is just one format -0.1

(domain) for analysing

• Data can also be analysed in -0.3

a ‘Spectrum’ – (Amplitude -0.4

process known as the FFT

• Raw vibration signal is rarely sinusoidal and forms a complex