You are on page 1of 393

Intermediate Vibration

Analysis
IMRAN AHMAD
Director Technical
SUMICO Technologies (Pvt) Ltd
+92 321 427 6092
imran.ahmad@sumico.com.pk
www.sumico.com.pk

Timings

1st Session

9000-1030

Tea Break

1030-1045

2nd Session

1045-1300

Lunch Break(Prayer)

1300-1400

3rd Session

1400-1530

Tea Break

1545-1600

4th Session

1600-1700

Basics of Vibration
Introduction:
What is Vibration?

Principle of Vibration Analysis

Mechanical faults generate


unique vibration
Geometry of the machine
diameter of the shaft,
number of bearing
elements, etc.

Turning speed (e.g. RPM)

Mechanical Defects detected


with vibration analysis
Belt drive faults
Imbalance
Misalignment
Bent shaft
Looseness
Machine resonance
Cavitation
Shaft Rub
Bearing Defects including:
cage defect
outer race defect
inner race defect
rolling element defect
Gear defects
Electrical faults

What is Vibration?
What is Vibration ?

Vibration is the motion of a body about a


reference position caused by a force

In simple terms vibration is :-

A response to some form of excitation

The free movement of the shaft in a


journal bearing will cause it to
vibrate when a forcing function is
applied

What is Vibration?

Vibration is a pulsating motion of a machine or a machine part from


its original position of rest and can be represented by the formula:

Vibration Amplitude Response =

Dynamic Force
__________________
Dynamic Resistance

Vibration from Mechanical Faults

Vibration from Mechanical Faults

Vibration from Mechanical Faults

Vibration from Mechanical Faults

Vibration from Mechanical Faults

Vibration from Mechanical Faults

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

ISO 10816-3

Vibration standards are guidelines

Just
Tolerable

Allowable

Just Tolerable
Allowable
Allowable

Good

Good

15 kW<
Medium Machines
<75kW

Small
Machines< 15 kW

<300 kW
foundations

Group K

Just Tolerable

on special

Group M

Good
Large Machines
with rigid and heavy
foundations whose
natural Frequency
exceeds
machine speed

Group G

45
28
18
11.2
7.1
4.5
2.8
1.8
1.12
1.71
0.45
0.28
0.18

Velocity mm/s RMS

Not
Permissible

10 times = 20dB

Not
Permissible

Not
Permissible

2.5 times = 8dB

ISO2372 ( BS 4675 , VDI 2056 )

Vibration Characteristics
Amplitude

How Much

Frequency

How Often

Phase

When

Vibration Characteristics
Amplitude

Amplitude is the magnitude of vibration expressed in terms of signal level (millivolts or milliamps) or in engineering units
( Micron, mils, milli meter per second or inch per second)
There are many ways of measuring vibration amplitude levels, the most common are: peak to peak, zero to peak, root mean
square (RMS), average and crest factor.

Zero to peak or peak is the measurement from the zero line to the top of the positive peak or the bottom of the negative peak.

Peak

1.414 x RMS

RMS

0.707 x PEAK VALUE

Peak to peak is the distance from the top of the positive peak to the bottom of the negative peak. This measurement is used
most often when referring to displacement amplitude

Pk-Pk

2 x PEAK VALUE

Vibration Characteristics

The average value is 0,637 times the peak of a sin wave;


average values are measured by most analog meters.

Avg

The crest factor is determined by dividing the peak value by the


RMS value. For a true sine wave

Crest Factor = 1/.707 = 1.414

0.637 x PEAK VALUE

0 to Peak

RMS
Average

Amplitude

Time t

Peak to Peak

Vibration Characteristics

There are three types of measurements used to display


amplitude. These are

Displacement

Velocity

Acceleration

Vibration Characteristics
Displacement
is the distance that shaft moves in relation to reference point. The total movement of the shaft is
measured in Peak to Peak.

Velocity
is the displacement of the shaft in relation to time? It is measured in RMS (Root Mean Square) or
Peak.

Acceleration
is defined as the change in velocity over time. With this value we want the maximum impact (Force)
generated, so we use the Peak or RMS measurement.

Vibration Characteristics

Amplitude Units Metric

Displacement

Pk-Pk

Velocity

mm/sec

RMS

Acceleration

gs

Pk

Vibration Characteristics

Amplitude Units English

Displacement

mils

Pk-Pk

Velocity

inch/sec

Pk

Acceleration

gs

RMS

Vibration Characteristics - 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

Ex15

0.35
5
140

Low frequencies require very


Forlittle
normal
speed
forceoperating
to move an
object
ranges, velocity data provides
the best indication of machine
Increasing the frequency that
condition

0.30
120
4
PK
Accelerationin
inMicrons
G-s
P-P
Displacement
RMS Velocity in mm/Sec

A8 - Example 15
-F1H Fan Inboard Horizontal

0.25
100
3
0.20
80

Route Spectrum
22-Aug-02 11:30:50

0.10
40
1
0.05
20

20000

40000
Frequency in CPM

Label: Large Fan Unit - Easy

High and low frequency


events can be seen

Displacement measures low


frequency events ignoring
high frequencies
Relative shaft motion

00
0

OVERALL= 3.45 V-DG


RMS==.3909
PK
3.44
P-P
104.98
LOAD = 100.0
RPM = 831. (13.85 Hz)

the objects move with the same


velocity, the force needed to
move it increases, thereby
reducing the distance it can
travel

0.15
60
2

Velocity is the default unit for


standard data collection
techniques

60000

Acceleration accentuates the


high frequencies ignoring the
low frequencies
Good for early bearing
detection (Whenever there
is Metal to Metal Impacting
involve)

Vibration Characteristics
Vibration Amplitude Measuring Units

Gs or in/s2
(180 deg phase lead)

Acceleration, a
Velocity,

Acceleration

v=a/2

Velocity

Mm/ses or inches/sec
(90 deg phase lead )

Displacement = a/4

f2

Displacement

m 0r .000inch or Mils
90

90

Time

Acceleration

Velocity

Displacement

Vibration Characteristics
Frequency
Measure of the number of cycles of vibration that occur in a specific period of time

Tells us at what rate the vibration is occurring

Reciprocal of the Period (T)

Measured in Hz /CPM

Converted by a factor of 60

CPM relates directly to machine RPM


The time required to complete one full cycle of vibration

Frequency =

1
1 =
=
Period
T

Cycles
Second

Vibration Characteristics

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

2.

Hz Cycles Per Second

3.

CPM / 60

Orders Multiples of Turning Speed

1CPM = 1RPM

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)

Vibration Characteristics
Frequency
The table below demonstrates the relationship between the different frequency units over a range of frequencies.

Motor Turning Speed = 1500 RPM


CPM

1500

2250

3000

6000

12000

Hz

25

37.5

50

100

200

Orders

1.5

Vibration Characteristics what to use ?

Stress = Displacement
0-600 CPM

Fatigue = Velocity
600-120,000 CPM

Force = Acceleration
Above 120,000 CPM

Vibration Institute
recommendation

Vibration Characteristics
Significance of Frequency

Essential to pinpoint the cause of a machine problem

The forces that causes vibration are usually generated through


the rotating motion of the machine parts. These changes in
direction and amplitude according to rotational speed of the
machine components, most vibration problem will have
frequencies that directly related to the rotational speed.

Vibration frequency is an analysis or diagnostics tool

Vibration Characteristics - What is Phase?

Phase is the measure of time


difference expressed in degrees
between two events occurring at the
same frequency

Phase is the relationship of vibration


motion with respect to an other
vibration part or fixed reference point

Vibration Characteristics - What is Phase?


180
B

A & B are
180 degrees
out of phase

1 Cycle = 360
90

A
B

B
A

A & B are
90 degrees
out of phase

Two Types of Phase

Absolute phase

Absolute phase is the relationship of the Peak of vibration and a fixed reference Signal
(once per revolution)

Relative Phase

Relative phase is the relationship between two Peaks of vibration signals

Absolute Phase
Phase lag angle between once per turn marker and
first positive peak in a vibration waveform

Express in degrees phase lag


Must be filtered to multiple of turning speed

b
Absolute Phase

= a/b X 360o

Relative Phase
o

110

Phase lag angle between positive peaks of two


separate vibration signals (equivalent events)
The two signal must be same vibration unit
(eg. vel & vel or displ & displ)

Velocity
signal B

Velocity
signal A

Signal B lags signal A by 110o

Significance of Phase
-

Phase measurements are not taken during routine data collection of


predictive maintenance

However, when developing problems are found comparative phase


readings can provide valuable information pinpointing the specific
problem

Conversion of Units

Metric Units

where: D=Peak-Peak Displacement (m Pk-Pk)


V= Peak Velocity (mm/sec Pk)
A= Peak Acceleration (gs Pk)
F= Frequency (CPM)

V = DF / 19,100

V = 3690A / F

A = DF2 / 70,470,910

D = 19,100V / F

A = VF / 3690

D = 70,470,910 / F2

Conversion of Units

English Units

where: D=Peak-Peak Displacement (mils Pk-Pk)


V= Peak Velocity (inch/sec Pk)
A= Peak Acceleration (gs Pk)
F= Frequency (CPM)

V = DF / 19,100

V = 93,640A / F

A = DF2 / 1,790,000,000

D = 19,100V / F

A = VF / 93,640

D = 1,790,000,000 / F2

Time Waveforms
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

Ex15

0.4

The plot shown on the right is


a complex time waveform.
Amplitude versus Time
This is just one format
(domain) for analysing
vibration data.
Data can also be analysed in
a Spectrum (Amplitude Vs
Frequency) through a process
known as the FFT

Route Waveform
22-Aug-02 11:33:16

0.3

0.2
EX 8

1.0
RMS Velocity in mm/Sec Acceleration in G-s

A8 - Example 15
-F2V Fan Outboard Vertical

PK = .1495
LOAD = 100.0
RPM = 832. (13.86 Hz)

Los - Example 8
-P2V Pump Outboard Vertical

PK(+) = .3263
Analyze
PK(-) Spectrum
= .3572
15-Nov-95
CRESTF=10:00:16
3.38

0.1

RMS = 1.27
LOAD = 100.0
RPM = 737.
RPS = 12.28

0.0
0.8

-0.1
0.6

-0.2

-0.3
0.4
-0.4
0

50

100

0.2

150
200
Time in mSecs

250

300

350

Time: 120.44
Ampl: -.07595

0
0

6000

12000

18000

Frequency in CPM
Label: Looseness

24000

30000

Freq: 736.86
Ordr: 1.000
Spec: .245

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 sine's and cosines.
Meaning if you take a time waveform and mathematically calculate the vibration
frequency, it can be converted to a more familiar format

Amplitude

How the Vibration Spectrum is Created

Amplitude

Amplitude

Tim
e

y
enc
u
eq
Fr

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.

Ex15

0.4

A8 - Example 15
-F2V Fan Outboard Vertical

0.3

PK = .1495
LOAD = 100.0
RPM = 832. (13.86 Hz)
PK(+) = .3263
PK(-) = .3572
CRESTF= 3.38

0.1

0.0

-0.1

Los - Example 8
-P2V Pump Outboard Vertical
Analyze Spectrum
15-Nov-95 10:00:16
RMS = 1.27
LOAD = 100.0
RPM = 737.
RPS = 12.28

0.8

RMS Velocity in mm/Sec

Acceleration in G-s

0.2

EX 8

1.0

Route Waveform
22-Aug-02 11:33:16

0.6

0.4

-0.2
0.2

-0.3

-0.4
0

50

100

150
200
Time in mSecs

250

300

350

Time: 120.44
Ampl: -.07595

0
0

6000

12000

18000

Frequency in CPM
Label: Looseness

24000

30000

Freq: 736.86
Ordr: 1.000
Spec: .245

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
EX3

1.0

Analyze Spectrum
15-Nov-95 10:00:16

Here the primary cursor is at 1


Order (1xTs). All the other cursors
are harmonics (exact multiples of
the primary cursor)

RM S Velocity in mm/Sec

0.8

Los - Example 3
-P2V Pump Outboard Vertical

0.6

RMS = 1.27
LOAD = 100.0
RPM = 737.
RPS = 12.28

0.4

Therefore:
0.2

When the primary cursors is located on 1Order all the harmonics


will be synchronous
0

6000

12000

18000

Frequency in CPM

24000

30000

Freq: 736.86
Ordr: 1.000
Spec: .245

Harmonic cursors can be used to show non-synchronous and subsynchronous harmonics depending upon the energy of the primary
frequency

Energy in the Spectrum


E4

0.5

C1 - Example 4
-MOH MOTOR OUTBOARD HORIZONTAL
Route Spectrum
09-Feb-00 12:41:33
OVRALL= .5785 V-DG
RMS = .5716
LOAD = 100.0
RPM = 2937.
RPS = 48.95

RMS Velocity in mm/Sec

0.4

0.3

0.2

0.1

0
0

20

40

60

80

100

Frequency in kCPM

120

140

160

Freq: 2.937
Ordr: 1.000
Spec: .01038

Synchronous Energy
EX 8

1.0

Synchronous energy - related to


turning speed.
We can see from the spectrum
that the first peak is at 1 Orders
(which means it is 1 x turning
speed)

RMS Velocity in mm/Sec

All the other peaks are harmonics


off, which means they are related
to the first peak

Los - Example 8
-P2V Pump Outboard Vertical
Analyze Spectrum
15-Nov-95 10:00:16
RMS = 1.27
LOAD = 100.0
RPM = 737.
RPS = 12.28

0.8

0.6

0.4

0.2

0
0

6000

12000

18000

24000

Frequency in CPM
Label: Looseness

Examples of synchronous energy:


1) Imbalance
2) Misalignment

3) Gearmesh

30000

Freq: 736.86
Ordr: 1.000
Spec: .245

Non-Synchronous Energy
E5

2.0

Non-synchronous energy - not


related to turning speed

BF - Example 5
-R4A ROLL BRG. #4 - AXIAL
Route Spectrum
12-Jul-96 17:16:42

1.8

OVRALL= 2.63 V-DG


RMS = 2.69
LOAD = 100.0
MPM = 3225.
RPM = 380.

1.6

We can see from the


spectrum that the first peak
is at 10.24 Orders. This is
not related to turning speed.

RMS Velocity in mm/Sec

1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
0

6000

12000

18000

Frequency in CPM

24000

30000

Freq: 3888.9
Ordr: 10.24
Spec: .748

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:

Imbalance
Misalignment

Looseness

Vane pass frequency

1.0

Gears etc

0.8

RMS Velocity in mm/Sec

EX 8

Los - Example 8
-P2V Pump Outboard Vertical
Analyze Spectrum
15-Nov-95 10:00:16
RMS = 1.27
LOAD = 100.0
RPM = 737.
RPS = 12.28

0.6

0.4

0.2

0
0

6000

12000

18000

Frequency in CPM
Label: Looseness

24000

30000

Freq: 736.86
Ordr: 1.000
Spec: .245

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

Electrical

System resonances

Multiples of belt frequency

E5

2.0

BF - Example 5
-R4A ROLL BRG. #4 - AXIAL
Route Spectrum
12-Jul-96 17:16:42

1.8

OVRALL= 2.63 V-DG


RMS = 2.69
LOAD = 100.0
MPM = 3225.
RPM = 380.

1.6
1.4

RMS Velocity in mm/Sec

1.2
1.0
0.8
0.6
0.4
0.2
0
0

6000

12000

18000

Frequency in CPM
Label: Outer Race Defect
Priority: 1

24000

30000

Freq: 3888.9
Ordr: 10.24
Spec: .748

Lines of Resolution

Lines of resolution determine the clarity of the spectral data

The number of lines of resolution selected are divided into the maximum frequency
scale (Fmax) to arrive at the bin width (BW).

The better the resolution the more accurate the frequency displayed

BW = Fmax/LOR

The lines are actually centre frequencies of bins of energy

All the energy within the bin is summed up to give a single amplitude
frequency

Resolution
BW = Fmax / LOR
Bandwidth
Centre Frequency

Energy is summed up within a


Bin and plotted at the centre
frequency

Lines of Resolution
TA16

0.20
0.5

L2 - TA 16
-M1H Motor Outboard Horizontal
Analyze Spectrum
13-Mar-01 09:14:16
09:13:53
PK = .3852
.7078
LOAD = 100.0
RPM = 1497.
1496.
RPS = 24.95
24.94

PK Acceleration in G-s
PK Acceleration in G-s

0.16
0.4

The spectrum shown displays data


at 800 L.O.R with an Fmax of 1600
Hz

The second spectrum displays the


same data but with 3200 L.O.R over
the same Fmax

0.12
0.3

0.08
0.2

0.04
0.1

0
0

400

800
Frequency in Hz

1200

1600

VIBRATION MEASURING INSTRUMENTS

OVERVIEW

Choosing Your Instrumentation

What do you want to achieve?

What is your present and future budget for equipment & training?

Person power? Knowledge level?

Number of machines to be monitored?

Type of machines to be monitored?

Environment?

INSTRUMENT TYPES

Overall Level Meters

Quick Check Analyzers

FFT Data Collector/ Analysers

Full Feature Analyzers

Real Time Spectrum Analyzers

Instrument Quality Tape Recorders

Dedicated Balancing instruments

Real Time Rate

The Highest rate at which data can be captured and displayed without leaving
any gaps in the analysis.

Vibration Transducers

Overview
and
Selection

Transducer Types
Seismic:- Bearing relative to space.

Velocity Pickups

Accelerometers

Piezoelectric velocity pickups

Relative:- Shaft relative to bearing.

Non-contact Eddy Current Displacement Probes

Absolute:- Shaft relative to space.

Shaft Contact Displacement Probes (including Shaft Sticks and Shaft Riders)

Seismic Transducer
VELOCITY PICKUP

Velocity Pickups
ADVANTAGES

Self- Generating

No On-board Electronics

Strong signal

Note :- There are two types of velocity pickups the above advantages do not apply to
piezoelectric velocity transducers.

SESMIC TRANSDUCERS
ACCELEROMETERS

AMP

AMP

Accelerometers - advantages

No moving parts, no wear.

Rugged.

Very large dynamic range.

Wide frequency range.

Compact, often low weight.

High stability.

Can be mounted in any orientation.

Accelerometer Types
The three most common are :

Compression Type

Inverted Compression Type

Shear Type

Compression type accelerometer

Acoustic Shield

Seismic Mass

Piezoelectric Material

Preload Stud

Electric connector

Base

Mounting Stud
Receptacle

ICP Amplifier

Compression Type Accelerometers


Advantages

Relatively low cost

Disadvantages

Sensitive to base strain

Sensitive to Thermal transients

Can cause over-saturation and transducer settling problems

Widely used

Inverted compression type

ICP Circuit

Piezoelectric Material

Seismic Mass
Preload Sleeve

Mounting stud receptacle

Shear Type Accelerometer

Acoustic Shield

Piezoelectric Material

Seismic Mass
Electric connector
Post

Base

Mounting Stud
Receptacle

ICP Circuit

Advantages - Shear Type

Lower sensitivity to base strain

Large dynamic range

Much less sensitive to temperature transients

Stabilizes quickly when taking measurements at low frequencies.

Disadvantage:

Generally higher cost due to added components

Piezoelectric Velocity Pickup


Connector
INTEGRATOR

Mass

Piezoelectric Disks

Base

Non-contact Eddy Current Probe (Relative)

Non-contact Pickups
MAGNETIC
FIELD

NON-CONTACT PICKUP
PICKUP
COIL

SHAFT

OSCILLATOR

DETECTOR

DC
GAP
METER

AMPLIFIER

SIGNAL SENSOR

DISPLACEMENT
SIGNAL - TO
ANALYSER OR
MONITOR

Non-contact Eddy Current Displacement


Probes
USED FOR:

Relative Shaft Vibration.

Radial & axial shaft position.

Differential expansion between case and rotor.


Especially effective on machinery with high mass rigid casings and relatively low
mass rotors supported in journal type bearings.

N.C.P Problems & precautions

Only Measures Displacement - Sensitive Only to low frequency defects.

Subject to Mechanical and Electrical Run-out .

Units must be pre calibrated for specific shaft materials.

Shaft Contact Displacement Probes


(Absolute)

Shaft Sticks

Hardwood, fish-tail, fixed to accelerometer or velocity pickup

Measures vibration amplitude & phase

Shaft Riders

Shaft Rider
PICKUP MOUNTING STUD

SHAFT RIDER ASSEMBLY

MACHINE HOUSING

NON-METALLIC TIP

SHAFT SURFACE

Direct Contact : Absolute Measurements

Shaft Riders (permanently installed)

Shaft Sticks or Fishtails

safety issue

very useful below coupling of vertical pumps

Vibration Transducers
SensorsTransducersProbesWhat is it?
.It basically converts mechanical vibration
to an electrical signal

Accelerometer
Charge Type &
Line Drive
Constant Voltage &
Constant Current

Velocity Probe

Displacement
Shaft Riders
Proximity Probes
(Eddy Current Probes)

Accelerometers

Velocity Sensors

Displacement Probes (non-contact, eddy


current probes)

Typical Uses of Vibration Transducers


Radial Shaft
Vibration & Position

Radial Casing
Vibration

Axial Shaft
Vibration & Position

Monitoring Techniques
Relative Amplitude
100000000

Useful Operating Ranges

10000000
1000000
Accelerometer

100000
10000
1000

Velocity Transducer

100
Prox. Probe

10
0.1

10

100

1000

10000

Hz

Measurement Parameter

Acceleration

Velocity

Displacement

Find the flattest spectrum


Normally velocity is used
For very slow running machine
(< 600 RPM) displacement is
preferred.
For High frequency diagnostics use
acceleration
Always use acceleration for Envelope
analysis.

Monitoring Techniques
Vib

Acceleration
accentuates HIGH frequencies,
and attenuates LOW frequencies.

Acceleration

Velocity
flat treats all frequencies equally.

Velocity
Displacement
accentuates LOW frequencies,
and attenuates HIGH frequencies.

Displacement
Freq

Comparison of Transducers

Proximity Probes

Accelerometers

Advantages

Disadvantages

Easy to install
Good for detecting high
frequency faults
No moving parts
Good dynamic/frequency
range
Small/light weight
Withstands high temperatures

Measures directly shaft motion


Same transducers for axial
thrust, speed and radial
vibration
Measures directly in
displacements units
Measures DC (shaft position)
No moving parts

Needs double integration to


displacement
Needs external power source
Provides limited information on
shaft dynamic motion
Not good for slow speed machines

Runout problems
Sensitive to shaft materials
Installation
Limited freq. range. No detection of
rolling element bearing faults
Temperature restrictions
External proximitor needed

Frequency Range

Relative Sensitivity

30
dB
20

Useful Frequency Range


10% limit 0.3f0 3dB limit 0.5f0

10
Frequency Response of Sensor

0
-10
-20
-30
-40

2Hz

10KHz
Frequency [xf0]

Vibration Pickups

Journal
Bearings
instability

Shaft
Rotating
Speed

Blades

2x

Rolling Element
Bearings

Gear
3x

1 KHz

Non Contact Displacement

3KHz

Velocity Probe
Accelerometer

25KHz

Sensor Mounting - Frequency Range

Sensitivity

Mounting
Pad

Dual Rail
Magnet

Stud
Mount

Flat
Magnet

Hand
Probe

Freq
1.5KHz

10KHz

32KHz

Frequency Range
What is the frequency range of yourInstrumentCables
Sensor Sensor Coupling
Cable length ?

Sensor freq = 12 KHz


Instrument freq
= 80 KHz
Sensor coupling ?

What is the fault frequency you are looking for ?

Monitoring Techniques
Types of Bearings
Journal
JournalBearings
Bearings

Stationary
StationarySignals
Signals
Relative
RelativeLow
LowFrequency
Frequency

Use Proximity probes


Rolling
RollingElement
ElementBearings
Bearings

Modulated
ModulatedRandom
RandomNoise
Noise
Pulsating
Pulsatingsignals
signals
High
HighFrequency
Frequency

Use Accelerometers

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.

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

P = PumpF = Fan

The second character represented by a number indicates the location on the machine

M = Motor

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
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

Locating Turning
Speed

Turning Speed

When performing analysis on spectrums and waveforms, it is of utmost importance to


set the turning speed (running speed) correctly

When the turning speed has been located, the software will re-calculate all the
frequencies to this exact speed.

Once the turning speed has been set, it is now possible to determine what is
Synchronous/Non-synchronous and Sub-synchronous energy.

Turning Speed
The spectrum is showing numerous
impacts appearing at different
frequencies.
Ex6

12

By locating the turning speed, it is very


clear that the impacts are Nonsynchronous

BFI - Example 6
-P2V PUMP OUTBOARD VERTICAL
Route Spectrum
30-Jan-96 15:14:51
OVRALL= 13.52 V-DG
RMS = 13.46
LOAD = 100.0
RPM = 2987.
RPS = 49.78

10

RMS Velocity in mm/Sec

0
0

20

40

60

80

100

Frequency in kCPM

120

140

160

Freq:
Ordr:
Spec:

2.987
1.000
.115

Analysis Techniques Test


Have a look at the spectrum below.

- Where was the data taken?

EX3

1.0

When the turning speed has been


located

- What type of energy is present?


Los - Example 3
-P2V Pump Outboard Vertical
Analyze Spectrum
15-Nov-95 10:00:16
RMS = 1.27
LOAD = 100.0
RPM = 737.
RPS = 12.28

0.8

RMS Velocity in mm/Sec

0.6

P2V

Synchronous Energy

0.4

0.2

0
0

6000

12000

18000

Frequency in CPM

24000

30000

Freq: 736.86
Ordr: 1.000
Spec: .245

Things to Remember about a Route

A route includes information from one area only

A route does not have to include all the equipment defined in that area

The order of the equipment in the route can differ from that of the database

Equipment can appear in more than one route BUT can not appear in the same route
twice

Route measurement points may not include all the points configured on the
equipment

Route measurement points do not have to be in the same order as they appear in the
database

Data is not stored at the route level but in the database with the measurement points,
there for routes can be deleted but will not loose data from the database

A maximum of 50 routes can be stored to each area

Each equipment has a maximum of 144 points

And one route can only contain 1044 measurement points

Important: A route file contains the equipment and measurement


point IDs and definitions / speeds. For this reason the route does
not recognise points if their IDs are altered in the database

Frequency Bands

Divide spectrum in frequency bands based on the


types of mechanical faults that might appear on the machine

Frequency Bands

Divide spectrum in frequency bands based on the


types of mechanical faults that might appear on the machine

Imbalance

Frequency Bands

Divide spectrum in frequency bands based on the


types of mechanical faults that might appear on the machine

Imbalance
Misalignment

Frequency Bands

Divide spectrum in frequency bands based on the


types of mechanical faults that might appear on the machine

Imbalance
Misalignment
Looseness
Bearing Band 1

Bearing Band 2

Frequency Bands

Divide spectrum in frequency bands based on the


types of mechanical faults that might appear on the machine

Imbalance
Misalignment
Looseness
Bearing Band 1

Frequency Bands

Divide spectrum in frequency bands based on the


types of mechanical faults that might appear on the machine

Imbalance
Misalignment
Looseness
Bearing Band 1

Bearing Band 2

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

Anti-Friction Bearings (Peakvue)


Electrical

Resonance

Looseness

Gears

Sleeve Bearings

Belts

Unbalance

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.

Unbalance

Primary Types

Static or Forced

Dynamic

Coupled

Imbalance (Types)
Static Imbalance

Couple Imbalance

Dynamic Imbalance

Static Imbalance

The radial vibration readings are the highest amplitudes with the axial vibration
generally much lower in amplitudes

Static Imbalance will show a 0 phase shift across the rotor (vertical to vertical or
horizontal to horizontal) and 90 phase shift from vertical to horizontal at the same
bearing location

The phase angle will change the same amount the heavy spot changes if the system is
linear

Dynamic Imbalance

Any thing other than static

Requires more than one correction planes

Rule of thumb

If D/W < 4 two plan is required


D = Diameter of rotor, w = width of rotor

Two unequal/equal heavy spots 180 apart in separate planes on the same rotor or
located at some spacing other than 180.Amplitudes will differ or will be related to the
location and amount of heavy spot

Unbalance

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 unbalance

Periodic non-impacting sinusoidal waveform

Spectral peak at 1xTs (1 Order)

Very little axial vibration in case of static imbalance but high in case of overhung rotor

Similar amplitudes between horizontal and vertical plains for static imbalance and differ in case of
dynamic imbalance

90 phase shift from vertical to horizontal

Synchronous fault type

Amplitudes will increase with speed

Very low harmonics of 1xTs

Static unbalance

Force unbalance will be in-phase and steady

Amplitude will increase with the square of speed

1X RPM always present and normally dominates

Can be corrected by the placement of one weight in one plane

Dynamic/coupled unbalance

0-180 out of phase on the same shaft for dynamic & 180 out for coupled

1X RPM always present and normally dominates

Amplitude varies with square of increasing speed

Can cause high axial as well as radial amplitudes

Balancing requires Correction in two planes

Overhung Rotor unbalance

1X RPM present in radial and axial directions

Axial readings tend to be in-phase but radial readings might be


unsteady

Overhung rotors often have both force and couple unbalance each of
which may require correction

Eccentric Rotor

Largest vibration at 1X RPM of eccentric rotor

Horizontal and vertical phase readings differ by 0 or 180

Attempts to balance will cause a decrease in amplitude in one direction


but an increase may occur in the other direction

Unbalance Spectral Data

The spectrum shown represents a simple unbalance state

Single peak at 1xTs (1 Order)

Little indication of harmonics


Ex2

IF - Example 2
-F1H Fan Inboard Horizontal
Route Spectrum
16-Sep-99 08:36:29

What should the


waveform show?

RMS Velocity in mm/Sec

OVRALL= 4.58 V-DG


RMS = 4.56
LOAD = 100.0
RPM = 3000.
RPS = 50.00

0
0

20000

40000
Frequency in CPM

60000

80000

Freq: 3000.0
Ordr: 1.000
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


1.0

Ex2

IF - Example 2
-F1H Fan Inboard Horizontal
Waveform Display
02-Feb-00 15:13:51

0.8

PK = .5289
LOAD = 100.0
RPM = 2985.
RPS = 49.76

0.6
0.4

PK(+) = .8332
PK(-) = .8893
CRESTF= 2.38

Acceleration in G-s

0.2
-0.0
-0.2
-0.4
-0.6
-0.8

Changing the
units to velocity would reduce the amount of high frequency
-1.0
noise residing0 on 0.5the1.0waveform
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Revolution Number

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


Date: 07-Apr-00

RMS Velocity in mm/Sec

10

8
FAULT
6

0
0

100

200

300

Days: 07-Apr-00 To 21-May-01

400

500

Date: 21-May-01
Time: 14:24:29
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
Ex2

R M S V e lo c i ty i n m m /S e c

IF - Example 2
-F1H Fan Inboard Horizontal
Route Spectrum
16-Sep-99 08:36:29
OVRALL= 4.58 V-DG
RM S = 4.56
LOAD = 100.0
RPM = 3000.
RPS = 50.00

0
0

20000

40000
Ex2

1.0

60000

80000

Frequency in CPM
IF - Example 2
-F1H Fan Inboard Horizontal

3000.0
1.000
4.539

Waveform Display
02-Feb-00 15:13:51

0.8

PK = .5289
LOAD = 100.0
RPM = 2985.
RPS = 49.76

0.6

A c c e le r a t io n in G - s

Freq:
Ordr:
Spec:

0.4

PK(+) = .8332
PK(-) = .8893
CRESTF= 2.38

0.2
-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?

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

Operational Deflection Shape (ODS) is a technique that machine movement based


upon the phase and magnitude of data collected from the analyser. Shown below is
an image from the ODS illustrating the forces that are exerted onto the machine and
components when running in a misaligned condition

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(often .5-2 times the radial readings)

High radial levels of vibration at1xTs and/or 2xTs, 3x & 4x may also be present

Repeatable period sine waveform showing 1, 2,3,4 clear peaks per revolution (Most likely M or
W shape)

Data can usually be seen across the coupling

Phase reading will differ by 180 in axial or radial directions

Other visual observations may also be present like:

Excessive bearing temperature

Oil Leakage around the seals

Coupling worn /wear

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

Route Spectrum
15-FEB-93 11:04:18

OVRALL= 6.50 V-DG


RMS = 6.47
LOAD = 100.0
RPM = 2976.
RPS = 49.61

5
RMS Velocity in mm/Sec

0
0

10000

20000

30000

Frequency in CPM

40000

50000

Freq: 2925.0
Ordr: .983
Spec: 2.046

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

Route Spectrum
04-Aug-04 08:49:05

OVERALL= 6.33 V-DG


RMS = 6.31
LOAD = 100.0
RPM = 1071. (17.84 Hz)

RMS Velocity in mm/Sec

0
0

30

60
Frequency in kCPM

90

120

Freq: 1.071
Ordr: 1.000
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.
RPS = 49.93

20

Velocityin mm/Sec

PK(+) = 30.66
PK(-) = 26.81
CRESTF= 1.82

10

-10

-20

-30
0

0.5

1.0

1.5

2.0

2.5

3.0

Revolution Number

3.5

4.0

4.5

5.0

Misalignment
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
CRESTF= 3.42

0.2
Acceleration in G-s

0.0

-0.2

-0.4

-0.6

-0.8
0

0.4

0.8

1.2

1.6

2.0

Revolution Number

2.4

2.8

3.2

3.6

Rev :
Ampl:

.680
-.306

Angular Misalignment

Characterized by high axial vibration

180 phase change across the coupling

Typically high 1 and 2 times axial vibration

Not unusual for 1, 2 or 3X RPM to dominate

Symptoms could indicate coupling problems

Parallel Misalignment
1x

2x
4x

Radial

High radial vibration 1800 out of phase

Severe conditions give higher harmonics

2X RPM often larger than 1X RPM

Similar symptoms to angular misalignment

Coupling design can influence spectrum shape and amplitude

Bent Shaft

Bent shaft problems cause high axial vibration

1X RPM dominant if bend is near shaft center

2X RPM dominant if bend is near shaft ends

Phase difference in the axial direction will tend towards 180


difference

Bearing Misalignment

Vibration symptoms similar to angular misalignment

Attempts to realign coupling or balance the rotor will not alleviate the
problem.

Will cause a twisting motion with approximately 180 phase shift side to side
or top to bottom

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 0.5xTs and/or 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

May also truncation in the waveform

Phase varies and unsteady

Raised noise level around the 1xTs + harmonics

Half harmonics may also be present

Can be present in all Directions but often high in vertical direction

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.

M4441

R M S V e l o c i t y i n m m /S e c

4.0

40
-G2H

- Kiln Main Drive


Sh aft 01 Outboard Horizontal
Route Spectrum
06-Nov-02 11:02:11

3.5

OVERALL= 5.22 V-DG


RMS = 5.22
LOAD = 100.0
RPM = 635. (10.58 Hz)

3.0

2.5

2.0

1.5

1.0

0.5

0
0

200

400
600
Frequency in Hz

800

1000

Freq:
Ordr:
Spec:

10.58
1.000
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

1.5

L1 - Example 9
-P2A Pump Outboard Axial
Route Spectrum*
17-Aug-01 08:52:02
OVERALL= 6.62 V-DG
RMS = 6.13
LOAD = 100.0
RPM = 2974. (49.57 Hz)

RMS Velocity in mm/Sec

1.2

0.9

0.6

The raised noise level around the vane


pass frequency is common to a pumping
problem known as Cavitation

This would be the likely cause


of the impeller problem

0.3

0
0

40

80

120
160
Frequency in kCPM
Label: Centrifugal Pump - Medium

200

240

Freq: 14.88
Ordr: 5.002
Spec: .742

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 nonperiodic pattern.

M4441

1.2

40 - Kiln Main Drive


-G2H Shaft 01 Outboard Horizontal
Route Waveform
06-Nov-02 11:02:11
RMS = .3174
LOAD = 100.0
RPM = 635. (10.58 Hz)

0.8

PK(+) = .9797
PK(-) = .9874
CRESTF= 3.11

Acceleration in G-s

0.4

0.0

-0.4

-0.8

-1.2
0

50

100

150

200
250
Time in mSecs

300

350

400

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.
M4441

40 - Kiln Main Drive


-G2H Shaft 01 Outboard Horizontal
Trend Display
of
3-15xTS

-- Baseline -Value: .837


Date: 28-Feb-02

6
RMS Velocity in mm/Sec

5
FAULT
4
ALERT
3

0
0

10

20
30
Days: 28-Feb-02 To 16-Apr-02

40

50

Mechanical Looseness

Caused by structural looseness of machine feet

Distortion of the base will cause soft foot problems

Phase analysis will reveal aprox 180 phase shift in the vertical direction
between the base plate components of the machine

Mechanical Looseness

Caused by loose pillow block bolts

Can cause 0.5, 1, 2 and 3X RPM

Sometimes caused by cracked frame structure or bearing block

Mechanical Looseness

Phase is often unstable

Will have many harmonics

Can be caused by a loose bearing liner, excessive bearing clearance or


a loose impeller on a shaft

Rotor Rub
Truncated waveform

Similar spectrum to mechanical looseness

Usually generates a series of frequencies which may excite natural


frequencies

Sub harmonic frequencies may be present

Rub may be partial or through a complete revolution.

Gear Boxes

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.

Bevel Gear Advantages

These gears are most commonly seen on right angle gearboxes (where the
input shaft is at 90 degrees to the output shaft)

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

1.2

40 - Kiln Main Drive


-G1V Shaft 01 Inboard Vertical
Route Spectrum*
08-Jun-02 23:11:51
OVERALL= 2.22 V-DG
RMS = 2.14
LOAD = 100.0
RPM = 1548. (25.80 Hz)

0.9
RMS Velocity in mm/Sec

0.6

0.3

0
0

200

400
600
Frequency in Hz

800

1000

Freq: 386.98
Ordr: 15.00
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

Gear 2

= 71 Teeth

OUTPUT

GMF = #teeth x turning speed


GMF = 44teeth x 1490 RPM
GMF = 65560 CPM
1092.6 Hz

or 65560/60 =

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:

INPU
T

OUTPU
T

Input
1490RPM

Gear 1
teeth
Gear 2
teeth

= 15

Gear 3
teeth
Gear 4
teeth

= 19

= 21

= 54

Calculating GMF Multi Reduction


INPU
T

OUTPU
T

Input
1490RPM

Gear 1
teeth
Gear 2
teeth

= 15

Gear 3
teeth
Gear 4
teeth

= 19

= 21

= 54

Gear Ratio 1
Speed Out

= 15 teeth / 21 teeth
= 1490 RPM x 0.714

= 0.714
= 1064.28 RPM

Gear Ratio 2
Speed Out

= 19 teeth / 54 teeth
= 1064.28 RPM x 0.351

= 0.351
= 374.47 RPM

GMF 1
GMF 2

= 1490 RPM x 15 teeth


= 22350 CPM
= 1064.28 RPM x 19 teeth = 20221.32 CPM

GMF Calculation Exercise


Calculate

Speeds of all shafts

All GMF from the following gearbox arrangement


Input
RPM

INPU
T

= 1000

Gear 1
teeth
Gear 2
OUTP
teeth
UT

= 10

Gear 3
=teeth
0.25
Gear 4
teeth
0.25

= 10

= 40

Gear Ratio 1

= 10/40

Shaft 2 speed

= 1000 x

Gear Ratio 2

= 10/20

Shaft 3 Speed

= 250 x 0.5

GMF 1 = 1000 x 10

= 10000 CPM

GMF 2 = 250 x 10

= 2500 CPM

= 20

= 250 RPM

= 0.5
= 125 RPM

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.
X401A

1.0

FPP - SAND MILLS (OLD)A


-G3A Shaft 02 Inboard Axial
Route Spectrum
07-Nov -02 09:11:53
(SST-Corrected)

RM S V e lo c ity in m m /S ec

0.8

GM
F

OVERALL= 2.18 V-DG


RMS = 2.17
LOAD = 100.0
RPM = 310. (5.17 Hz)

0.6

0.4

Sideban
ds
0.2

0
0

8000

16000
Frequency in CPM

24000

Freq: 18363.
Ordr: 59.23
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
X401A

1.5

FPP - SAND MILLS (OLD)A


-G3A Shaft 02 Inboard Axial
Route W aveform
07-Nov -02 09:11:53

1.2

PK = .4580
LOAD = 100.0
RPM = 311. (5.19 Hz)

0.9

A c c e le ra t io n in G - s

0.6

PK(+) = 1.27
PK(-) = 1.13
CRESTF= 3.91

0.3
0
-0.3
-0.6
-0.9
-1.2
-1.5
0

3
Revolution Number

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:

Rolling Element Bearings

Outer Race

Inner Race

How Bearing Faults Generate Vibration

BALL SPIN (BSF)

CAGE (FTF)

INNER RACE (BPFI)

OUTER RACE (BPFO)

FTF & BSF

BPFI & BPFO

How Bearing Faults Generate Vibration

How Bearing Faults Generate Vibration

Rolling Element Bearings

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

0.376

1.894 3.009

4.991

25372

SKF I-26313

19

0.433

3.568 8.219

10.781

Rolling Element Bearings

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

Sidebands surrounding BPFO are much more serious than sidebands


surrounding BPFI (for fixed outer race)

The appearance of defect frequencies USUALLY starts with

BPFO & BPFI

Followed by BSF

Followed by FTF

These scenarios presume the absence of manufacturing errors on rolling element


bearing components

When a roller or ball defect is present from the start, BSF may well appear in the
spectrum WITHOUT any progression similar to these scenarios

Frequencies Generated By REBs

Random HF to ultrasonic
5KHz to 60 KHz

Component Fn
30KCPM to 120KCPM
54K to 96K for most

Defect Frequencies

Sum & Difference / Sidebands

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).

Stage 1: Fault Onset


Vibration Analysis (typical):
1) Standard FFT: no visible indication in spectrum
2) Spike Energy: slight increase in value (e.g. 0.25 gSE)
3) PeakVue: bearing frequency peak(s) corresponding to fault type
amplitude at 2-7 gs depending on type and location

Oil Analysis (typical):


1) Readings: Slight increase in elemental Fe, particle count, and WPC
2) Visual Ferrography:
Small platelet shaped particles (<30 ) from contact fatigue
Small spherical shaped particles (<5 ) from surface fatigue

3X

2X

1X

STAGE 1

ZONE C
BEARING
COMPON.
NATURAL FREQ.
REGION
30K

ZONE B
BEARING
DEFECT
FREQ. REGION

120K

STAGE 1

ZONE
A

BPFO/BPFI
2x
3x

ZONE C
BEARING
COMPON.
NATURAL FREQ.
REGION
120K

ZONE B
BEARING
DEFECT
FREQ. REGION

30K

ZONE
A

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).

Stage 2: Intermediate Wear


Vibration Analysis (typical):
1) Standard FFT: bearing defect rings in Zone C (natural freq.)
2) Spike Energy: increase in value (e.g. 0.50 gSE)
3) PeakVue: bearing frequency peak(s) with increasing harmonics
amplitude at 3-10 gs depending on type and location

Oil Analysis (typical):


1) Readings: elemental Fe stable but increase in particle count, WPC, and PLP
2) Visual Ferrography:
Platelet shaped particles (30-50 ) from contact fatigue
Possible spherical shaped particles (<5 ) from surface fatigue

STAGE 2

120K

ZONE C
BEARING
ZONE C
COMPON.
BEARING
NATURAL
COMPON.
FREQ.
REGION
NATURAL FREQ.
REGION

30K

120K

BPFO/BPFI
2x
3x
4x
5x
6x

3X

2X

1X

STAGE 2

ZONE B
ZONE
ZONE B
A ZONE BEARING
DEFECT
BEARING
A
FREQ.
DEFECT
REGION
FREQ. REGION

ZONE C
BEARING
COMPON.
NATURAL FREQ.
REGION
Bearing f

ZONE B
BEARING
DEFECT
FREQ. REGION

30K

ZONE
A

Failure Mode 3

The fundamental frequency normally appears at this stage

(First calculable frequency of the bearing towards the lefthand 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).

Stage 3: Severe Wear


Vibration Analysis (typical):
1) Standard FFT: bearing frequencies with harmonics and sidebands in Zone B
2) Spike Energy: increase in value (e.g. 1.0 gSE)
3) PeakVue: bearing frequency peak(s) with increasing harmonics and sidebands
amplitude climbs to 5-10 gs or higher (depending on type and location)

Oil Analysis (typical):


1) Readings: small change in elemental Fe, substantial increase in WPC and PLP
2) Visual Ferrography:
Sharp increase in large particles (>30), both platelets and cutting wear
Increased three-dimensional appearance to wear particles

ZONE C
BEARING
COMPON.
NATURAL FREQ.
REGION

STAGE 3

120K

ZONE B
BEARING
DEFECT
FREQ. REGION

BPFO/BPFI
2x
3x
4x
5x
6x
7x
8x
30K

120K

30K
BPFI

BPFI

BPFO

ZONE
A

3X

2X

1X

STAGE 3

ZONE C
BEARING
COMPON.
NATURAL FREQ.
REGION
Bearing f

ZONE B
BEARING
DEFECT
FREQ. REGION

ZONE
A

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).

Stage 4: Imminent Failure


Vibration Analysis (typical):
1) Standard FFT: discrete bearing frequencies replaced by broadband noise
2) Spike Energy: falling levels until just before failure, then levels rise sharply
3) PeakVue: bearing frequency peak(s) with increasing harmonics and sidebands
amplitude climbs to 10 gs or higher (depending on type and location)

Oil Analysis (typical):


1) Readings: small change in elemental Fe, substantial increase in WPC and PLP
2) Visual Ferrography:
Broad range of huge particles (75+) from fatigue and adhesion
Particle counts/ferrous density are excessive

2X
3X

RANDOM HIGH
FREQ. VIBRATION

ZONE C
BEARING
COMPON.
NATURAL FREQ.
REGION

STAGE 4

120K

ZONE B
BEARING
DEFECT
FREQ. REGION

BPFO/BPFI
2x
3x
4x
5x
6x
7x
8x
30K

ZONE
A

120K

STAGE 4

ZONE C
BEARING
COMPON.
NATURAL FREQ.
REGION
30K

ZONE B
BEARING
DEFECT
FREQ. REGION

1X

ZONE
A

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

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

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.

TBT

P-P Dis pla cement in Microns

16

Fu - Turbine Brg Thrust End


-R1Y Radial 'Y' Direction

Route Spectrum*
27-Jul-04 14:08:21
OVERALL= 2.93 V-DG
P-P = 22.71
LOAD = 100.0
RPM = 941. (15.69 Hz)

12

0
0

6
Frequency in Orders

12

Ordr:
Freq:
Spec:

1.000
15.68
7.494

As the clearance increases then the


harmonics of 1xTs will increase and
can go up to 1015xTs.

Like looseness the more


harmonics there are the more
severe the problem will be.
A good sleeve bearing will still
show a few harmonics as there
is a small clearance between
the shaft and bearing

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.

The dominant peak within the spectral data


will be typically at 0.4 orders. (.40-.48)

Fu - Turbine Brg Thrust End


TBT -R1Y Radial 'Y' Direction

16

Oil Whirl at 0.4 orders


P-P Displacement in Microns

12

Route Spectrum*
27-Jul-04 14:08:21

This defect is sub-synchronous data.

OVERALL= 2.93 V-DG


P-P = 22.71
LOAD = 100.0
RPM= 941. (15.69 Hz)

When the amplitude of the oil whirl is


equal to or greater than the 1xTs
peak a problem exists

In this instance oil whirl can be corrected


by:

0
0

6
Frequency in Orders

12

Ordr:
Freq:
Spec:

1.000
15.68
7.494

Properly loading the bearing

Change the oil viscosity

Change the oil pressure

Plain Bearings Spectral Diagnostics

Oil Whirl

Usually occurs at 42 - 48 % of running speed

Vibration amplitudes are sometimes severe

Whirl is inherently unstable, since it increases centrifugal forces


therefore increasing whirl forces

Oil Whip Instability


oil whip
oil whirl

Oil whip may occur if a machine is operated at 2X the rotor


critical frequency.

When the rotor drives up to 2X critical, whirl is close to critical and


excessive vibration will stop the oil film from supporting the shaft.

Whirl speed will lock onto rotor critical. If the speed is increased
the whip frequency will not increase.

Bearing Defects
Rolling Element
Plain Bearings

Peakvue

Peakvue

What is Peakvue

What is Peakvue?

Peakvue is a technology unique to CSI and means Peak Value

Such as the Peak Value of an impact generated by a bearing defect in a time waveform
- (True Peak Value)

If you have a 21XX analyzer you have the capability to acquire Peakvue Data

The True Peak Value is obtained by concentrating on Stress Wave Analysis rather than
conventional vibration data.

These stress waves travel further than conventional vibration signals so a truer
indication of fault severity is obtained.

What is Peakvue

What is a Stress Wave?

Stress waves accompany metal-metal impacting. These stress waves are short-term
(fractional to a few milliseconds) transient events, which introduce a ripple effect on
the surface machinery as they propagate away from the initial event.

If you think of a stone being dropped into a pool of water. The stone is the
initial impact generated by the fault. The effect of the stone being dropped
into the water cause a ripple on the surface of the water which, spreads
over a wide area.

Initial Impact

What is Peakvue

If a bearing has a sub-surface defect (early bearing wear), when a rolling element
passes over the defect it bends the race slightly and then as the rolling element
passes it restores back to its natural state.

This event causes a high frequency (1-50KHz) short duration stress wave.

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
Signals that are filtered from
the Peakvue Signal
Imbalance
Misalignment
Gears
Bearings
Resonance

2. Peakvue filter removing


low frequency noise from
the stress wave data
This is to prevent low
frequency noise
consuming the stress
wave activity

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


The band pass filter removes all the data above and below the filter corner values

High Pass Filter


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.

How Does It Work?


Vibration
Signal

High

Full

Pass

Wave

Filter

Rectify

Digital
Peak

FFT

Impact
Detection

A comparison can be made of the sampling to show how data is collected through
both methods of data acquisition, normal and Peakvue.

How Does It Work?

The diagram below shows sampling of data using normal data collection.
Stress wave- this is
missed under normal
conditions

Instantaneous Samples

How Does It Work?

The diagram below shows sampling of data using Peakvue data collection.

Stress wave- this is


missed under normal
conditions

Peakvue Samples

How Does It Work?

Peakvue measures the highest amplitude found in a stress waves (Pk Value) and
holds that data

The waveform data is then passed through a high pass filter to remove the
unwanted, low frequencies

Imbalance, Misalignment, Looseness, resonance etc.

This just leaves us with the high frequency impacting data (Peak) above the
machine noise level

The data is then brought back to fundamental frequency. (this allows analysis of
the data to be done quicker and easier)

PeakVue How does it work?


The waveform should contain enough time to include at least 15 shaft revolutions to
resolve cage frequency in the spectrum for rolling element bearings.

(The waveform time length is determined by the lines of resolution divided by the f-max)

PeakVue How does it work?


The f-max should be set at least 3 or 4 times the highest expected defect frequency
(usually inner race defect for rolling element bearings)

One average should be used when taking PeakVue data

PeakVue How does it work?


Transducer mounting should be consistent for trend able data.

At minimum the surface should be clean (free of paint, dirt, etc.), stress waves are
easily attenuated.

Filters
Types of filter available
Filter Calculations
Filter Guidelines

Filters Options

One of the key elements in acquiring meaningful peakvue data is the selection of filters

Selecting the wrong type of filter will result in poor quality data

To much noise filtered through (the spectrum becomes very noisy)

To much is filtered out (The stress wave is not allowed to pass through)

There are two types of filter options in Peakvue, these are:

1. Band Pass Filter

2. High Pass Filter

Each of the filters are designed to remove unwanted data out of the signal at the
appropriate levels

Filter Options - High Pass Filter

High Pass Filters remove all frequencies from the data below the filter setting but
allow the high frequency stress wave to pass through.

All low frequencies are


removed from the
input signal

Stress Wave data is allowed to


pass through the filter

High Pass Filter

Filter Options - Band Pass Filters

Looks for stress waves within a parameter defined by the filter setting. Frequencies
above and below this setting are removed from the data

Data passes
through filter

Data is filtered
out of the
signal

Data is filtered
out of the
signal

Filter Selection

To select the correct filter we need to consider the highest operational defect frequency
that we want to measure/detect. Then select the next available filter above that
frequency

E.g.

Consider a typical motor / pump arrangement. We have:


1 - 4 Pole A.C. Induction Motor
2 - 3 Jaw Coupling
3 - Centrifugal Pump

Typically the highest defect frequency


to emit from this machine would be?
1 - BPFI - Bearing Defect

Filter Selection

4 Pole Motor A.C Induction fitted with bearings SKF 6313

Defect Frequencies (Orders)


FTF - 0.384
BSF - 2.037
BPFO - 3.071
BPFI - 4.929

Typically we would want to see the 10th Harmonic of the BPFI

Highest defect frequency:


(BPFI x 10) x Turning Speed (Hz)
(4.929 x 10) x 25
1232.3 Hz

We would then select the next available filter setting above the frequency

Available filters

High Pass Filters

Band Pass Filters

500hz

20hz 150hz

1000hz

50hz 300hz

2000hz

100hz 600hz

5000hz

500hz 1khz

10000hz

20000hz

From our previous calculation of 1232Hz, What filter setting would we select?

Note: the meter will only allow you to select the next filter above
the specified Fmax.

Filter uses (Band Pass) - Guidelines

Band Pass Filters

20hz 150hz Felt problems on paper machines

50hz 300hz Certain structural resonance excitation, modulation of


gearmesh in low speed machinery

100hz 600hz Gearmesh modulation in intermediate speed


machinery.

500hz 1khz Gearmesh modulation

Tip: use bandpass filters when the event of interest is the excitation of a structural
resonance, or the modulation of known frequencies such as gearmesh.

Filter uses (Highpass) - guidelines

High Pass filters

500hz Low speed machinery having <125hz. Bearing & gearing problems

1000hz Intermediate speed machinery (<2000 rpm) with gear mesh <300hz

2000hz Medium speed machinery (<4000rpm) with gear mesh ,600hz

5000hz Machinery up to 9000rpm and gear mesh to 1500hz, Requires attention be


paid to how the sensor is mounted as well as the sensors frequency response.

10000hz High speed machinery with gear mesh up to 3000hz sensor must be
permanently mounted with a frequency response of 3db in the 30kHz or higher
range.

20000hz High speed machinery with gearmesh up to 6000hz. Sensor must be high
frequency and permanently mounted.

Tip: Use highpass filters when the objective is to detect stress waves which are emitted by metal
on metal impacting.

Filter Selection - Question

Consider:

Motor running at a speed of 1000RPM

Driving a fan unit via pulley belts

Fan Speed is 1350RPM


Motor Bearings = SKF 3095 - BPFI 4.855
Fan Bearings = SKF 6210 - BPFI 5.907

Calculate what Filter setting would be required for both the motor and the fan
bearings?

Filters Available:
500 Hz, 1000Hz, 2000Hz, 5000Hz, 10000Hz, 20000Hz. (High Pass)
20-150Hz, 50-300Hz, 100-600Hz, 500-1KHz. (Band Pass)

Filter Selection - Answers

Motor Speed = 1000CPM / 60 = 16.667Hz

Fan Speed = 1350CPM / 60 = 22.5Hz

Motor.

BPFI = 4.855

Defect Frequency = (BPFI x 10) x Turning Speed (Hz)

Defect Frequency = (4.855 x 10) x 16.667

Defect Frequency = 809.18 Hz

Filters Available:
1000Hz 2000Hz, 5000Hz, 10000Hz, 20000Hz. (High
500 Hz, 1000Hz,
Pass)
20-150Hz, 50-300Hz, 100-600Hz, 500-1KHz. (Band Pass)

Filter Selection - Answers

Motor Speed = 1000CPM / 60 = 16.667Hz

Fan Speed = 1350CPM / 60 = 22.5Hz

Fan

BPFI = 5.907

Defect Frequency = (BPFI x 10) x Turning Speed (Hz)

Defect Frequency = (5.907 x 10) x 22.5

Defect Frequency = 1329.07Hz

Filters Available:
500 Hz, 1000Hz, 2000Hz
2000Hz, 5000Hz, 10000Hz, 20000Hz. (High
Pass)
20-150Hz, 50-300Hz, 100-600Hz, 500-1KHz. (Band Pass)

Peakvue Data
Spectrums and Waveforms
Diagnostics Techniques

Peakvue - Spectrum

Here is a typical Peakvue spectra plot.


2. Units should be acceleration (Very
high frequency analysis)

1. Broad band energy Filtered Noise

3. Amplitude values are low. Severity of


fault is not determined in the spectra

This is typically a GOOD spectrum

Peakvue - Spectrum

This is a Peakvue spectrum where high frequency stress waves are being detected

Notice the Impacts passing


through the filtered noise

2. Units still in acceleration (Very


high frequency analysis)

1. Broad band energy Filtered Noise

3. Amplitude values are low. Remember


severity of fault is not determined in the
spectra

This is indication of a fault developing

Peakvue Processing Spectral Data

Shown below is a typical Peakvue spectrum with a defect present


Stress waves are showing clearly in the
data at 4.6 Orders

The filter used is shown in the top


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

RMS Acceleration in G-s

0.8
0.7

Route Spectrum
09-Jul-03 09:50:49
(PkVue-HP 1000 Hz)
OVERALL= 1.37 A-DG
RMS = 1.37
LOAD = 100.0
RPM = 1342. (22.37 Hz)
>NTN 6217
N=BPFO -OB

0.6
0.5
0.4
0.3
0.2
0.1

For Peakvue analysis

Use the Spectrum

Diagnose the defect

0
0

200

400

600

800

1000

Frequency in Hz
Route Waveform
09-Jul-03 09:50:49
(PkVue-HP 1000 Hz)
RMS = 2.97
PK(+) = 8.35
CRESTF= 2.81

8
Acceleration in G-s

7
6
5
4
3
2
1
0
0

12

16

20

Revolution Number
Label: Bearing Fault - BPFO NTN6217

24

28

32

36

Freq: 1.250
Ordr: .05587
Spec: .01367

Use the Waveform

Determine the severity

Peakvue - Waveforms

Waveforms can be confused with spectrums, as the waveform is only plotting the peak
value and does not show a full wave.
EX 1

0.6

A1 - Example 1
-D3P Tail Roll Non D/S Peakvue
Analyze Waveform
16-Mar-01 12:03:14
(PkVue- HP 500 Hz)

2. Peak Value Impacts

0.5

PK = .0556
LOAD = 100.0
RPM = 80.
RPS = 1.33

Acceleration in G-s

0.4

PK(+) = .5599
PK(-) = .0397
CRESTF= 14.25

0.3

0.2

1. Filtered Noise Level

0.1

4. Acceleration as
default units

-0.1
0

6
Revolution Number

Label: Easy

3. No Peak Negative Value

12

Peakvue - Diagnostics

Diagnosing a Peakvue spectrum and waveform is not to dissimilar to that of


conventional data.

However there are a few differences which can be a bit confusing at first, these are:

1. Do not try to locate 1xTurning Speed, as this is low frequency data and will be
filtered out.
Turning speed should be entered using the conventional spectral data.

2. Multiple harmonics are often present within a spectrum due to the way peakvue
samples the data.
These do not indicate Looseness

3. Spectral amplitudes are always low in amplitude but should not be used to judge
severity. Use the spectrum to diagnose the fault.

4. Waveforms indicate the severity of the problem.

Peakvue - Diagnostics

Continued..

5. Ensure the same filter setting is used in both the spectrum and waveform.
Potential faults can be missed or overlooked if different filters are used.

6. Cage Defects show up well in peakvue data and is normally an indication the
bearing is under stress.

7. All low frequency faults are removed from the data and will not be seen in a
Peakvue spectrum and waveform
Imbalance, Misalignment, Looseness, Resonance - All Gone.

Peakvue - Diagnostics
EX 1

PK Acceleration in G-s

0.016
0.012

A1 - Example 1
-D3P Tail Roll Non D/S Peakvue
ANALYZE SPECTRUM
16-Mar-01 12:03:14
(PkVue- HP 500 Hz)
PK = .0484
LOAD = 100.0
RPM = 80.
RPS = 1.33

1.Spectral data indicating a


defect at 5.463 Orders

0.008
0.004

3. Very Slow
RPM

0
0

20

40

60

80

100

Frequency in Hz

Acceleration in G-s

0.6
0.5

2. Impacting also being


detected at 0.6G-s

0.4

ANALYZE WAVEFORM
16-Mar-01 12:03:14
(PkVue- HP 500 Hz)
PK = .0556
PK(+) = .5599
PK(-) = .0397
CRESTF= 14.25

0.3
0.2
0.1
0
-0.1
0

6
Revolution Number

Label: Easy

12

Freq: 7.284
Ordr: 5.463
Spec: .01018

Peakvue - Diagnostics
EX 1

PK Acceleration in G-s

0.016
F

4.Fault Frequencies Indicate


A1 - Example 1
a BPFI
Defect
-D3P
Tail Roll Non
D/S Peakvue
F

ANALYZE SPECTRUM
16-Mar-01 12:03:14
(PkVue- HP 500 Hz)
PK = .0484
LOAD = 100.0
RPM = 80.
RPS = 1.33

0.012
0.008
0.004

>NSK 6207
F=BPFI -IB

0
0

20

40

60

80

100

Frequency in Hz

Acceleration in G-s

0.6
0.5

ANALYZE WAVEFORM
16-Mar-01 12:03:14
(PkVue- HP 500 Hz)
PK = .0556
PK(+) = .5599
PK(-) = .0397
CRESTF= 14.25

0.4
0.3
0.2
0.1
0
-0.1
0

6
Revolution Number

Label: Easy

12

Freq: 7.284
Ordr: 5.463
Spec: .01018

Peakvue Amplitudes - Rolling Element Bearings

For machines running between speeds of 900 - 3600RPM recommended guidelines


for setting initial warning levels in the Peakvue time - waveform are as follows:

Alert Value

Fault Value

Inner Race

3.0g's

6.0g's

Outer Race

6.0g's

12.0g's

Rolling elements fault

4.5g's

9.0g's

Cage frequencies

If evident then the bearing is usually under


stress.

Peakvue Amplitudes - Rolling Element Bearings

For machines running at speeds <900RPM recommended guidelines for setting initial
warning levels in the Peakvue time- waveform are as follows:

Peakvue Vs
Demodulation

Peakvue Vs Demodulation

What is Demodulation?

This is a technique which concentrates on stress wave analysis, but is not as


effective.

How Does it Work?

Demodulation looks for the ringdown that follows an impact, and tries to measure
how quickly it fades. In order to do this the Time Waveform has to be manipulated
in such away that the waveform data becomes useless

Less than 1ms

Initial Impact

Ringdown

Peakvue Vs Demodulation

What are the Differences?

Peakvue samples the data much quicker enabling it to catch the very short
duration high frequency stress wave. It then holds that Peak Value throughout
its parameter.

Due to the Analogue filtering system used by Demodulation, results in a delay


in response and the stress wave impact is missed

Demodulation registers

Peak Impact Detection registers

Peakvue Vs Demodulation
The Process!

Standard Demodulation

High

Low

Full

Remove

Pass

Pass

Wave

DC

Filter

Filter

Rectify

Bias

A/D
Converter

FFT

Enveloping Stage

Peak Impact Detection


High

Full

Pass

Wave

Filter

Rectify

Digital
Peak
Impact
Detection

FFT

Peakvue Vs Demodulation

Equipment
A conveyor system consisting of six rolls is driven by a motor/gearbox unit .
The motor speed is 1500RPM reduced through the gearbox giving the roller speed
to be 98.5RPM

Peakvue Vs Demodulation

Data was collected on each bearing of the conveyor system

Due to the slow speeds Peakvue and Demodulation Filters were both set to
500Hz High Pass using 1600 Lines of Resolution
Ex1

0.016
0.07
Max Amp
.0490

Direct Comparison
Demodulation
Between the Two
Spectra
Peakvue Spectra

0.06

PK Acceleration in G-s

0.012
0.05
PK Acceleration in G-s
PK Acceleration in G-s

PvD - Example 1
PvD - Example 1
Ex1
-R1P Roller Bearing Drive End (18-Dec-01)
-R1P Roller Bearing Drive End

0.05
Spectrum
Analyze Spectrum
14:22:30
18-Dec-01 14:10:26
0.04
(Demod-HP
HP500
500Hz)
Hz)
(PkVue-

.0295
PK = .1238
0.03
100.0
LOAD == 100.0
RPM == 99.
97.
RPM
RPS = 1.64
1.62
RPS
0.02

0.04

0.008

0.01

Peakvue

0.03

0
18-Dec-01
14:10:26

0.02
0.004

Demodulation

0.01

18-Dec-01
14:10:26
0
0

50
50

100
100 in Hz
Frequency
Frequency in Hz
Label: Peakvue
500Hz
HighHigh
PassPass
Demodulated
500Hz

150
150

200
200

Peakvue Vs Demodulation
Ex1

PvD - Example 1
-R1P Roller Bearing Drive End

Plot
Span

Demodulation
Waveform

Acceleration in G-s

1.0

18-Dec-01 14:22

Peakvue Waveform
-0.2
18-Dec-01 14:10

Revolution Number

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

0.6

Ex7 - Example 7
-M1H Motor Outboard Horizontal
Route Spectrum
08-Nov-00 14:27:35
OVERALL= .5613 V-DG
RMS = .5607
LOAD = 100.0
RPM= 2967. (49.44 Hz)

0.5

RMS Velocity in mm/Sec

0.4

0.3

0.2

0.1

0
0

500

1000
Frequency in Hz

1500

2000

Freq: 100.00
Ordr: 2.023
Spec: .386

The spectral plot is showing a


peak at 100Hz (6000cpm)

2xLf

This can be mistaken


for misalignment

Electrical Defects Waveform Data


The waveform data from a 100Hz peak will show a sinusoidal pattern like the
waveform shown below

Ex7

1.5

Ex7 - Example 7
-M1H Motor Outboard Horizontal
Route Waveform
08-Nov-00 14:27:35

1.0

RMS = .5291
LOAD = 100.0
RPM= 2967. (49.44 Hz)

Velocity in mm/Sec

0.5

PK(+) = 1.50
PK(-) = 1.77
CRESTF= 3.31

-0.5

-1.0

-1.5

-2.0
0

3
Revolution Number

Again this type of pattern can


be associated with
misalignment.

Usually misalignment
would produce higher
force (Higher waveform
levels) than those from
electrical defects due
to the stress being
applied to the machine

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.

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

Belt Tension (tension on the belt holds it tightly against the sheave)

Friction between the belt and sheave

The arc of contact between the belt and sheave (Wrap)

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

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 crosssection, 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 subsynchronous 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
Belt Length
= 2000mm

= 300 mm

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)

1 x Belt Frequency
showing with
harmonics
Dominant 2 x Belt
Frequency

There are a lot of sub-synchronous


peaks showing in this data.

The first peak is the


fundamental frequency of
the belt rotation.

The second peak is the


2xbelt 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.

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

3
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

180 phase change occurs when shaft speed passes through resonance

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 dont
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

Route Spectrum
13-Feb-03 10:14:46

24

OVERALL= 19.95 V-DG


RMS = 19.85
LOAD = 100.0
RPM = 1484. (24.73 Hz)

21

RMS Velocity in mm/Sec

18

In order to help resolve this issue we


need to check the amplitude of the
1xTs 90 degrees to this point
(horizontal to vertical)

15
12
9
6
3
0
0

500

1000
Frequencyin Hz

1500

2000

Freq: 24.72
Ordr: 1.000
Spec: 19.50

This can easily be done by


using the multi point plot in
the software

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.

M4551

40 - No 1 GCTCompressor
- Multiple Points (13-Feb-03)

24
20

Max Amp
22.0

16
12
8
4
0
M2V 10:15

RMSVelocity in mm/Sec

M2H 10:14

M1V 10:14
Point= M2H
13-Feb-03
10:14:46
RPM= 1484.
M1H 10:14
0

500

1000
Frequency in Hz

1500

2000

Freq:
Ordr:
Sp 3:

25.00
1.011
19.35

Rotary Pumps , Fans &


Compressors Analysis

Hydraulic and Aerodynamic Forces

CAVITATION

Cavitation will generate random, high frequency broadband energy


superimposed with VPF harmonics

Normally indicates inadequate suction pressure

Erosion of impeller vanes and pump casings may occur if left unchecked

Sounds like gravel passing through pump

Hydraulic and Aerodynamic Forces


VPF = VANE PASS
FREQUENCY

If gap between vanes and casing is not equal, Blade Pass Frequency
may have high amplitude

High VPF may be present if impeller wear ring seizes on shaft

Eccentric rotor can cause amplitude at VPF to be excessive

Hydraulic and Aerodynamic Forces

FLOW TURBULENCE

Flow turbulence often occurs in blowers due to variations in pressure or


velocity of air in ducts

Random low frequency vibration will be generated, possibly in the 50 2000 CPM range

Hydraulic and Aerodynamic Forces

SURGE

Surge can occur if the pressure developed by the compressor is not


equal to or greater than the downstream pressure

Random low frequency vibration will be generated, possibly at 30-45 %


of compressor of compressor speed

Digital Signal Processing

Digital Signal Processing

Signal Integration Mode

Analog

Analog integration integrates data in the time domain as collected , then performs FFT.

Digital

Performs the FFT first then integrates in the frequency domain

The Vibration Meter

Accelerometer

Raw Signal

Amp

Integrator
1x, 2x

High Pass
Filter

Amp

Low Pass
Filter

Detector
P-P or RMS

AC
Output

Display
Reading

DC
Output

Digital Signal Processing


Machine Vibration Signal

Time Signal

Machine Vibration Signal


Time Signal

Absolute Vibration
with Free-Space

Machine Vibration Signal


Relative Vibration with
mounting position of Prox. Probe

AC Signal

DC
Signal

Bandpass Measurement
Peak
RMS
Avg
Peak
to
Peak

Freq. = 1/Time

Freq. = Hz
= rev. per second

Always ask.... Are you measuring RMS or Peak , etc ??


What is the frequency range ??
How much averaging?

Machine Freq are function of RPM


ie. rev. per minute

Detector
Apeak peak

ARMS

A peak

RMS

For Sine waves only:

Apeak

T = averaging period

True peak - peak

1
2
a
(t )dt
T 0

A peak
Apeak peak
RMS

2 * ARMS

Machine Signal Types


Stationary Signals
- Vibration from rotating machines

Non - Stationary Signals


- Vibration from reciprocating machines (short term)
- Vibration from run-ups and coast-down

Frequency Analysis
Types of filters:
f

High-Pass filters

- As the name imply, a high pass filter allows high


frequencies to pass. (lower frequency limit)

Low-Pass filters

- Allow low frequencies to pass through


(upper limit)

Bandpass filters

- Allows only frequencies within the band


f

Anti-aliasing filters

- Low pass filter at half the sampling frequencies

Vibration Amplitude ?

Frequency Range Selection?

Lower
Freq.
limit?

No. of lines? Avgs?

Upper
Freq.
limit?

Freq / Orders?

BW = Fmax /LOR

*BW, bandwidth ,the spacing between each line


The bandwidth shoul not be no greater than 5Hz or 300CPM

Time to Collect one average (sample time) = 1/BW

Tmax (Time waveform length) = Sample size / Sample rate


Where sample size = 2.56*LOR & sample rate = 2.56*Fmax

Discrete Fourier Transform

The signal that comes to the analyzer is analog signal. It must be digitally sampled by
the analyzer. This process is a variation of FFT and is known as DFT.

For DFT the waveform is re-created in the analyzer by digitally sampling and then
transformed into the frequency domain.

To under stand the FFT digital sampling process ,we must have the under standing of:

L.O.R

Fmax

Length of Waveform

Digital Sample Size

FFT (DFT) - Pitfalls


Discrete Fourier Transform (DFT) - Pitfalls
FFT - Fast Fourier Transform is an efficient means of calculating a DFT (Discrete
Fourier Transform). Basically, it transform a time signal into a frequency spectrum.

1.1. Aliasing
Aliasing-- high
highfrequencies
frequenciesappearing
appearingas
aslow
lowfrequencies
frequencies
2.2. Leakage
Leakage-- Memory
Memorycontents
contentsforced
forcedtotobe
beperiodic.
periodic.

Can
Cangive
givediscontinuities
discontinuitieswhen
whenends
endsjoined
joined

3.3. Picket
Picketfence
fenceeffect
effect

Actual
Actualspectrum
spectrumsampled
sampledat
atdiscrete
discretefrequencies.
frequencies.
Peaks
Peaksmay
maybe
bemissing
missing

FFT pitfalls - Aliasing Effect

Sampling rate too slow


High frequency analysis results in false low frequency signal

Solution: Use Anti-aliasing filter


Typically a 1K (1024 point) transform, 512 frequency components are calculated
and 400 lines displayed. Similarly a 2K transform 800 lines are displayed.

FFT pitfalls - Leakage


+ve
1st Sample

-ve

..give discontinuities
when ends joined

2nd Sample
Use Hanning Window

+ve

-ve

FFT - Picket Fence Effect


Actual
Spectrum

Measured
Spectrum

Waveforms

Waveforms

Just like the spectral there are certain patterns and characteristics to look for when
conducting waveform analysis.

Once the characteristics have been identified, the analyst can rule out certain faults

e.g: if the waveform is periodic faults like Looseness, Bearing defects, Cracks could be
ruled out.

Some Characteristics are:


1.Amplitude

2.Periodic

5.Asymmetry 6.Modulation

3.Impacts
7.Restrictions

4.Discontinuities

Waveform Analysis

Time - seconds

Waveform Analysis
Periodicity

Waveform Analysis

Periodicity

Waveform Analysis

Asymmetry

Waveform Analysis

Impacting

Waveform Analysis

Distortion

Vibration Analysis

Complex

Waveform Analysis
Electrical vs Mechanical

Waveform Analysis
Noise

Waveform Analysis
Extended time

Waveform Analysis
Extended time

Waveform Analysis
Low frequency

Waveform - Beats
WIDEBAND SPECTRUM

F1 F2
ZOOM
SPECTRUM

A beat is the result of two closely spaced frequencies going into and out of
phase

The wideband spectrum will show one peak pulsating up and down

The difference between the peaks is the beat frequency which itself will be
present in the wideband spectrum

Crest Factor

Introduction
The Crest Factor is equal to the peak amplitude of a waveform divided by the
RMS value. The purpose of the crest factor calculation is to give an analyst a
quick idea of how much impacting is occurring in a waveform. Impacting is
often associated with roller bearing wear, Cavitation and gear tooth wear.

In a perfect sine wave, with an amplitude of 1, the RMS value is equal to .707, and
the crest factor is then equal to 1.41. A perfect sine wave contains no impacting and
therefore crest factors with a value higher than 1.41 imply that there is some degree of
impacting

Crest Factor

The Problem with the Fast Fourier Transform (FFT)


The definition of the Fast Fourier Transform implies that any signal can be
approximated by the sum of a set of sine waves. Unfortunately, this doesnt work
so well when one has a signal that consists of non-periodic events, impacts or
random noise . Both impacts and random noise appear the same in the spectrum
although they mean different things in the context of machinery vibration analysis.
The crest factor is therefore useful in giving the analyst a quick idea of what is
occurring in the time waveform.

Crest Factor

Comparison of 2 Waveforms
In below figures we can see an example of the use of the Crest Factor. The
waveform in figure on left has a crest factor of 3.01. The waveform in figure on right
has a crest factor of 1.61. The data in figure on left represents a machine with
serious rolling element bearing wear, and the crest factor is relatively high due to
the amount of impacting occurring within the bearing. The data in figure on right
represents a machine with an unbalance, but no impacting related to bearing wear.

Crest Factor

Conclusion
The Crest Factor is a quick and useful calculation that gives the analyst an idea of how
much impacting is occurring in a time waveform. This is useful information that is lost if
one is only viewing a spectrum as the FFT cannot differentiate between impacting and
random noise. Impacting in a time waveform may indicate rolling element bearing wear,
gear tooth wear or Cavitation. Quite often, the Crest Factor is trended over time in order
to see if the amount of impacting is increasing or not.

Advance Analysis

Rotor Dynamics

Terminology & Introduction

Types of bearings

Tilting Pad

elliptical

Standard Tilting Pad Bearing

grooved

Turbine Supervisory Instrumentation

HP Turbine

LP Turbine

A typical 250MW Steam Turbine

Generator

Exciter

Turbine Supervisory Instrumentation


Valve Position

Axial Thrust

(LVDT/Potentiometer)

(Dual-voting)

Radial Vibration
XY Probes
(at each bearing)

Phase
(once per turn)

HP Turbine
Thrust
Collar

Turning
Gear

Eccentricity

Casing
Expansion

Generator

LP Turbine
Y

Differential
Expansion Probes

(LVDT)

Prox. Probe

Typical TSI Configuration for a Large Steam Turbine

Exciter
Y

Instrumentation, Measurement Types

Thrust

Eccentricity

Case expansion

Differential
expansion

Radial vibration

Speed

Phase

Temperature

Valve position

Differential Expansion
6210

Bearing
Vibration
6110

Valve
Position
6410

Bearing
Vibration
6110

Eccentricity
6220

Bearing
Vibration
6110

Case Expansion
6410

Speed/Phase
6312

Bearing
Vibration
6110

Dual Thrust Position


6210
Temperature
6220

Thrust

Thrust is a position measurement, and is one of the most critical


measurements on a high speed steam turbine

At least two thrust sensors, for redundancy, and for voting logic, are installed
to measure axial rotor position

Thrust bearing deterioration, failure, or sudden changes in steam pressure


could quickly move the rotor axially and could cause collisions with rotor and
stator

Thrust or Rotor Position


Tach

HP

LP

IP

HP

GEN

EXC

IP

Thrust bearing
Thrust
failures
bearing
can failures
result inare
extremely
one of the
costly
most
repairs or even
machine replacement
catastrophic failures

Eccentricity

Eccentricity is a measure of rotor bow


during start-up

Eccentricity is measured with at least one


displacement sensor in the HP section
with an eccentricity collar as the target

This parameter is typically monitored


from 0 to 600 rpm

When a turbine is stopped during an


outage, uneven heating at the top of the
case and the weight of the rotor will
introduce bow in the rotor

During start-up, this bow is worked out


before the turbine is brought online

Eccentricity
Tach

HP

IP

LP

GEN

EXC

Over 600 RPM, Eccentricity can be set to zero by the monitor


Or, over 600 rpm, vibration measurements are monitored

Case Expansion

Case expansion is a position measurement of the external


case of the turbine as it expands during start-up

LVDTs (Linear Variable Differential Transformers), the position


sensors, are mounted on both sides of the foundation with
sensor tips connected to the case. The LVDTs measure
case expansion relative to the foundation.

Large steam turbines, greater than


250MW, are not fixed at the HP section.
The turbine pedestal is allowed to slide on
the sole plate as the case expands when
heated during start-up. While the case
expands, if one side of the turbine binds,
the case may bend, which could cause
the rotor to contact the stator.

Differential Expansion

Upon start-up, due to mass differences, the rotor expands at a different rate
than the case

Differential expansion (D.E.) is a position measurement to insure that the rotor


does not collide with the case

By mounting the D.E. sensors on the case, and then measuring the position of
the rotor, the result is the difference in position of the rotor relative to the case

D.E. is recommended for turbines 250MW and greater

SHAFT

c
d = BEVEL ANGLE
a = DETECTED RANGE
c = INDICATED RANGE

Differential Expansion
Case Expansion
Tach

HP

IP

LP

GEN

EXC

Differential Expansion (difference between


case and rotor expansion)

Shaft Vibration

Shaft vibration is monitored for rotor


position, rotor motion and bearing
housing motion

For turbines up to 250MW, where


case is massive as compared to
rotor, relative vibration
measurements are recommended

For turbines 250MW to 650MW or


greater, absolute vibration
measurements are recommended

For aeroderivative turbines where


rolling element bearings are used,
accelerometers with RMS
processor and PeakVue processors
are recommended

Turbo Machines Malfunction Detection

Rub

Bow

Imbalance

Looseness

Misalignment

Couplings

Runout

Fluid instabilities

Shaft cracks

Turbo Machinery Instrumentation


1

Y
2

3
X

Machine Case
4
Emerson TIE Seminar May 2007

Prox. Probe Systems

Proximity Probe Systems


Pitfalls and Hints

Displacement Probes
Extension
Cable
-24V
Com
Signal

Oscillator/
Demodulator

Connector

Probe
Also known as non-contact, eddy current probes, proximity probes

Integral
Cable

Mounting Blocks

API 670 gives examples of probe mounting blocks which employ a split clamp
housing design, as shown below. However, these can be a little difficult to use, so
alternative designs are available
10mm slots
for mounting bolts

Holes for
Mounting bolts

10

20

10

25

40

20

Clearance hole
for probe

60

60

Drilled and tapped M10 for probe

Whatever the design, however, the mounting must be rigid enough so as not
to cause excessive vibrations. It can therefore be necessary to calculate the
natural frequency of the adapter. It should lie at least 10 times higher than
the machine rpm

Probe Installation
Internally Mounted Probes

Externally Mounted Probes


Bushing screwed
into housing

Cable
Conduit

Cable
Ties
Cable
Conduit

Probe Holder
screwed into
Bushing

Reverse threaded
Probe

Probes can be mounted externally to the machine with the probe mounted through
the machine case or bearing housing, or they can be mounted inside the machine
and the probe leads exited through the machine case.

API 670 shows examples of typical mounting techniques.

Probe Installation

Displacement Probes

The probe, integral cable, extension cable and oscillator/demodulator make up a tuned resonant
circuit.
In order to maintain a proper ratio between gap and voltage the various parts must be
calibrated and matched

Never indiscriminately substitute one cable type for another. The systems are matched and cable
length is critical

A typical probe calibration curve looks as the following:


Output,V

Probe too far from shaft

Gap the probe at


its mid point
Probe in contact or
very near shaft

+
Linear Range
Typicall y 10 to 90 mils gap
(250 to 2250um)

Probe Sensitivity
= 200mV/mil
(8mV/um)

Gap

Gapping Probes

First inspect the surface of the shaft.


Remember to inspect the complete area
that the probe can look at, i.e. take into
account the thermal expansion of the shaft

With the machine stopped, gap the probes


to their mid point of operation

The probes can be gapped mechanically


using feeler guages or more accurately by
measuring the DC output from the
proximitor (typically 50mils shaft-to-probe
gap, giving -10VDC output. API
recommends that the probe be gapped to
0.2V)

-10V DC

50mils
50mils

Grapping Axial probes


Total Thrust Movement = (V2 - V1) x Scale Value
Zero Point Offset =
Midpoint Voltage +- (Total Thrust Movement)/2

Gap the probe so that the DC output


is at the midpoint of its linear range
when the rotor is in the middle of its
normal thrust float zone.

Mech. & Elect. Run Out

Probe cannot tell the difference between vibration and runout caused by defects such
as scratches, dents and variations in conductivity and permeability

Runout can be measured by observing shaft displacement at low speed (typically 300 to
600 RPM on a high-speed machine)

To avoid runout, ensure that the probe is always mounted to observe a smooth journal
surface

Surface finish is specified by API 670 at 16 to 32 inches (0.4 to 0.8um) rms


obtained by honing and burnishing.
Demagnetized so that the total mechanical and electrical runout do not exceed 25%
of the max. allowed peak-to-peak vibration or 0.25mils (6um) which ever is greater.
The gauss level of the probe surface should not exceed +-2 gauss with a
variation of less than 1 gauss.

Side Clearance & Cross talk

Probe Calibration Kit

Machine Operating Stages

The need for Adaptive Monitoring Strategy


Steady State 1
Start
Up

Coast
Down

Stop
Process
Variable

Simple Machineries Motors, Pumps, Fans etc

Steady
State 2

Machine Operating Stages

The need for Adaptive Monitoring Strategy


Steady State
Start Up

Bump
In the night

Slow
Roll

Coast Down

Stop
Process
Variable

Step
Up
(holding)

Turbo Machineries Compressors, turbines etc.

Monitoring Key Vibration Indicators


Aux. machines

Turbo-machines

Roller Element Bearing Machines

Journal Bearing Machines

Overall vibration trends


Analysis bands trends
Spectrums
Waveforms shape
Fault Freq components
Harmonics & sidebands

Vibration vectors
Slow roll vectors
Orbit / Mode shapes
Synchronous amplification factor
Anisotropic Stiffness
Rotor position
Heavy Spot

Vibration monitoring plots


Aux. machines

Turbo-machines

Roller Element Bearing Machines

Journal Bearing Machines

Vibration trends (overall & bands)


Multiple trends
FFT Spectrum
Order Spectrum
Order Tracking
TSA
Zoom Analysis
Waveform
Bearing Analysis (PeakVue, Evelop,etc)
CPB Analysis
Cepstrum

Orbits
Shaft centerline (ASCL plots)
Polar or Nyquist plots
Position plots (Acceptance region)
Bode plots (Amplitude & phase vs RPM)
APHT plots for crack detection
(Amplitude & phase vs Time)

Cascade plots (FFT vs RPM)


Waterfall plots (FFT vs Time)
Order Tracking
Long time waveform
Waveform
FFT Spectrum (half & Full)
Bump in the night
Live & post processing of data
(Extraction)

Vibration Signals important components

Amplitude
Phase
Frequency
Position
Form / Shape
Precession

Types of vibration measurements


Accelerometer

Absolute Casing Vibration


Absolute motion of bearing (casing)
Typically Better indicator of roller bearing forces
Gives early indication
velocity and acceleration,

Casing
Vibration

Relative shaft vibration

Relative Shaft
Vibration

Relative motion of shaft in bearing (wrt mounting position)


Related to rotor dynamics / oil film
Typically displacement proximity probes

Displacement
probe

Absolute shaft vibration


Sum of casing + relative vibration
Typically Dual probe
Absolute
Shaft
Vibration

Shaft

Basic Principles
Force
Vibration =
Dynamic Stiffness

Resonance Freq. res =

K (stiffness)
M (Mass)

Dynamic stiffness is a combination of


Mass Stiffness, Quadrature Stiffness and Spring Stiffness.
Mass Stiffness
Rotor
Quadrature Stiffness Lubrication, Fluid/Bearings,
Process Fluid
Spring Stiffness
Bearing, shaft, Pedestals, Foundation

Rotor
Mass

F spring
K
Support
Stiffness

Vibration Vectors
Shaft rotation

0o
ax
i

270

Vibration
Sensor

75 um pp

V axis
225

phase lag

180

The vibration vector contains both


the amplitude and absolute phase
information from a filtered raw vibration signal.

ax

is

The axis in line with the sensor is also known as the U


axis . The V axis is always 90o from the U axis
from the opposite the direction of rotation.

90o

225o

75 um pp

Filtered vibration signal

A one pulse per turn signal is required for phase and


the phase (angular position) is measured in the
direction opposite to the shaft rotation and
wrt to the sensor as ref.

Monitoring Basics

Roller bearings vs Journal Bearings


1. Are you able to get a repeatable measurement?
If machine is operating under a fix process condition (eg. Same speed & load)
you should be able to get a repeatable reading. If not ask WHY?
2. How can we tell if a machine is good or bad?
A good healthy machine should have a stable vibration amplitude under
the same operating and process condition.
When the machine condition deterioratesmost of the time we will
See an increases in vibration. Some situation you will see a decrease
in vibration.
Now we have a machine and when we measure its vibration 6 months ago
..it amplitude was 3 mils with a phase angle of 0o.
Today the vibration reading is still 3 mils with a phase angle of 180o.
Since vibration amplitude 6 months back and now remains constant is the machine
Good or bad?

Why monitor vectors?


What is the difference between a Vibration = 3 mils @ 0o and Vibration = 3 mils @ 180o

6 mils
Change !!
ie. Increase by 2x

Important to watch for phase changes in addition to amplitude

Rotation & Precession


Shaft
Rotation

Precession

+
+

+
+

Rotation is the angular motion of the rotor.


Precession is the lateral motion / vibration /orbit / shaft center line of the rotor.
Precession can be in opposite direction of rotation

Rotation & Precession


Precession
Bright spot = trigger

Blank = notch

Shaft
Rotation

Forward precession happens when the direction is the same as rotation.


If direction of precession is opposite to rotation than it is known as reverse precession

Heavy & High spot


Heavy spot (HS) is the angular location of
unbalance of the rotor.
High spot is the location of the shaft which is closest to the
Vibration probe. High spot is the response of the rotor system
to the unbalance within the system.

Heavy spot
High spot

Shaft
rotation
+

Isotropy & Anisotropic Systems

Isotropic (Symmetric)
No Split Resonances
Circular 1X Orbit
Vibration not orientation dependent

Anisotropic (Asymmetric)
Resonances can be split
Elliptical 1X Orbits
Vibration depend on probe orientation

Slow Roll Speed / Slow Roll Vector


Slow Roll Speed is the rotor speed low enough that the dynamic effects of
unbalance are negligible. Typically less than 10% of the first balance resonance
speed.

Slow Roll Vector is a constant (or very slowly varying) vibration vector that
represents the nondynamic vibration observed by the vibration sensor.

Important as it provides information on thermal, electrical and mechanical


effects, rotor bow, mechanical & electrical runout, coupling problem.

Slow roll compensation = Run out correction

Slow Roll Compensation


Bode & Polar plot

1X
Phase

0o
Phase Slope

Uncompensated

90o

270o
Compensated

Slow roll speed range

Amplitude (p-p)

180o

If full scale is 4mil pp then


SR=1 mil pp @ 130o

Slow roll range select region when both Phase


& amplitude is relative flat (min. changes)

Uncompensated

Compensated

Frequency (rpm)

Slow roll compensation removes mechanical and electrical


Run-out thus represents only the dynamic response of the rotor

Transient Information

Slow roll speed & runout vector

Resonance frequency (critical speed, split resonance)

Vectors information (Amplitude & phase, preloads)

Synchronous Amplification Factor

High & heavy spot relationship

Mode shape / deflection shape

Mode shapes
1st Mode

2nd Mode

HP Turbine

3rd Mode

LP Turbine

Generator

Exciter

Amplitude (pp)

Synchronous Amplification Factor (SAF )


a

Peak
Ratio

70%

SAFPR= a/b

Half-power Bandwidth

b
c d e

SAFhp= e-c/d

RPM

The Synchronous Amplification Factor is a measure of


effective damping in the rotor system

Amplitude (pp)

Phase

Synchronous Amplification Factor (SAF )

The Synchronous Amplification Factor is a measure of


effective damping in the rotor system

SAF < 5 enough damping in system


SAF 5~8 moderate damping in system
SAF > 8 potential problem

RPM

1st & 2nd Mode


0o

1st mode

Phase

2nd Mode
Heavy Spot

Amplitude (pp)

2nd mode

90o

270o

RPM
1 Mode
Heavy Spot
st

180o

1st mode

2nd mode

Where should be the shaft center?


Shaft
rotation

Bearing
Center
++

Shaft
Center

++
+
+
+
++

Shaft center line plots


Can we get a shaft center line plot with only one probe?

Overhung rotor
1st Mode

At rest

++
+
+
+
+
+

++
+
+
+
+
++

Shaft center line plots

Shaft Center line

Shaft misalignment

Estimated shaft deflection shape

View direction, X-Y position?

Driver to Driven is the standard view convention

Transducer that is most CCW is Y regardless of rotation or


viewing direction
Y
Y

CW

CCW
Shaft
Rotation

Shaft
Rotation

Driver
Driver
* However you will need to chose one direction and stick to it

Vibration Data Plots

Timebase Plots

Direct Orbit

Average Shaft Centerline Plots

Polar Plots

Bode Plots

APHT Plots

Trend Plots

XY Plots

Half Spectrum

Full Spectrum

Cascade

Waterfall

Compensated / Filtered Orbit

Compensated / Filtered Timebase

Analysis Display Plots

FFT Spectrum (Half spectrum)


Raw Machine Time Signal
Sample 1

Sample 2

Filtering

Filtering

Window Function

Window Function

Detectors

Detectors

FFT

FFT

Also known as
Narrow band analysis,
constant bandwidth analysis
Half spectrum
Autospectrum
Spectras, Signature

/n =
Avg

FFT Spectrum 1

FFT Spectrum 2

FFT Spectrum

Time Synchronous Averaging


Tacho

Sample 2

Sample 1

Raw Machine
Time Signal

1. Samples triggered by tacho


2. Averaging waveforms
3. FFT is done on averaged waveform

Filtering

Filtering

+
Window Function
Detectors

FFT
Averaged Time Signal
Enhanced synchronous components while Non synchronous and
noise will be averaged out.
Reduced vibration effect from nearby machine.
Isolate specific machine components

TSA Spectrum

Cascade, Waterfall, Campbell plots


What is their difference?

Amp

Rpm
Time

Hz

Cascade is spectra vs speed


Waterfall is spectra vs time

Campbell displace 3D Cascade/waterfall


Plot in 2 Dimension.
The size of the circle represent the vibration amplitude

Full Spectrum - Spectrum of an Orbit


Y Probe
Waveform

X Probe
Waveform
FFT

FFT

X Probe
half spectrum

Y Probe
half spectrum

Shaft
Rotation
Direction

Full Spectrum
Transform

1x

Reverse
Precession
-2x

-1x

+
Forward
Precession
2x

Orbits

Shape can help


determine rubs, oil
instabilities,
imbalance

AC, dynamic
motion or the path
of the shaft
centerline as seen
by the transducers

Orbits

The orbit represents the path of the shaft centerline relative to a pair of perpendicular, coplanar sensors.
Information:

Determine P-P amplitude


Determine precession direction
Estimate Phase
nX & frequency information
Freq. ratio information
Shape (loading direction)
Shaft deflection shape (multiple orbits)

Direct Orbits
Filtered Orbits
Compensated Orbits
Absolute Orbits

Unbalance

3:1 could still be unbalance


relative to structural mobility. UNUSUAL cases as high as 5:1

At 5:1 and above, the problem is too directional

3:1 and above


5:1 / 6:1 usual

Indications of ...

MISALIGNMENT

Misalignments
With looseness

5:1 to 6:1
Up to 8:1

P. 8-8

8-8

A ratio of 8:1 to 10:1


tells us...
that a HIGHLY DIRECTIONAL problem is at work
Either a rotor critical or structural resonance.
Eccentricity could also be the cause !!!

Light Rub

Heavy Rubs

Rubs

Characterized by slight or
heavy truncations of the
orbit

Each hit is marked by a


flattening of the orbit curve...

A Parallel with
Waveforms
Rubs cause truncation both in the
time domain (waveform) and the orbit...

From Bad to Worse...

8-12

Oil Whirl

Internal loop ROTATES about the axis of the orbit

Similarity ?

Extreme looseness DOES lead to rubbing

Looseness, rub or whirl ?

Looseness or rub

Looseness

orbit & secondary loop locked in-phase

Rub

displacement orbit versus shaft rotation

orbit rotation opposite to shaft rotation

Whirl

orbit loop rotates in forward precession

8-14

Orbits

Average Shaft Centerline (ASCL Plot)


Y
2

X
3

Average,
DC position
thefluid
shaft
Rotor should
rise inofthe
film
centerline
and load will influence final position
Should rest at the center and bottom of
Shaft centerline can help
the bearing at zero RPM
determine misalignment

Average Shaft Centerline (ASCL Plot)


The average shaft centerline plot is an XY plot of the average position of the shaft in the measurement plane.
These plots are constructed using the DC part of the vibration signal (gap voltage).
Information:

Shaft motion
loading on shaft
Eccentricity condition
Shaft alignment visualization

Shaft at
running speed

+
+
+
+
++

++
+

Bearing

Clearance Circle

Shaft at rest

Bode & Nyquist

Information:

Slow roll speed & runout vector

Resonance frequency (critical speed, split resonance)

Vectors information (Amplitude & phase, preloads)

Synchronous Amplification Factor

High & heavy spot relationship

Mode shape / deflection shape

Long Time Waveform

You might also like