Eng.

Islam Ramy

Introduction to Machinery Vibration Analysis

Eng. Islam Ramy Zakaria
Senior condition monitoring engineer
CAT IV, CMRP, MLE

Course Contents
• Introduction
• Maintenance strategies
• Condition monitoring technologies
• Vibration principles
• Data acquisition & transducers
• Understanding Signals
• Spectrum Analysis
• Time waveform
• Introduction to fault analysis
• Resonance & natural frequencies

1
 Eng. Islam Ramy

Introduction

Road to Mastery
The most reputed entities that introduces vibration
analysis course & certifications are :
• Mobius Institute

• Vibration Institute

• Technical associates of charlotte

2
 Eng. Islam Ramy

Road to Mastery
Vibration certification levels:
• CAT I (You can skip it & enters CAT II Directly)

• CAT II (Intermediate)

• CAT III (Advanced)

• CAT IV (Master)

Why do machine fails ?

• Poor design
• Poor maintenance practices
• Poor operating conditions
• Normal wear

o Failure starts with initial specifications & purchase,

balancing & alignment, routine maintenance, lubrication,
overhaul procedures and acceptance testing.

3
 Eng. Islam Ramy

Why vibration analysis ?

• Simply machines with faults that can ultimately result in
failure will vibrate in characteristic way the vibration
measurement can be studied for patterns to detect the fault
can be detected early to act.
• Vibration analysis is part of the condition monitoring &
predictive maintenance program.
• Majority of machinery faults can be detected by vibration
analysis.
• Measurements tells us about the nature of the problem &
severity of the fault.

Why vibration analysis ?

• Gives us enough information so that we can plan our
maintenance activities.
• Help in finding the root cause of the problem.
• Most machines are supposed to last longer than they do,
vibration analysis gives us the information so that we can
change the way machine installed, operated, maintained or
even designed so that it become more reliable.
• Vibration is the language of the machine if you learn it you
can under stand what your machine tells you.

4
 Eng. Islam Ramy

Why vibration analysis ?

when there is time to plan the maintenance the
following can be avoided:
• Catastrophic failure

• Down time

• Additional spare time costs

• Safety issues

• Secondary damage

Maintenance
strategies

5
 Eng. Islam Ramy

Maintenance
• The most important reason to have a maintenance
program with a structured approach is to ensure that
assets don’t fail prematurely, that they keep producing or
providing service as intended.

• Every industry is different from another & different

organizations have their own priorities & measures of their
success you have to understand your company vision so
that you can do your part to help in achieving company
goals.

Why Have a Structured Maintenance

Program ?
• Reducing production downtime — the result of fewer asset
failures.

• Increasing life expectancy of assets, thereby eliminating

premature replacement of machinery and asset.

• Reducing overtime costs and providing more economical

use of maintenance personnel due to working on a scheduled
basis, instead of an unscheduled basis, to repair failures.

• Reducing cost of repairs by reducing secondary failures.

When parts fail in service, they usually damage other parts.

6
 Eng. Islam Ramy

Why Have a Structured Maintenance

Program ?
• Reducing product rejects, rework, and scrap due to better
overall asset condition.

• Identifying assets with excessive maintenance costs,

indicating the need for corrective maintenance, operator
training, or replacement of obsolete assets.

• Improving safety and quality conditions.

Overall Equipment Effectiveness (OEE)

• OEE is a measure of equipment or process effectiveness
based on actual availability, performance, and quality of
product or output.

7
 Eng. Islam Ramy

Overall Equipment Effectiveness (OEE)

• Availability (A) = Actual production time / planned
production time
 Equipment failure
 Set-up & adjustments

• Performance (P) = Actual run rate / planned rate

 Idling & minor
 Reduced speed
• Quality (Q) = good product / total product
 Defects in process & rework
 Start up losses
OEE = A*P*Q
- World class 85 %
- Typical plant 65 %

Common maintenance practices

• Breakdown maintenance (corrective maintenance)

• Preventive maintenance

• Predictive maintenance (condition based)

• Proactive maintenance

8
 Eng. Islam Ramy

Breakdown maintenance

“ Just let it break”

• Maintenance repair work done as an immediate response
to an asset failure, normally without planning and
unscheduled, Synonymous with reactive and emergency
maintenance.

Breakdown maintenance

• It is possible to make a sound economic decision to allow

certain non-critical machines to run to failure.
• In this case we call it run to failure philosophy.

Run-to-Failure (RTF)
A maintenance strategy (policy) for assets where the cost and
impact of failure is less than the cost of preventive actions. It
is a deliberate decision, based on economical effectiveness,
not to perform PM but let asset run to fail.

9
 Eng. Islam Ramy

Breakdown maintenance

Advantages
• No costs that relates to predictive & preventive
maintenance.

Disadvantages
• Un-planned down time
• Catastrophic failure
• Secondary damage
• Loss of production
• High repair costs
• Safety risk
• Lost of control

Preventive maintenance

“ fix it before it breaks”

• A maintenance strategy based on inspection, component
replacement, and overhauling at a fixed interval, regardless
of its condition at the time. Usually, scheduled inspections
are performed to assess the condition of an asset.
Replacing service items, e.g., filters; oils, and belts and
lubricating parts are a few examples of PM tasks. PM
inspection may require another work order to repair other
discrepancies found during the PM.

10
 Eng. Islam Ramy

Preventive maintenance
• Also known “scheduled maintenance” “planned
maintenance” “calendar based maintenance”.
• The problem is we don’t know the best time to schedule
the maintenance actions to be performed.

Preventive maintenance
• Infant Mortality: the high probability of failure
immediately after an overhaul due to poor lubrication,
incorrect parts being installed, parts installed incorrectly,
poor alignment & balance ….etc.
• The dilemma is that we have to schedule the maintenance
in the probable life period and before the machine fails,
unfortunately we don’t what this period is or how quickly
the machine may fail once the wear out phase begins.
• So either we will perform maintenance more frequently
while the machine is running normally & make matter
worse (increase the risk of infant mortality) or lengthen
the period of maintenance & machine fails without being
maintained.
• Majority of machines doesn’t follow the normal wear
pattern & fails randomly

11
 Eng. Islam Ramy

Preventive maintenance

Advantages
• Unexpected machinery failure should be reduced.
• More control over parts storage & costs.
• Less interruption to production.

Disadvantages
• Unnecessary maintenance actions performed.
• Infant mortality risk.
• The maintenance schedule is the same for all machines,
not tailored to individual machine requirement or life
expectancy.

Predictive maintenance

“ if it isn't break, don’t fix it”

• Maintenance based on the actual condition (health) of an

asset as determined from non-invasive measurements and
tests. PdM allows preventive and corrective actions to be
optimized by avoiding traditional calendar or run-time
directed maintenance tasks. The terms Condition Based
Maintenance (CBM), and Predictive Maintenance (PdM)
are used interchangeably.

12
 Eng. Islam Ramy

Predictive maintenance

• It is a maintenance program based on the ability to predict

the future condition of the machines using information on
the current and past condition.
• The detailed maintenance work will be based on the
results of the machine condition analysis using several
technologies.
• The ultimate goal of PdM is to identify proactive
maintenance actions to be performed at a scheduled point
in time when the maintenance activity is most cost
effective and before the asset fails in service.
• The goal of predicting the future trend of the asset’s
condition. This approach uses principles of statistical
process control, trend analysis, and preselected thresholds
to determine at what point in the future maintenance
activities should be scheduled.

Predictive maintenance

There are a number of different CBM / PdM technologies

that can be used to evaluate assets condition. A few of the
more common technologies (or data) are:
• Vibration analysis
• Infrared (IR) thermography
• Acoustic / Ultrasonic — sound level measurements
• Oil analysis
• Electrical — amperage plus other data
• Operational performance data — pressure, temperature,
flow rates, etc. Basically, in the CBM

13
 Eng. Islam Ramy

Predictive maintenance

Advantages
• Unexpected down time is reduced.
• Inventory control (parts are ordered only when they are
needed).
• Maintenance is only performed when convenient.

Disadvantages
• Cost of technologies, training, systems & personnel.
• Does not ensure that the machine life is extended.

Proactive maintenance

“ fix it once, & fix it right”

• The sum of all maintenance work that is completed to

avoid failures or to identify defects that could lead to
failures (failure finding).It includes routine preventive and
predictive maintenance activities and work tasks identified
from them.
• Predictive maintenance gives us warning for example of
bearing failure we can the order the replacement &
schedule the repair, but why did the bearing fail if you find
the root cause & fix it that is a proactive maintenance
approach.

14
 Eng. Islam Ramy

Proactive maintenance

Examples of proactive maintenance approach:

• Determine the root cause analysis (RCA)
• Improve reliability (change in design)
• Precession balancing
• Precession alignment (laser alignment)

Proactive maintenance

Advantages
• Extended machine life
• Improve reliability
• Reduced overall maintenance cost
• Less secondary damage
• Reduce downtime
Disadvantages
• Requires culture shift & change of management philosophy
• Additional time investment & additional skills requires.
• High cost of instrumentation , systems, personnel

15
 Eng. Islam Ramy

P-F curve

P: Potential failure (detectable)

F: Functional failure (can no longer perform function(

P-F curve

• P-F interval is not the same it changes depending on the

nature of the failure historical data helps us to determine
this interval.
• If there is a short P-F interval we need to carry spares, if
the P-F interval is long we have time to order.
• Poor planning increases risk of catastrophic failures,
increase risk of safety incident, increases inventory costs,
wastes labor & increases stress & frustration.

16
 Eng. Islam Ramy

Condition monitoring
techniques

Condition monitoring

• The start of condition monitoring may have been when a

mechanic first put an ear to the handle of a screwdriver,
touched the other end to a machine, and pronounced that
it sounded like a bearing was going bad.

17
 Eng. Islam Ramy

Condition monitoring

• We have come a long way since then with a variety of

technologies for analyzing what’s happening inside the
asset. However, the need for a knowledgeable,
experienced person to use the technology has not
changed.

Condition monitoring

Any condition-based maintenance program can be

characterized by a combination of three phases:
• Surveillance — monitoring machinery condition to detect
incipient problems. (condition monitoring)
• Diagnosis / Prognosis — isolating the cause of the problem
and developing a corrective action plan based on its condition
and remaining life.
• Remedy — performing corrective action.

Predictive maintenance

Predictive maintenance ≠ condition monitoring

18
 Eng. Islam Ramy

Condition Monitoring Data Collection

Asset condition data is collected basically in two ways:

1. Spot readings — route based with portable instruments

2. Permanently installed data acquisition equipment for

continuous online data collection. Typically, permanent data
collection systems are installed only on critical and expensive
assets and systems used in production processes.

Condition monitoring technologies

• Vibration analysis
• Acoustic emission (Ultrasound)
• Infrared thermography
• Electric motor testing
• Oil & wear particle analysis

We will talk briefly about each technology & its application

each one of them is a science & needs more study & practice
to master it.

19
 Eng. Islam Ramy

Acoustic emission

• Most of the sounds sensed by humans range between 20

Hertz and 20 kilohertz (20 cycles per second to 20,000
cycles per second). The average human high frequency
threshold is actually 16.5 kHz.
• The ultrasound sensor is used to measure the signal &
heterodyne (demodulate) to a frequency range within the
human hearing range.

Consideration while measuring Ultrasound:

• The sounds are directional

• The volume depends on the distance from the source. (for
distances a concave dish is used)
• Sounds travels through air, liquid & solid objects but not in
vacuum.
• There are two modes of detecting ultra sound :
1. Air borne.
2. Direct contact.

20
 Eng. Islam Ramy

Consideration while measuring Ultrasound:

There are two ways to use the ultra sound data:

• Listen to it, search for leaks, listen to faults…etc.
• Measure it either then units are displayed in dB or view
time wave form & spectrum.

Ultrasound:
The basic advantages of ultrasound and ultrasonic
instruments are:
1. Ultrasound emissions are directional.
2. Ultrasound tends to be highly localized.
3. Ultrasound provides early warning of impending
mechanical failure.
4. The high frequency, short wave characteristic of ultrasound
enables users to accurately pinpoint the location of a leak,
electrical emission or of a particular sound in a machine.
5. The instruments can be used in loud, noisy environments
6. They support and enhance other PDM technologies or can
stand on their own in a maintenance program.

21
 Eng. Islam Ramy

Ultrasound Applications:
• Air leaks
• Steam leaks
• Defected Steam trap
• Electrical faults (Corona)
• Bearing faults & lubrication

Infrared thermography:
• An infrared camera measures and images the emitted
infrared radiation from an object.
• The fact that radiation is a function of object surface
temperature makes it possible for the camera to calculate
and display this temperature.
• The device translates the detected level of radiated energy
into temperature based on information entered by user.

22
 Eng. Islam Ramy

Infrared thermography:
• Spot radiometers

• Infrared Cameras

“Note: infrared camera does not see temperature is calculated from inputs by
the user in the camera’s software”

Infrared thermography:
Typically used to measure:

• Mechanical faults (coupling, bearings…..etc.)

• Electrical (transformers, motors, control panels)

• Steam systems

• Refractory plant

23
 Eng. Islam Ramy

Infrared thermography:
Infrared camera detect three forms of radiation:

• Emitted Heat
• Reflected Heat
• Transmitted Heat

Infrared thermography:
Blackbody Radiation
• All objects with a temperature above absolute zero (0 K, -
273.15 C °) emit energy in the form of electromagnetic
radiation. A blackbody is a theoretical or model body which
absorbs all radiation falling on it, reflecting or transmitting
none. It is a hypothetical object which is a “perfect”
absorber and a “perfect” emitter of radiation over
all wavelengths(emissivity=1).

24
 Eng. Islam Ramy

Infrared thermography:
Emissivity:
• Emissivity is defined as the ratio of the energy radiated by
an object at a given temperature to the energy emitted by
a perfect radiator, or ‘black body’ at the same temperature.
• Emissivity is the measure of an object's ability to emit
infrared energy. Emitted energy indicates the temperature
of the object. Emissivity can have a value from 0 (shiny
mirror) to 1.0 (blackbody).

Infrared thermography:

• All of these bars are the same temperature:

 Black body = 1
 Oxidized copper = 0.68
 Polished copper = 0.02

25
 Eng. Islam Ramy

Infrared thermography:
Reliable measurement:

• Emissivity greater than 0.6

• Known controlled background

• Controlled or no wind

• No reflected background radiation

Infrared thermography:
Applications:

26
 Eng. Islam Ramy

Electric motor testing

Tests fall into two categories:

• Static off-line tests

• Dynamic on-line tests

Electric motor testing

Static off-line tests:

• Winding resistance
• Meg-ohm
• Polarization index
• High potential
• Surge testing

27
 Eng. Islam Ramy

Electric motor testing

Dynamic on-line tests:
 Motor signature analysis
• Dynamic voltage & current treated like vibration.
• View spectrum & time wave form

Electric motor testing

Dynamic on-line tests:
 Motor signature analysis

Good motor Bad motor

28
 Eng. Islam Ramy

Oil & Wear particle Analysis

Three categories of oil analysis:
• Fluid properties (physical & chemical properties of used oil)
• Contaminants (fluid & machine destructive contaminants)
• Wear particles (presence & identification of wear)

Fluid properties:
This type of oil analysis focuses on identifying the
oil’s current physical and chemical state as well as
on defining its remaining useful life (RUL). It
answers questions such as:
• Does the sample match the specified oil identification?
• Is it the correct oil to use?
• Are the right additives active?
• Have additives depleted?
• Has the viscosity shifted from the expected viscosity? If so,
why?
• What is the oil’s RUL?

29
 Eng. Islam Ramy

Contamination
By detecting the presence of destructive
contaminants and narrowing down their probable
sources (internal or external), oil analysis can help
answer questions such as:
• Is the oil clean?
• What types of contaminants are in the oil?
• Where are contaminants originating?
• Are there signs of other types of lubricants?
• Is there any sign of internal leakage?

Wear debris
This form of oil analysis is about determining the
presence and identification of particles produced as
a result of mechanical wear, corrosion or other
machine surface degradation. It answers questions
relating to wear, including:
• Is the machine degrading abnormally?
• Is wear debris produced?
• From which internal component is the wear likely
originating?
• What is the wear mode and cause?
• How severe is the wear condition?

30
 Eng. Islam Ramy

Oil analysis & wear particle analysis:

Wear particle analysis :

• There are many wear debris analysis techniques Some of
these methods can even analyze wear debris right from
the oil sample. Analytical ferrography requires wear debris
to first be isolated from the collected oil sample.
• Photomicrograph images of the ferrograms or filtergrams
are then studied using an optical microscope. The main
advantage of analytical ferrography is its ability to
determine particle shapes, sizes and textures as well as
identify elements.

Chemical corrosive wear

31
 Eng. Islam Ramy

Oil analysis tests

*Gas compressors only ** Air compressors only ***For

phosphate ester fluids, consult the fluid supplier and/or turbine
manufacturer. R = Routine testing E = Exception test keyed to a
positive result from the test in parentheses

Vibration Principles

32
 Eng. Islam Ramy

What is vibration:
• Vibration is a mechanical phenomenon whereby
oscillations occur about an equilibrium point. The
oscillations may be periodic, such as the motion of a
pendulum—or random, such as the movement of a tire on
a gravel road.
• The forces within the machines cause vibration which is
transferred to the bearings.
• By acquiring measurements on the bearing that we can
analyze we will know the nature of the fault & its severity.

“Vibration is the machine language learn it & you

will understand what the machine is trying to tell
you”

Simple harmonic motion:

• Repetitive movement back and forth through
an equilibrium, or central, position, so that the maximum
displacement on one side of this position is equal to the
maximum displacement on the other side. The time
interval of each complete vibration is the same.

• The force responsible for the motion is always directed

toward the equilibrium position and is directly
proportional to the distance from it. That is, F = −kx,
where F is the force, x is the displacement, and k is a
constant. This relation is called Hooke’s law.

33
 Eng. Islam Ramy

Simple harmonic motion:

Amplitude:
• The magnitude of vibration expressed in signal level (mV or
mA) or engineering units (microns , mils, mm/sec , g’s)

34
 Eng. Islam Ramy

Amplitude:
• Amplitude is measured in either displacement , velocity or
acceleration units.
• And defined by one of these terms
 Root Mean Square (rms)
 Peak to peak (pk-pk)
 Peak (pk)

Amplitude:

For pure sine wave :

 RMS = Peak/ 𝟐 = 0.707*Peak

 Peak to peak = 2*peak

These equations are only for pure sine wave it doesn’t apply
for complex signals.
Peak
RMS NOT = 0.707*pK

35
 Eng. Islam Ramy

Frequency:
• Frequency is the repetition rate of vibration per unit of
time.
 Cycles per minute (CPM)
 Revolution per minute (RPM)
 Cycles per second (HZ)
 CPM = RPM = 60*HZ

Frequency = 1/T (period)

Period:
• The amount of time it takes to complete on cycle of
vibration.
• The reciprocal of frequency

 Seconds
 Milliseconds
 1 millisecond = 0.001 seconds

Period (T) = 1/FREQUENCY

36
 Eng. Islam Ramy

Period :

The period is ??????

Fan vibration example:

• Imagine a Fan with 8 blades and we attached a coin to one
of the blades what would happen ?

37
 Eng. Islam Ramy

Introducing phase:
• Phase relates to timing between two signals.

Lets go back again to the fan with a coin example:

One cycle is equal to 360˚

Introducing phase:
• Phase relates to timing between two signals.

Now let us add another fan with same speed (same frequency) as our fan.

To measure phase between two signals they must be at the

same frequency & synchronized with each other

38
 Eng. Islam Ramy

Introducing phase:

The two fans are out of phase (180˚)

Introducing phase:

The phase between the two fans is 90˚

39
 Eng. Islam Ramy

Introducing phase:
Why phase is important?

• It is very useful in diagnosing faults.

• Machine balancing.
• Understand the relative motion of different point on the
structure.

How to measure phase:

- Absolute phase:
• It is measured using a tachometer signal (or any reference
signal as key phasor, stroboscope).

An optical Laser tachometer generates a pulse every shaft revolution

using reflective tape attached to the shaft and aiming a laser towards it
as a result it will create a pulse every shaft revolution.

40
 Eng. Islam Ramy

How to measure phase:

- Absolute phase:
• Optical Tachometer with reflective tape

How to measure phase:

- Absolute phase:

Vibration signal

Tach signal

- In vibration the absolute phase is a measurement of phase lag which means that
we measure the time difference starting from tach signal to the peak of the
vibration wave peak so we can say that the vibration signal lag the tach signal by
say 90˚ so the phase will be +90 ˚.
- Phase lag (+ve)

41
 Eng. Islam Ramy

How to measure phase:

- Absolute phase:
• Example:

 Absolute phase is used to measure phase in single channel analyzers.

 Absolute phase is used in balancing machines.

How to measure phase:

- Absolute phase:
• Stroboscope

42
 Eng. Islam Ramy

How to measure phase:

- Relative phase:
• Measured on a multi-channel vibration analyzer using two
or more (similar type) vibration sensors.
• Relative phase is the time difference between the
waveforms at a specific frequency converted to degrees
• Relative phase does not require a tachometer so phase can
be measured at any frequency.

How to measure phase:

- Relative phase:

 As shown in these figure here the reference point is A which reaches

its peak 90˚ before point B so the phase is 90˚.
 Please note that the phase is measuring the relative timing between
signals not angle but as the shaft rotates 360˚ every cycle we
represent time in degrees.
 All phase values measured are in the range of plus or minus 30°.

What happens if B is the reference signal? What will be the

phase readings???

43
 Eng. Islam Ramy

How to measure phase:

- While acquiring phase be aware of the position of the sensor in the below
image the two sensors are in the axial position but there position is 180˚
different so you must compensate the reading you get by adding 180 ˚.

Uncompensated

Compensated

Representing phase:
- Bubble diagram:

44
 Eng. Islam Ramy

Coherence :
• Coherence is a number between one and zero, and is a
measure of the degree of linearity between two related
signals, such as the input force of a structure related to the
vibration response to that force. Coherence is thus a two-
channel measurement, and does not apply to single-
channel measurements of vibration signatures. In a
frequency response measurement of a mechanical
structure, if the structure is linear, the coherence will be
one, but if there is some non-linearity in the structure or if
there is noise in one or the other measurement channel,
the coherence will be less than one.

• The dual-channel FFT analyzer is able to measure the

coherence between the two channels, and this is a useful
tool in determining good from noisy or meaningless data.

Displacement, Velocity & acceleration:

• Displacement: describes the distance traveled by the
mass, how far up & down it is moving.

• Velocity: the rate of change in displacement, how fast the

mass is moving at any distance (how quickly it will cover
the distance).

• Acceleration: the rate of change in velocity, how quickly

the mass is speeding up or slowing down.

45
 Eng. Islam Ramy

Displacement, Velocity & acceleration:

Displacement, Velocity & acceleration:

 Acceleration Leads Velocity by 90˚

 Velocity Leads Displacement by 90˚
 Acceleration Leads Displacement by 180˚

46
 Eng. Islam Ramy

Vibration Amplitude units:

• Vibration can be directly measured using Accelerometers
(acceleration), Velocity meters (Velocity) & proximity
probes (displacement).
• It can be converted from one unit to another

Commonly it is specified as follows:

 Acceleration- g’s Peak (1 g = 9.8 m/sec^2)
 Velocity- mm/sec RMS (root mean square)
 Displacement- microns PK-PK (peak to peak)

Vibration Amplitude units:

• Displacement: proportional to stress, strain is the
fractional deformation produced in a body when its
subjected to a set of deforming force, strain produced in a
body is directly proportional to stress.

• Commonly used when you are interested in low frequency.

• Commonly used on machines with proximity probes

installed.

47
 Eng. Islam Ramy

Vibration Amplitude units:

• Velocity: proportional to fatigue.

• Generally velocity is a good measurement of vibrations

across most machine speeds & frequency of interest (2 to
2000 HZ).

Vibration Amplitude units:

• Acceleration: the rate of change in velocity, how quickly
the mass is speeding up or slowing down.

• Most sensitive to high frequency vibration.

• Low speed is very little acceleration.

• Used in high speed machined & high frequency analysis

(bearings & gearboxes).

48
 Eng. Islam Ramy

Comparing units:
To convert from:
• displacement to velocity we perform differentiation.
• velocity to acceleration we perform differentiation.
• acceleration to velocity we perform Integration .
• velocity to displacement we perform Integration .
• acceleration to displacement we perform double
Integration .

As the most used transducer is accelerometer, Integration is mostly used

Displacement = D*sin(wt)
Velocity = D*w*sin(wt + 𝜋/2)
Acceleration D*w^2*sin(wt + 𝜋/2)
Where w = 2*𝜋 ∗f

Unit conversions:

49
 Eng. Islam Ramy

Comparing units:

Comparing units:

50
 Eng. Islam Ramy

Complex Vibration:

Let’s consider a real Machine:

• Motor, pump , fans etc….
• Each part in the machine will generate vibration.
• As a result the vibration signal won’t be just a simple sine
wave.

Complex Vibration:

Now Let’s play with our coin & fan :

• As we learned if we attach a coin on one of the fan’s blade
it will generate unbalance force at the frequency of shaft
revolution.
• Assume the fan speed is 6 HZ the wave form now has 6
cycles per second.

51
 Eng. Islam Ramy

Complex Vibration:

• Also as we will learn the fan blades generates a vibration

force called blade passing frequency as each blade will
create a pressure wave as it displaces the air.
• Assume the fan has 8 blades it will create 8 pulses every
shaft revolution it is occurring 8 times faster (6*8= 48 HZ).

Complex Vibration:

• The signal we measure will show the combination of the

two signals.

52
 Eng. Islam Ramy

Complex Vibration:

• In real application we don’t have only two signals, it’s a

combination of many signals the machine components
generate.
• Unfortunately it will be some thing like these.
UT 1 - EFFLUE NT WATE R PUM P
10P-9174 C-P1H Pump Inboard Horizontal
20
Route Wave form
16/Jan/17 11:08:08
15
PK = 7.13
LOAD = 100.0
RPM = 1501. (25.02 Hz )
10
PK(+) = 17.24
PK(-) = 15.86
5 CR ESTF= 3.42
Acceleration in G-s

-5

-10

-15

-20

0 50 100 150 200 250 300 350 400

T ime in mSe cs

Spectrum:

• The spectrum enables us to easily see which frequencies

are present in the vibration.

• The Fast Fourier Transform (FFT) converts time waveform

to spectrum. (time domain to frequency domain)

53
 Eng. Islam Ramy

Spectrum:

• Back to the fan example:

Orders:

• We have mentioned that frequency is measured in HZ or

CPM.
• Now lets introduce orders.

Orders is the normalization of all the spectrum

frequencies by the speed of the machine.

For example in our fan the speed is 6 HZ.

 6/6= 1 order or 1X of fan speed.
The Blade pass frequency is 48 HZ
 48/6= 8 order or 8X of fan speed.

54
 Eng. Islam Ramy

Orders:

• So the spectrum will look as these.

These will help us so much when analyzing the spectrum as

we will learn later.

Spectral regions:

• Spectrum can be grouped into 3 different regions.

Synchronous : energy that is integer (whole number) multiple of

running speed.

Non-Synchronous: energy that is Fractional multiple of running speed.

Sub-Synchronous : energy that is below running speed.

55
 Eng. Islam Ramy

Forced vibration:

‘The oscillation of a system under the

action of a forcing function. Typically
forced vibration occurs at the frequency of
the exciting force’

Forcing frequencies:

Examples of forcing frequencies:

• Blade passing rate (synchronous)
 (number of blades * speed)
• Vane passing rate (synchronous)
 (number of vanes * speed)
• Bearing frequencies (non- synchronous)
(depends on speed & bearing geometry )
• Belt rate (sub- synchronous)
 ( Pi * Sheave diameter * sheave RPM) / (length of belt)
• Gear mesh frequency (synchronous)
 ( number of teeth * shaft RPM )
• Etc…

56
 Eng. Islam Ramy

Free vibration:

• Vibration of a mechanical system following an initial force -

typically at one or more natural frequencies.
• When the forcing frequency is near the natural frequency
of a mechanical system, the vibration response amplitude
can become highly amplified. This phenomenon is called
resonance.
• Natural frequency : The frequency of free vibration of a
system when excited with an impact force.

Overall levels:

• Overall vibration levels in RMS is indication of the power of

the signal.
• A single number that can be trended & compared to alarm
limits.
• RMS values can be computed in different ways so be
careful that two overall meters can measure different
values as it is dependent on the way it is calculated (not
universal).

57
 Eng. Islam Ramy

Overall levels:

RMS Overall levels can be computed by 3 different

Methods:

• Analog method

• Digital method

• From the spectrum

Overall levels:

Analog Method:
• Recall that RMS is equal to 0.707*Peak but this is only true
for a pure sine wave.
• Most vibration are not sinusoidal.
• The image below shows the analog method.

58
 Eng. Islam Ramy

Overall levels:

Digital Method:
• Analog signal is digitized & broken up into number of
discrete samples(n).
• The sum of squares is divided by the number of samples.
• The square root of these values is computed to get RMS.

Overall levels:

From the spectrum:

• Square each discrete frequency RMS value in the spectrum,
Add them and then get the square root of that value to get
the overall RMS value
• Note that the calculated value is dependent on the
frequency range we set.

Overall RMS = (𝑅𝑀𝑆𝑎)2 + (𝑅𝑀𝑆𝑏)2 + (𝑅𝑀𝑆𝑐)2 + (𝑅𝑀𝑆𝑑)2

59
 Eng. Islam Ramy

IS0-10816:

• Defines the severity of the vibration measured over a

frequency range from 2 HZ to 1000 HZ.
• It is used in acceptance testing, trending & alarm limits.
• It must be considered that if the vibration is out of the Iso
range it will missed so it is not suitable for the assessment
of bearings & gearboxes.

Crest factor:

• Crest factor is the ratio of the Peak to RMS value.

𝑷𝒆𝒂𝒌
Crest factor = 𝑹𝑴𝑺
• It gives us an idea of how much impacting in the time wave
form.
• High crest factor indicate impacting.

60
 Eng. Islam Ramy

Data Acquisition

Acquiring data:

Data must be acquired correctly:

• Speed & load
• Type of sensor used
• Measurements locations
• Sensors mounting
• Measurement fault identified & corrected

61
 Eng. Islam Ramy

Data collector:

Data collector have a very important role:

• He must get good data.
• Observations are extremely important.
• Keep your eyes & ears open & record the observations.
• Look for leaks , hear the noise , know the operating
conditions , feel hot surfaces carefully.
• Report problems correctly.
• These observations will extremely help the vibration
analyst.

Data collector:

Data collector routine work procedures:

• Load the data to the vibration analyzer.
• Collect the route measurements & record your
observations.
• Dump the data to the computer.
• You can use the software to generate quick exception
reports.
• Report your observations to the analyst.
• Analysis is performed.

Machines normally measured routinely every 30-90 days we look for changed
vibration patterns, our target is to catch a problem before the risk of failure is
too high.

62
 Eng. Islam Ramy

Types of transducers:

• Proximity probes (Non-contact) (displacement)

• Velocity meters (Contact) (Velocity)

• Accelerometers (Contact) (acceleration)

Proximity probes:

• Proximity probes also called eddy current probes.

• Measures the relative displacement of the shaft in fluid
film (journal) bearings.
• Permanently mounted (drilled into the bearing).
• Units are micron pk- pk or mils pK-pk.

63
 Eng. Islam Ramy

Theory of operation:

Components of the eddy current probe:

• Demodulator/ Oscillator
• Extension cable
• Probe

Theory of operation:

• As the gap between the shaft and the probe decrease the
voltage will increase.
• Common sensitivity of bently nevada probes is 200 mV/mil
or 7.87 V/mm(0.00787 mV/micron)

64
 Eng. Islam Ramy

Theory of operation:

The output of the demodulator will has two

outputs:
• DC Proportional to the gap depth. (shaft position, center
line plot).

• AC signal the dynamic output (time waveform , spectrum ,

orbits)

Velocity meters:

• Often Measure Bearing Housings (Rolling element bearing)

or Machinery Casing Vibration.
• Measure velocity directly (mm/sec RMS).
• Self Generating Devices (no external power needed it
generates electricity).
• Sensitivity is expressed in terms of mv/mm/sec or in
mv/ips.
• Big size and need a big space to install.

65
 Eng. Islam Ramy

Accelerometers:

• The most common used sensor in routine measurements.

• Measure acceleration directly (g’s pK).
• Need powering to function
• Available in different designs for several applications.
• Are used for permanent monitoring of machines.(gear
boxes)

Accelerometers:

• Made mainly of piezoelectric materials (crystal).

• Measure acceleration directly, velocity through single
integration and displacement through double integration
process.
• Sensitivity is measured in terms of mv/g.
• Mostly have built in amplifier which have to be supplied
with 4 to 20 mille-amps from an external supply or the
instrument.

In comparison with velocity pickups accelerometers are:

 Lighter weight.
 Wider frequency response.

66
 Eng. Islam Ramy

Accelerometers:
ICP Integrated circuit Charge Mode
piezoelectric
 Powered from Analyzer /  Needs an external charge
Data Collector. amplifier.

 Limited temp. because of  Ability to operate at high

built in electronics. temperature.

 Internal amplifier.  External charge amplifier

Accelerometers designs:

Three most common designs:

Compression
 Compress the crystal under a mass. (commonly used in our
applications)

67
 Eng. Islam Ramy

Accelerometers designs:

Three most common designs:

Shear
 Subjects the sensing elements to shear (most commonly used).

Accelerometers designs:

Three most common designs:

Flexural
 The piezoelectric element is secured to the seismic mass in the form
of double cantilever beam (not commonly used).
 Most common used because it is not affected by temperature
changes

68
 Eng. Islam Ramy

ICP sensors settling time:

• When ICP accelerometers is connected to the power

source (Analyzer) it takes a few seconds for the amplifier to
stabilize.
• Sensor must settle before taking measurements.

Triaxial accelerometers:

• Standard accelerometer can measure in one axis.

• Triaxial accelerometer there are three accelerometers
mounted orthogonal to each other .
• The data collector can then capture the three axis at once.

69
 Eng. Islam Ramy

Condition monitoring goals:

• Collect enough data to identify machinery faults.

 Too much data more time to analyse also higher cost.
 Too little data not enough information.

• Collect repeatable data for trending(Repeatability).

• Collect good quality data.

Condition monitoring goals:

• Important issues:
 Limitations of the sensor
 Mounting method
 Mechanical transmission path
 Mounting locations

• Goals:
 Repeatability
 Safety
 Cost
 Time

70
 Eng. Islam Ramy

Transducer frequency response:

• Ideal respnse

Unfortunately life is not always good

Transducer frequency response:

• In reality it is not flat

 Lower limit (Roll off): low frequency limit amplifier noise swamps
data.
 Upper limit (saturation): there is a resonance around 30 KHz, output
beyond amplifier capability susceptible to overload from high
frequencies(some special bearing analysis techniques utilizes these
range).

71
 Eng. Islam Ramy

Transducer operating range:

• Operating range commonly from 1 to 10 KHz.

Sensitivity:

• The ratio between electrical signal (output) and

mechanical quantity (input) in a mechanical to electrical
transducer.
• Low speed machines generate low vibrations so we use
high sensitivity sensors. (500 mV/g or 1V/g).
• Machines that generates very high levels of vibrations use
low sensitivity sensors (10 mV/g).
• Most machines in our applications use standard sensitivity
of 100mV/g .

72
 Eng. Islam Ramy

Sensitivity & frequency response:

Mounting methods:

• It is important to mount the sensor correctly to get useful

data.
• If not:
 Repeatability & frequency response will be an issue.

73
 Eng. Islam Ramy

Mounting methods:

• Screw (stud) mount.

• Adhesive mounting
• Magnetic mounting base
• Hand-held or probe-type mounting

Mounting methods:

• As you can see that the amounting method will highly

affect the frequency response.

74
 Eng. Islam Ramy

Mounting methods:

PROBE TIPS ADHESIVE MOUNTING

MAGNETS STUD MOUNTING

Mounting methods:

• In di-pole magnet mounting set one edge of the magnet

and then roll it into full contact with the machine.

75
 Eng. Islam Ramy

Mounting methods:

• Quick mounts are now offered these devices are

permanently mounted to the machine and the employ a
locking method.
• The accelerometer is then attached to the pad with a half
turn.

Selecting sensor:

The sensor should be selected based on:

• Frequency range required.

• Sensitivity.

• Environmental conditions (temperature , moisture).

• Mounting Method

76
 Eng. Islam Ramy

Repeatability:

The

Collecting data:

• Mechanical transmission path.

 There all a lot of vibration sources inside the machine.
 The bearings carrying the machine load so the forces are transmitted
to the bearing.
 The measurement location should have a good mechanical path
(vibration should travel along solid metal with no joints, gaps)

77
 Eng. Islam Ramy

Measurement Axis:

• As we will learn in the analysis section every fault in the

machine will generate a force in a certain axis.
• So you must know the axis you need to measure to identify
the machine fault generally we measure on the three axis:
 Horizontal (X)
 Vertical (Y)
 Axial (Z)

Measurement Axis:

• For vertical machines as a rule of thumb the horizontal axis

will be perpendicular(90°) to the discharge line (pipe).
• The most important issue is repeatability you must fix &
identify your measurement locations.

78
 Eng. Islam Ramy

Measuring inaccessible locations:

• There may inaccessible locations due to physical , safety ,

or environmental reasons, that a sensor must be
permanently installed on the machine and wired back to a
junction box.

Naming conventions & Measurement locations:

 Motor Non Drive End (NDE) or Motor outboard (H,V,A) (M1(H,V,A)).
 Motor Drive End (DE) or Motor Inboard (H,V,A) (M2(H,V,A)).
 Pump Drive End (DE) or Pump Inboard (H,V,A) (P1(H,V,A)).
 Pump Non Drive End (NDE) or Pump outboard (H,V,A) (P1(H,V,A)).

79
 Eng. Islam Ramy

Measurement locations:

Recognizing bad data:

• While acquiring measurements you must view the data.

• Bad data can be easily recognized as it will appear as noise
& ski-slope on the spectrum.

Ski-slope: If the accelerometer is either mechanically loose on

the block, has a loose wire, is overheated or experiencing a
temperature transient, or physically overloaded when data is
collected, the problem will manifest itself in the spectrum as
a low frequency "ski slope" at the left end of the low
frequency range graph.

80
 Eng. Islam Ramy

Recognizing bad data:

• While acquiring measurements you must view the data.

• Bad data can be easily recognized as it will appear as noise
& ski-slope on the spectrum.

Ski If the accelerometer is either mechanically loose on the block, has a loose wire, is

Recognizing bad data:

Sources of bad data:

• Thermal transients
• Mechanical shock
• Defected Cable
• Loose mounting
• Poor setup

81
 Eng. Islam Ramy

Understanding signals

Understanding signals:

• As we described earlier machine generates various

vibration signals which mix together & produce a complex
timewave form.
• The analog signal is then digitized & and FFT is performed
on the digital time waveform to get the spectrum.
• These signals mix together & generates many peaks in the
spectrum.
• It is important to know how the signals in the time
waveform translated in the spectrum to get a sense of
what is going on your machine.
 Sine waves
 Truncation
 Transients & impulses
 Amplitude Modulation
 Beating

82
 Eng. Islam Ramy

Sine Wave:

• Pure sine wave is a single producing a single peak in the

spectrum.

Sine Wave:

• Now we will add another source of vibration which is also

pure sine wave & mix it with another sine wave.
• The time waveform shape will dramatically change.
• In the spectrum we will have two peaks at the two
frequencies.

83
 Eng. Islam Ramy

Sine Wave:

• Two signals in phase & have the same amplitude &

frequency.
• They will add together & the spectrum will show one peak
at the signals frequency and the addition of there
amplitude.

Sine Wave:

• Two signals out of phase & have the same amplitude &
frequency.
• They will cancel each other (phase cancelation).

84
 Eng. Islam Ramy

Beating:

• When two signals are very close to each other in frequency

(less than 4 HZ) eventually the will go in & out of phase
these will produce a beating cycle.
• In the spectrum they will appear as two peaks.

Beating:
Example:
• Two signals 11 HZ & 10.5 HZ.
• Amplitude of both signals = 5
• Frequency difference = 0.5 HZ
• Time of one beat cycle = 1/0.5 = 2 seconds

85
 Eng. Islam Ramy

Beating:
The figure below shows 10 cycles of the previous beating
example:

2 seconds

Beating:
The spectrum of previous beating example:

In the spectrum if there is a good resolution there will be two peaks at 10.5 HZ & 11 HZ

In the spectrum if there is low resolution it will appear as one peak in the range of 10.5
to 11HZ and its amplitude rising and falling every beat cycle

86
 Eng. Islam Ramy

Beating:
In the previous example the two were having the same
amplitude that’s why the amplitude goes to zero in the out of
phase region in real applications the two signals mostly will
have different amplitudes so they will add & subtract when
they go in & out of phase but it will not goes to zero:

Two signal beats with different amplitudes

Beating:

Examples of beating:
• Two machines of very close speed and there is good
transmission path between them.
• Fan on belt or gearbox beating could occur between motor
speed & fan blade passing frequency.
 (ex: motor speed = 24.7 HZ, Fan speed = 3.6 HZ the fan has 6 blades
so the fan blade passing frequency = 3.6 * 6 = 21.6 HZ).
 Therefore beating could occur between motor speed 24.7 HZ & Fan
blade passing frequency 21.6 HZ (beat frequency= 24.7-21.6 = 3.1 HZ ,
Beat period = 0.32 sec)
• As the difference between the two frequencies decrease
the time of on beat cycle increase so we need more time
while collecting data to capture it.
• You can hear a droning sound when beating is present.

87
 Eng. Islam Ramy

Beating:

• Beating is not considered as a machine distress but if the

amplitude when the two peaks added together is high it
must be considered.
• To diagnose it first of all you probably will hear a droning
sound in the spectrum because the two peaks are so close
to each other they will appear as one peak (low resolution)
then you must acquire live data to see this peak rising &
falling then you may increase your resolution to be
sufficient to separate the two peaks.
• To correct this situation you may:
 Isolate the two machines with close frequencies.
 Change one of the close frequencies (in the fan example change fan
speed if it is practically possible)
 Find the frequency of the two that its amplitude is high and solve its
problem (may be one machine needs to be balanced, belt tension ,
blade angle)

Sum & difference frequencies:

• When to peaks beats together there sum & difference may

appear in the spectrum.
EX:
 Two peaks ( 100 & 99 HZ) you may see there sum 199 HZ & there
difference 1 HZ.

88
 Eng. Islam Ramy

Impulse:

• The integral of force over a time interval.

• The waveform will have just a spike on it & the spectrum
will be a flat line as the impulse injects energy to all
frequencies in the spectrum.
• Its application is the impact test.

Time waveform

Spectrum

Clipped wave distortion:

• When the waves appears to chipped or clipped in the time

waveform.
• This happens commonly when the motion is restricted in
one direction (looseness).
• Clipped wave will generate a multiple harmonics (series of
evenly spaced peaks integer multiples of the first in the
series) in the spectrum.

89
 Eng. Islam Ramy

Transients:

• Is a very steep increase in amplitude followed by an

equally steep decrease in level in the time wave form
(ringing).
• These will produce more harmonics in the spectrum in
sever cases it may produce also half & quarter harmonics
(0.5X, 1X , 1.5X etc…).
• It may be generate from severe rotating looseness ,
impacting (bearings).

Distortion VS. Actual signals:

• Actual sources of vibration : Misalignment , blade pass ,

coupling etc….

• Distortion : FFT produces Harmonics in the spectrum.

Beware you must be able to know the sources of the peaks in your

90
 Eng. Islam Ramy

Amplitude modulation:

• Amplitude modulation, or AM for short, is the fluctuation

in amplitude of one signal component due to the influence
of another signal component called the modulating
frequency. The modulating frequency is usually much
lower in frequency than the (modulated)carrier frequency.
Amplitude modulation is a non-linear process, and gives
rise to new frequency components in the spectrum which
would not be there without the modulation. These new
spectral components are called sidebands.

Amplitude modulation:

• Sidebands are spectral components that are the result of

amplitude modulation. The frequency spacing of the
sidebands is equal to the modulating frequency, and this
fact is used in diagnosing machine problems by examining
sideband families in the vibration spectrum. For instance, a
defective gear will exhibit sidebands at the gear rpm
around the gear mesh frequency.

91
 Eng. Islam Ramy

Amplitude modulation:

• Carrier frequency: Is a signal which is generated by

modulation, the frequency being modulated is called the
carrier frequency, by analogy to radio broadcasting, where
a very high frequency signal called the carrier is modulated
by the audio signal. In machinery vibration analysis, an
example of a carrier might be a gear mesh frequency
which is being amplitude modulated by the turning speed
of the gear.
• Modulating frequency : Is rate at which the carrier
frequency fluctuates in amplitude it appears in the
spectrum as equally spaced side bands around the carrier
frequency.

Amplitude modulation:

The spectrum of the amplitude modulated signal will show

the carrier frequency with two equally spaced sidebands
around it. (spaced by modulating frequency)

Wave repeats it self at frequency X; however Sidebands of X-Y and X+Y appear to either side
amplitude of wave varies at frequency Y. of peak at frequency X.

92
 Eng. Islam Ramy

Amplitude modulation:

Amplitude modulation:

• If the amplitude modulated signal is distorted there will be

sidebands due to modulation. But there will be also
harmonics due to distortion & family of sidebands.

93
 Eng. Islam Ramy

Amplitude modulation:
• Distorted Amplitude Modulation in gearbox.

Frequency modulation:
• Frequency modulation differs from amplitude modulation
in that the frequency varies in frequency modulation.
• Occurs when machine changes speed periodically (it could
be due to process loading or other causes).
• In gearboxes when it allows to much play the gears rock
back & forth as they turn changes the frequency
periodically.
• The spectrum is similar to amplitude modulation but much
more sidebands.

94
 Eng. Islam Ramy

Noise:
• Random vibration-time waveform that doesn’t repeat.
• It cause spectrum noise floor to be raised.

Noise floor

Noise:
• Causes of raised noise floor:
• Cavitation.
• Rotating looseness (bearing)
• Turbulence
• Poor lubrication
• Impacts & rubs
• Internal source
• External source

95
 Eng. Islam Ramy

Spectrum Analysis

Spectrum Analysis:
Before analyzing data you must :
• Understand the machine components & failure modes.
• Calculate forcing frequencies (BPF , GMF , Bearing
frequencies ).
• Defining standard test conditions (speed , load , operating
conditions )
• Test the machine in the correct condition with the correct
data collector setup.

96
 Eng. Islam Ramy

Spectrum Analysis:
Validate the data :
• Ensure that the data collected properly
• Look for the classic ski-slope

Spectrum Analysis:
Identify the speed:
• The spectrum regions & forcing frequencies are all related
to the speed of the machine so we normalize all
frequencies with respect to machine speed.

97
 Eng. Islam Ramy

Spectrum Analysis:
Identify the speed:
• For machines with belt-drives or gearbox take care that
you will have more than one speed you must identify them
all in order to identify the frequencies related to each
speed.

Spectral regions:

• Spectrum can be grouped into 3 different regions.

• Quick scan to spectrum helps you to identify the fault.
Synchronous : energy that is integer (whole number) multiple of
running speed.

Non-Synchronous: energy that is Fractional multiple of running speed.

Sub-Synchronous : energy that is below running speed.

98
 Eng. Islam Ramy

Sub-synchronous:
• Lower than running speed.
 Belt frequencies.
 Hydraulic instability (oil whirl).
 Rotor rub , shaft rub.
 Turbulence.
 Cage frequencies in bearing.
 Severe looseness.

Synchronous:
• Multiple integer of running speed.
 Imbalance  Gear mesh
 Misalignment  Coupling
 Bent shaft  Rotor bars (motor)
 Looseness  eccentricity
 Vane or blade pass

99
 Eng. Islam Ramy

Non-synchronous:
• Non-Multiple integer of running speed.
 EX: 3.1X , 5.65X etc..
 Rolling element bearings defect.
 Another component in the machine
 Multiples of belt frequency
 System resonances
 Cavitation.
 Electrical. ( 2X line frequency).

Harmonics :

• Look for harmonics that indicates clipping or transients.

100
 Eng. Islam Ramy

sidebands:
• Sidebands indicates modulation.

Noise:

• Noise indicates random events as cavitation & looseness.

101
 Eng. Islam Ramy

Display units:

• Displacement : very low frequency machines & shaft

vibration.

Display units:

• Velocity : sensitive across the entire spectrum however not

ideal for very low & very high frequencies.

102
 Eng. Islam Ramy

Display units:

• Acceleration: sensitive for very high frequencies.

Trending :
• Look for values changes compare it with international
standards (ISO , API).
• However trending will be an important issue to compare
patterns changes, the different in values between
measurement axis.
• Water fall plots are good for quick look at pattern changes

103
 Eng. Islam Ramy

Logarithmic scale :

Logarithmic scale :
• It is used to see very small amplitudes in the presence of
very large amplitudes.

104
 Eng. Islam Ramy

Time Waveform
Analysis

Time Waveform :
• In order to have the complete picture you must also have a
look at the time waveform.
• Time waveform is complex but there is patterns that you
can recognize to help you to catch faults.
• It is very useful especially in gearbox analysis.
• Look for :
 Transients
 Impacts
 Clipping
 Modulation
 beating

105
 Eng. Islam Ramy

Time Waveform :
• Similar to spectrum for medium frequency range that of
unbalance, misalignment etc… display it in velocity
(mm/sec)
• For high frequency bearings, gearboxes & when looking for
impacting display it in acceleration.

Time Waveform :
• Dominant sine wave.
• Possibly unbalance, bent shaft, eccentricity etc…

106
 Eng. Islam Ramy

Time Waveform :
• To see beating in TWF you must take long measurement
beating looks similar to amplitude modulation in time
waveform but you can differentiate between them in the
spectrum beating will show two peaks while amplitude
modulation will generate a carrier peak with two equally
spaced sidebands.

Difference bet. these two peaks is 2HZ

Time Waveform :
• Amplitude modulation

107
 Eng. Islam Ramy

Time Waveform :
• Amplitude modulation

BPFI
1X Sidebands 1X Sidebands

Time Waveform :
• Periodic impacts at 1X indicating severe rotating looseness.
• Because in looseness the motion will be restricted in one
of the directions probably the -ve values in the time wave
form is much greater the +ve ones or viceversa.

108
 Eng. Islam Ramy

Time Waveform :
• Look in the time waveform for peaks in the spectrum.

Time Waveform :
• Here we observed impacts in the time waveform when we
get the time between the impacts it was 0.1441 sec.
• Then get the frequency = 1/0.1441= 6.9 HZ
• Motor speed is 30 HZ so theses frequency is 0.23X (Order)
of motor speed which matches the belt rate frequency.
• So the belt of these machine is probably damaged.

109
 Eng. Islam Ramy

Time Waveform :
• Non periodic (Random) pulses may be cavitation.
• The spectrum noise floor will be raised.

Time Waveform :
• Wear in the gearbox as we see the time waveform shows
the impacts of the gear teeth better than the spectrum.

110
 Eng. Islam Ramy

Auto-correlation:
• Function used on the time waveform to show periodic
events. (from 0 to 1).

Normal time waveform

Autocorrelated time waveform

Introduction to fault
Analysis

111
 Eng. Islam Ramy

Unbalance (imbalance)

Unbalance:
• Occurs when the geometric (shaft centerline) and the mass
center of rotor do not coincide.
 Static unbalance .
 Couple unbalance.
 Dynamic unbalance.

112
 Eng. Islam Ramy

Unbalance:
• Unbalance creates a centrifugal force.
• 𝑈 = 𝑀𝑒𝑤 2 /𝑔.
 𝑀- Rotor mass
 e- eccentricity between mass & geometric centerlines.
 W- the angular velocity of the rotor
 g- gravitational force

Unbalance:
• Unbalance occurs at 1X in the radial direction (H , V) & its
time waveform is sinusoidal.
• Phase is a great tool to differentiate unbalance from other
faults

113
 Eng. Islam Ramy

Static Unbalance:

Static Unbalance:

• Static unbalance phase relationship.

114
 Eng. Islam Ramy

Couple Unbalance:

Couple Unbalance:

• Couple unbalance phase relationship.

115
 Eng. Islam Ramy

Dynamic Unbalance:
• In reality there is no pure couple or static unbalance
however it will be a combination between them called
dynamic unbalance.

Overhung machines Unbalance:

• In overhung machines the unbalance force will create a
bending moment in the shaft causing the bearing housing
to move axially so vibration at 1X will be in radial & axial
direction.

116
 Eng. Islam Ramy

Overhung machines Unbalance:

• Overhung machines unbalance phase relationship.

Causes of Unbalance:

Imbalance can be caused by a number of factors

including:
• improper manufacture.
• An uneven build up of debris on rotors/vanes/blades.
• The addition of shaft fittings without appropriate counter
balancing.
• With pumps, uneven wear on impellers is indicated as
imbalance.
• Incorrect key.
• Machining errors.

117
 Eng. Islam Ramy

Misalignment

Misalignment :

• Shafts are misaligned when their rotational center lines are

not collinear when the machines operating at normal
conditions.
• Misalignment is the root cause of majority of machine faults
(bearings failures, damaged seals, couplings & shafts)

118
 Eng. Islam Ramy

Parallel (offset) Misalignment :

• When the misaligned shaft centerlines are parallel but not

coincident.

Angular (gap) Misalignment :

• When the misaligned shafts centerlines meet at a point but

are not parallel.

119
 Eng. Islam Ramy

Common Misalignment :
• In reality the misalignment will be a mix of parallel & offset
misalignment.
• Symptoms are combined
• Severe misalignment will cause higher order components to
show on the spectrum (up to 8x).
• It could go even to a rich spectrum similar to that caused by
mechanical looseness.
• Coupling design would greatly influence the shape of the
spectrum in case of severe misalignment.

Coupling:
• Coupling type influence the shape of the spectrum for
example jaw coupling with 3 jaws the 3X peak & its
harmonics will be high.

120
 Eng. Islam Ramy

Time waveform of misalignment:

• Time waveform of misalignment might shows M & W shape.

Bent shaft

121
 Eng. Islam Ramy

Bent shaft:

Cocked bearing

122
 Eng. Islam Ramy

Cocked bearing:

• A cocked bearing is a situation where the inner or outer

race installed at an angle with the shaft or housing.

Cocked bearing:

123
 Eng. Islam Ramy

Cocked bearing:
• Cocked in the housing: The Axial phase shift around the face
of the bearing will be in one direction as the outer race
doesn’t rotate.
• Cocked in the shaft: The Axial phase shift around the face of
the bearing will be in all directions as the inner race rotates
with the shaft.

looseness

124
 Eng. Islam Ramy

Looseness:
• Rotating looseness: caused due to excessive clearance
between rotating & stationary elements of the machine
such as an bearing, while non-rotating looseness is
looseness between two normally stationary parts (bearing
pedestal , foot & foundation).

• Structural looseness: occurs where there is a weakness in

baseplate, foundation or feet & the machine is able to rock
from side to side.

Rotating Looseness:

125
 Eng. Islam Ramy

Rotating Looseness:
• Severe rotating looseness will produce sub harmonics
at 0.25X , 0.5X & also 1.5X , 2.5X etc….

Rotating Looseness:

126
 Eng. Islam Ramy

structural Looseness:

structural Looseness:

127
 Eng. Islam Ramy

Bearing pedestal Looseness:

Belt drive analysis

128
 Eng. Islam Ramy

Belt drives forcing frequencies:

• Determine the drive & driven speeds
 Output speed = input speed * (input sheave diameter / output sheave
diameter)

𝑫𝟏
𝑺𝟐 = 𝑺𝟏 ∗ ( )
𝑫𝟐

Belt rate frequency:

• Sub-synchronous from both input & output speeds.

𝝅∗𝑺𝟏∗𝑫𝟏 𝝅∗𝑺𝟐∗𝑫𝟐
Belt Rate= OR
𝑳 𝑳
 S1- Input speed
 S2- output speed
 D1- Input sheave diameter
 D1- output sheave diameter
 L- Belt length (if the belt is cut & laid out full length)

129
 Eng. Islam Ramy

Pulley eccentricity:

Pulley eccentricity:

130
 Eng. Islam Ramy

Belt resonance:

Sheave misalignment:

131
 Eng. Islam Ramy

Worn Belt:

Rolling element bearing

132
 Eng. Islam Ramy

Rolling element bearings:

Rolling element bearings fault frequencies:

• Ball Pass Frequency Inner race (BPFI).
• Ball Pass Frequency Outer race (BPFO).
• Ball Spin Frequency (BSF).
• Fundamental Train Frequency (cage frequency) ( FTF).

133
 Eng. Islam Ramy

Rolling element bearings fault frequencies:

• Ball Pass Frequency Inner race (BPFI).
• Non-synchronous.
• Modulated by 1X RPM.

Rolling element bearings fault frequencies:

• Ball Pass Frequency Outer race (BPFO).
• Non-synchronous.
• Not Modulated Unless in applications where outer race
rotating it will be modulated by 1X rpm.

134
 Eng. Islam Ramy

Rolling element bearings fault frequencies:

• Ball Spin Frequency (BSF).
• Non-synchronous.
• Modulated by Cage (FTF) frequency.

Rolling element bearings fault frequencies:

• Fundamental Train Frequency (cage frequency) ( FTF).
• Sub-synchronous.
• Usually around 0.4X running speed.

135
 Eng. Islam Ramy

Vibration the complete picture:

Stage 1 (early stage):

136
 Eng. Islam Ramy

Stage 2 (early stage):

Stage 2 (early stage):

137
 Eng. Islam Ramy

Stage 3 :

Stage 3 :
• Time waveform shows impacts & modulation
depending on the defect (inner, outer or balls).

Rollers defect

UT 1 - E FFL UE NT WAT E R PUM P

10P-9174 C-P1H Pump I nboard Horizontal
20
Route Wave form
16/Jan/17 11:08:08
15
PK = 7.13
L OAD = 100.0
RPM = 1501. (25.02 Hz )
10
PK(+) = 17.24
PK(-) = 15.86
5 CR E ST F= 3.42
Acceleration in G-s

-5

-10

-15

-20

0 50 10 0 15 0 20 0 25 0 30 0 35 0 40 0
T ime in mSe c s

138
 Eng. Islam Ramy

Stage 3 :

Stage 4 :

139
 Eng. Islam Ramy

Stage 4 :

Enveloping & Demodulation:

• Step 1: High pass filter.

140
 Eng. Islam Ramy

Enveloping & Demodulation:

• Step 2: Rectifying.

Enveloping & Demodulation:

• Step 3:Set low pass filter.

141
 Eng. Islam Ramy

Enveloping & Demodulation:

• Step 4:Perform FFT & analyze it.

Gearbox Analysis

142
 Eng. Islam Ramy

Gearbox forcing frequencies:

• Gear Mesh Frequency: The gear mesh frequency, also
called “tooth mesh frequency”, is the rate at which gear
teeth mate together in a gearbox. It is equal to the
number of teeth on the gear times the rpm of the gear.
A gearbox will always have a strong vibration
component at the gear mesh frequency, and it is one of
the fault frequencies used in machinery monitoring.
 GMF = Number teeth * Shaft RPM.
𝒊𝒏𝒑𝒖𝒕𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒕𝒆𝒆𝒕𝒉
 𝒐𝒖𝒕𝒑𝒖𝒕 𝒔𝒑𝒆𝒆𝒅 = 𝒊𝒏𝒑𝒖𝒕 𝒔𝒑𝒆𝒆𝒅 *
𝒐𝒖𝒕𝒑𝒖𝒕 𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒕𝒆𝒆𝒕𝒉
 𝑺𝟏 ∗ 𝑻𝟏 = 𝑺𝟐 ∗ 𝑻𝟐

 S1 – Input speed.
 T1 - 𝒊𝒏𝒑𝒖𝒕𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒕𝒆𝒆𝒕𝒉.
 S2 - Output speed .
 T2 - 𝒐𝒖𝒕𝒑𝒖𝒕 𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒕𝒆𝒆𝒕𝐡.

Gearbox forcing frequencies:

• Gear Mesh Frequency ex:

 S1 = 1482 RPM = 24.7 HZ

 T1 = 31 teeth
 T2 = 49 teeth
 T3 = 16 teeth
 T4 = 64 teeth
Calculate :
S2, S3, GMF1, GMF2

143
 Eng. Islam Ramy

Gearbox forcing frequencies:

• Gear Mesh Frequency ex:
 S1 = 1482 RPM = 24.7 HZ
 T1 = 31 teeth
 T2 = 49 teeth
 T3 = 16 teeth
 T4 = 64 teeth
Calculate :
S2, S3, GMF1, GMF2

 GMF1 = 24.7 * 31 = 765.7 HZ

 S2 = 765.7 / 49 = 15.6 HZ
 GMF2 = 15.6 * 16 = 250.02 HZ
 S3 = 250.02 / 64 = 3.9 HZ

Gearbox forcing frequencies:

• Hunting tooth:
 The hunting tooth frequency (HTF) is the rate at which a particular tooth on one gear
mates with a particular tooth on the other gear. If the numbers of teeth on the gears
are a simple ratio such as 1:2 or 1:3, the HTF will be equal to the RPM of the larger
gear, but if the numbers of teeth have no common factors, the HTF may be very low.
Gear pairs with low HTFs will wear more evenly and last longer than ones with a
relatively high HTF. The HTF is equal to the gear mesh frequency divided by the least
common multiple of the numbers of teeth on the gears.

144
 Eng. Islam Ramy

Gearbox forcing frequencies:

• Hunting tooth:
24 Tooth Gear 1 x 24 84 Tooth Gear 1 x 84
2 x 12 2 x 42
3x8 3 x 28
4x6 4 x 21
6 x 14
7 x 12
𝐺𝑒𝑎𝑟 𝑀𝑒𝑠ℎ 𝐹𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 ∗ 𝐻𝑖𝑔ℎ𝑒𝑠𝑡 𝑐𝑜𝑚𝑚𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟
𝐻𝑢𝑛𝑡𝑖𝑛𝑔 𝑡𝑜𝑜𝑡ℎ =
𝑇1 ∗ 𝑇2

Assume Input speed of 2000 RPM

(48000 x 12)/(24 x 84) = 285.7 CPM. That means that 285.7 times per minute,
those bad teeth (one on each gear) will enter mesh together and generate a very
high vibration pulse. That may seem rather high but usually the CF is not as high
as 12 - often it will be 1.

Gear load:

145
 Eng. Islam Ramy

Eccentric gear or Bent shaft & backlash :

Misaligned gears:

146
 Eng. Islam Ramy

Tooth Wear:

Cracked or Broken Tooth:

147
 Eng. Islam Ramy

Time Waveform:

Hydraulic & Aerodynamic

148
 Eng. Islam Ramy

Forcing frequencies:
• Vane pass frequency
• Blade pass frequency
• Pocket pass frequency
• Etc…

Rotor eccentric :

149
 Eng. Islam Ramy

Cavitation:

Recirculation & flow turbulence:

150
 Eng. Islam Ramy

AC induction Motors

Forcing frequencies:
• Line Frequency (LF) = Electrical line frequency - normally 60
Hz (3600 RPM) or 50 Hz (3000 RPM).
• 2X Line Frequency (2X LF) = This is a common frequency
found on a high resolution spectrum.
• P = # of poles on the motor. The number of poles is how the
speed of the motor is controlled. The greater the number of
poles, the slower the motor runs. The number of poles is
always an even number (2, 4, 6, etc.).
• FSynch = Synchronous electrical speed = 2 x LF/ P.
• FSlip = Slip frequency = FSynch - rotor RPM (actual speed)
• FPole = Pole pass frequency = P x FSlip
• WSPF = # Winding Slots x RPM
• RBPF = # Rotor Bars x RPM

151
 Eng. Islam Ramy

Eccentric Stator & soft foot:

Eccentric rotor:

152
 Eng. Islam Ramy

Loose rotor bars:

Cracked/Broken rotor bars:

153
 Eng. Islam Ramy

Resonance

Definitions:
• Resonance: A vibratory condition where a natural frequency and an
excitation frequency coincide. Resonance results in high vibration, and
may reach damaging levels. It is of paramount importance that a
machine not be operated at a speed that corresponds to a natural
frequency of the structure.

• Natural Frequency: The natural frequency is the frequency at which a

mechanical system will continue to vibrate after the excitation signal is
removed.

• Critical Speed: The critical speed of a rotor is an operating range where

turning speed equals one of its natural frequencies due to bending or
torsional resonances. If a rotor is operated at or near a critical speed,
it will exhibit high vibration levels, and is likely to be damaged. Much
rotating equipment is operated above its lowest critical speed, and this
means it should be accelerated relatively rapidly so as not to spend any
appreciable time at a critical speed.

154
 Eng. Islam Ramy

Natural Frequency:

Natural Frequency:

 Fn – Natural frequency
 K – stiffness
 M- mass

155
 Eng. Islam Ramy

Bode plot:
• A type of spectrum plot which consists of a graph of
amplitude vs frequency and a graph of phase vs
frequency. In most vibration analysis work the phase
spectrum is not important and is either ignored or not
recorded. In two-channel vibration measurements, such
as transfer functions and frequency response
measurements used for modal analysis, phase is of vital
importance

Structural resonance :

Fig 1 Fig 2

Fig 3

156
 Eng. Islam Ramy

Impact (Bump) test:

180° Phase shift

Natural frequency = 27.34 HZ

157