Fundamentals of Machinery

Condition Monitoring
Vibration Analysis
Fundamentals of Machinery
Condition Monitoring,
Vibration Analysis
 Contents
 Section 1: Introduction to Machinery Condition Monitoring
 Section 2: Vibration Theory and Measurement Transducers
 Section 3: ISO & API Standards
 Section 4: Machinery Fault Diagnosis
 Section 5: Prism4 Configuration
 Section 6: Supplementary Condition Monitoring Techniques
 Section 1:

Introduction to Machinery
Condition Monitoring
 Section 1.1:

Introduction to Condition
Monitoring
Maintenance Philosophy
For many years maintenance was simply based on keeping the plant running.
There was no real planning or thought involved; it was simply the case that if a
machine failed it was repaired or a spare was used. The more later day philosophy
on maintenance is based upon:

 Optimisation of Production

 Optimisation on Plant Availability

 No Compromises to Safety
Maintenance Strategies
There are four main strategies regarding the execution of maintenance activities:

 Opportunity Maintenance

 Planned Preventative Maintenance

 Breakdown Maintenance

 Condition Based Maintenance

Opportunity Maintenance
Adopts the philosophy ‘Fix It When The Opportunity Arises’. Opportunity
Maintenance is carried out, for example, when:

 There is a plant shutdown

 Other equipment is down for maintenance

 A seasonal or weather window is available

Opportunity Maintenance
Advantages Disadvantages
1. Maintenance does not cause additional plant downtime. 1. Machine condition may worsen before the
opportunity arises - causing secondary damage.
Planned Preventative Maintenance
Adopts the philosophy ‘Fix It Before it Fails’. Planned Preventative
Maintenance is carried out with a high level of planning at specific time intervals.

 Based upon generic machinery

reliability data ‘bath tub’ effect

 Time or running hours based e.g.

after 6 months or 10,000 hours

Preventative Maintenance
Advantages Disadvantages
1. Maintenance is planned and carried out a 1. Machines repaired when they may not have
convenient time. a problem.
2. Theoretically fewer catastrophic failures. 2. Repairs can cause more harm than good.
3. Control over storage of spare parts. 3. Unscheduled breakdowns still occur.
4. Not tailored to individual machines.
Breakdown Maintenance
Adopts the philosophy ‘Fix It When It Breaks’.

 No element of planning

 No forewarning of failure

Breakdown Maintenance
Advantages Disadvantages
1. Machines are not over maintained. 1. Unexpected machine downtime - potential loss
of production.
2. No costs associated with preventative or 2. Potential for secondary damage and
condition based maintenance. catastrophic failure - leading to higher repair
costs.
3. Lack of any form of control or planning
Condition Based Maintenance
Adopts the philosophy ‘If It’s Not Broken Don’t Fix It’. Based upon measured
parameters which are sensitive to the development of machinery faults
 Maintenance related to machine condition

 Predictive - forewarning of failure, diagnosis of faults and root causes

Condition Based Maintenance

Advantages Disadvantages
1. Unexpected downtime is reduced. 1. Expense of instrumentation and software,
training personnel and use of specialist contractors.
2. Optimised spares. 2. Does not identify all machine faults e.g. seal
leaks.
3. Causes of failure can be diagnosed.
4. Identifies many common machine faults.
5. Maintenance can be deferred and performed
when opportunity arises.
6. Tailored to individual machines.
 Section 1.2:

Condition Monitoring
Strategy
Condition Monitoring Strategy
A condition monitoring strategy is developed to focus condition monitoring
activities on business and safety critical machinery:

 A Criticality Assessment is undertaken, based upon the

likelihood of failure and consequences of failure, to
define the equipment for inclusion within the condition
monitoring programme.

 Fault Matrices are utilised to identify detectable

machine faults (for each type of critical machinery) and
the parameters that can be monitored which are
sensitive the development of these faults.

 The Condition Monitoring Strategy is then defined

based upon the critical machine list and failure modes
which can be realistically identified through condition
monitoring.
Sample Fault Matrix – Motor Driven Pump (Motor)
Sample Fault Matrix – Motor Driven Pump (Pump)

The fault matrix is used to define the measurements for inclusion in the condition
monitoring database.
Condition Monitoring Techniques
Condition monitoring techniques include:
 Vibration Analysis
 Lube Oil Analysis
 Thermography
 Electrical Motor Phase Current Analysis
 Rogowski Coil Analysis
 Performance Analysis
Condition Monitoring vs. Machinery Protection
Condition Monitoring

Trending of ‘Fault Sensitive’ machine parameters on a periodic basis, to provide

information regarding current and forecasted machine condition. Allows the
onset of fault conditions to be identified so that maintenance activities can be
planned.

This can prevent unscheduled machinery shutdowns.

Machinery Protection

Acquisition of vibration and temperature values from online systems to initiate

machinery shutdown once a pre-set value has been exceeded.
This avoids catastrophic failure.
 Section 2:

Vibration Theory &

Measurement Transducers
 Section 2.1:

Vibration Theory
 Section 2.1:

Vibration Theory
What is Vibration?
Vibration is defined as an oscillating
motion about an equilibrium point e.g.
a mass on a spring.

 The number of times a complete

cycle takes place per second is
called the Frequency (measured in
hertz (Hz)).
 The motion can be of a single
frequency, as with a tuning fork, or
a number of frequencies such as the
motion of a piston in an internal
combustion engine or a gearbox.
Acceleration, Velocity and Displacement
Vibration is described in one of three terms:

 Acceleration – is a vector quantity (i.e. it has magnitude and

direction) which defines the rate of change of velocity. It is measured in
units of m/s2 or more commonly g, where 1 g = 9.81 m/s2. An acceleration
signal can be integrated to give a velocity signal.

 Velocity – is a vector quantity which defines the speed of motion in a

particular direction i.e. the rate of change of displacement. It is normally
measured in units of mm/s (it can be found measured in Imperial units of
‘ips’ (millionths of an inch per second where 1 ips = 25.4 mm/s). A velocity
signal can be integrated to give a displacement signal.

 Displacement – is a vector quantity which defines the change of

position from a rest position. It is measured in units of mm or mm
(microns). It can be found measured in Imperial units of ‘mils’ where 1 mil
= 25.4 microns.
 Simple Harmonic Motion
Where vibration is of a single frequency the motion is sinusoidal and repeats in
an identical pattern over time. This is known as simple harmonic motion.
 Vibration is described in terms
of amplitude (the level or
magnitude of the motion) and
frequency (the number of
repetitions of one full cycle
per second).
 When viewed in the time
domain, the time waveform
exhibits a pure sine wave.
 When viewed in the frequency
domain simple harmonic
motion exhibits a single peak.
Simple Harmonic Motion – Mass on a Spring Example
The oscillating movement of a mass on a spring exhibits Simple Harmonic Motion :
Quantifying The Amplitude of Vibration
The amplitude of vibration describes it severity and can be quantified in several
ways: peak-to-peak level, peak level, average level and RMS level.
Quantifying The Amplitude of Vibration
The Peak-to-Peak value is the maximum excursion (positive to negative motion) of the
time waveform. Useful for measuring direct shaft displacement within a bearing housing
and is often applied to acceleration envelope measurements.
The Peak value (also termed true peak and zero-to-peak) is the absolute value from zero
to the maximum point in the time waveform . Useful for measuring short duration shocks.
The RMS (Root Mean Square) value is directly related to the energy content of the time
waveform and is an indication of the destructive properties of the vibration. Used
predominantly for machine casing (bearing cap) measurements.
The Average value is of limited practical use in describing vibration.

The following is true for purely harmonic

motion (single frequency):

Peak-to-Peak = 2 x Peak
Peak =1.414 x RMS
RMS = 0.707 x Peak
(Average = 0.9 x RMS = 0.637 x Peak)
Non-Sinusoidal (Complex) Waveforms
The mathematical relationships between peak-to peak, peak, average and RMS
levels are inaccurate once the time waveform is no longer sinusoidal.
Time vs Frequency Domain
Once more than one frequency is present within a time waveform it can become
very difficult to analyse. A vibration data collector (Fast Fourier Transform
analyser) will convert the time waveform into individual frequency components
for ease of analysis.
Phase Angle Measurement
Phase measurements are used to assess the relationship between two vibration
signals. It is commonly used in machine balancing and advanced diagnosis of
machinery fitted with proximity probes.

Phase information is taken in the time domain (carrying out a FFT of a time
waveform looses the phase information).

Phase Angle is the timing relationship between two signals of identical

frequencies. Phase is normally measured in degrees and can be either relative
or absolute:

 Relative Phase requires two vibration signals of the same frequency.

 Absolute Phase requires a vibration signal and a synchronous reference pulse
e.g. Keyphasor or optical tachometer.
Relative Phase
Relative Phase is the timing relationship between two signals measured in
degrees. It is measured from a point on one signal to the nearest corresponding
point on another signal.

Rules to follow when measuring

relative phase:

1. Two vibration signals are

required

2. Same frequency

3. Same units (e.g. mm/s, mm)

4. Either signal is chosen as the

reference signal

5. Relative phase is measured

as either lead or lag from 0 to
180 degrees
Absolute Phase
Absolute Phase is the number of degrees of a vibration cycle following the
triggering of a once-per revolution pulse (e.g. keyphasor, optical pickup, strobe
light or magnetic pickup).

Rules to follow when measuring

absolute phase:

1. Two signals are required (one

vibration signal and one reference
signal)

2. Measured from reference signal

thus always a phase lag angle from
0 to 360 degrees

3. 0o location is the point at which

the reference signal triggers

4.Vibration signal to be filtered to a

single frequency which is a integer
multiple of reference signal
Phase Relationships of Acceleration/Velocity/Displacement
It is very important to measure in consistent units when measuring phase. When
a signal is integrated from acceleration to velocity or from velocity to
displacement its phase angle changes. It is also essential that vibration
transducer orientation is taken into account.
 Velocity leads displacement by a phase angle of 90o.
 Acceleration leads velocity by a phase angle of 90o.
 Acceleration leads displacement by a phase angle of 180o.

Period T = 360o
Uses of Phase Angle Measurements
Examples of how phase angle measurements can be used:
 Shaft Balancing – vibration measurements are related to the phase reference
to calculate the placement of balance weights
 Shaft Crack Detection
 Shaft/structural resonance detection
 Shaft mode shapes
 Direction of shaft precession
 Confirming force or couple imbalance
 Confirming misalignment
Shaft Orbits
When two XY vibration signals are added together the resultant signal shows a
two dimensional picture of the vibration motion. This is known as an orbit.
 If the two vibration signals are from casing mounted transducers (e.g.
accelerometers) the orbit reveals the casing motion.
 If the two vibration signals are from proximity probes the orbit reveals the
actual shaft motion within the bearing clearance.
Shaft Orbit Shape
Shaft orbits must be taken from XY vibration transducers mounted orthogonally
(90o apart). They do not need to be true vertical and true horizontal but must be
90o apart in the radial plane.
 Figure 1 shows a typical orbit. As machines are typically more stiff vertically
than horizontally the orbit is elliptical in shape.
 Figure 2 shows a circular orbit. A circular orbit is normally indicative of an
imbalance condition.
 Figure 3 shows a figure of eight orbit. This shape of orbit is characteristic of
misalignment.

1. Typical Orbit 2. Orbit Indicating 3. Orbit Indicating

Imbalance Misalignment
Resonance
When a tuning fork is struck it rings at a single frequency. This frequency is
known as its resonant frequency and is the same every time the fork is struck.
The natural frequency of any system is a function of its stiffness and mass.
 Resonance occurs when the
forcing frequency is equal to
the natural frequency.
 At resonance very little
excitation is required to
produce a large response. The
same excitation above or
below resonance will produce
a greatly reduced response.
 Increasing or decreasing the 1 Stiffness
Natural Frequency Fn  
system’s stiffness or mass will 2 Mass
change the resonant
frequency.
[Note: the above equation is true for an undamped single degree of
freedom system]
Why Machines Vibrate?
All machines have the properties of mass and stiffness and therefore possess the
ability to vibrate.

Perfect engineering would produce machines with no vibration, however, in

reality all machines are built to tolerances and as such any rotating system will
have an element of unbalance. This unbalance force will produce vibration when
the machine rotates.
Rigid vs Flexible Rotors
 A machine is said to have a Rigid
Rotor if the rotating elements’ Rigid Support and Flexible Rotor
natural frequency is above the
running speed of the unit.
 If a machine runs at a speed above
the rotating elements’ natural 1st Critical
frequency it is said to have a
Flexible Rotor. On run up and
shutdown the machine will pass
through resonance (a critical speed).
2nd Critical
 A Critical Speed is a natural
frequency of the rotating element
and its support system including
bearings and lubricating film.
3rd Critical
Critical Speeds
Critical speeds of rotating machinery are speeds which correspond with the
systems resonant frequencies.
 It is important to know a machine’s
critical speeds so that they do not
coincide with normal operating
speeds.
 The response of a rotor at critical
speed will give an indication of the
system’s damping and hence the
condition of the journal bearings.
 The synchronous amplification factor
between operating speed vibration
and resonance should be determined
during commissioning to calculate
machine protection system alarms.
Synchronous Amplification Factor
When a unit is running at a speed above resonance (shaft critical speed) the
maximum allowable vibration should be governed by the level the vibration will
reach when it passes through resonance upon shutdown. The simplest
calculation of SAF is the Peak Ratio Method where the level of vibration at
resonance is divided by the steady state running speed vibration. This is often
used to calculate machinery trip levels in order that shutdown occurs at such a
level of vibration to avoid damage on rundown.

Example:
If the level of running vibration is 100
microns (pk-pk) and the bearing oil
clearance is 200 microns. The unit will
only safely run down if the SAF is
significantly less than 2.
 Section 2.2:

Vibration Transducers
Vibration Transducers
There are three main types of vibration transducer:

 Accelerometer – measures acceleration from machine casing (the signal

can be integrated to measure velocity or displacement).
 Velocity Transducer – measures velocity from machine casing.
 Proximity Probe (Eddy Current Probe) – non-contact transducer which
measures shaft relative vibration.
Transducer Sensitivities
The sensitivity of a transducer describes its electrical output per unit of
mechanical input. The higher the electrical output per unit input, the more
sensitive the transducer.

Sensitivities can be described in metric or imperial units. The nominal sensitivity

of a transducer is normally displayed on its casing. For vibration transducers
sensitivities are normally given in mV/EU (millivolts per engineering unit).

 Accelerometer sensitivities are defined in units of mV/m/s2 (millivolts per

metre per second squared) or more commonly mV/g (where g is gravity).
 Velocity Transducer sensitivities are defined in units of mV/ips (millivolts
per inches per second) or mV/mm/s (millivolts per millimetre per second).
 Proximity Probe (Eddy Current Probe) sensitivities are defined in units of
mV/mil (millivolts per millionth of an inch (1x10-6 inches)) or mV/mm
(millivolts per micron or micrometre (1x10-6 metres).
Sensitivity Conversion Factors
The following are the most common sensitivity conversion factors for Imperial to
metric units. When setting up a condition monitoring database it is good
practice to convert all sensitivities into metric units for ease of analysis:

Sensitivity Conversion
Divide by to Obtain
mV/m/s2 9.81 mV/g
mV/ips 25.4 mV/mm/s
mV/mil 25.4 mV/mm

Unit Conversion:
1 g = 9.81 m/s2 (meters per second squared)
1 ips (inches per second) = 25.4 mm/s (millimeters per second)
1 mil (millionth of an inch) = 25.4 mm (micrometers or microns)
Transducer Frequency Response
 Accelerometer – useable range 1 to
20,000 Hz (dependent upon mounting
arrangement), specialist
accelerometers exist for very low and
very high frequency applications.
 Velocity Transducer – useable range
10 to 1,500 Hz for electromechanical
type and 1 to 2,000 Hz for
piezoelectric type.
 Proximity Probe (Eddy Current
Probe) – stated useable range 0 to
10,000 Hz, typically used in the
frequency range 0 to 2,000 Hz as
higher frequencies can be influenced
by shaft surface imperfections.
Accelerometer
An accelerometer converts acceleration into an electrical output.

Typical Sensitivities are 25 mV/g, 50 mV/g and most commonly 100 mV/g.
Very sensitive accelerometers for low frequency, low amplitude applications
can be as high as 1000 mV/g (1000 V/g).
Accelerometer Mounting
An accelerometer’s frequency response
is highly dependent upon how it is
mounted to the machine surface.
 Stud mounted measurements provide
the best frequency response and
repeatability for machine condition
monitoring. This is critical for rolling
element bearing and gearbox vibration
analysis.
 Magnetic mounted measurements have
reduced frequency response.
 Hand Held measurements provide the
poorest repeatability and frequency
response (should only be used when no
other alternative e.g. very high
machine casing temperature).
Velocity Transducer
A velocity transducer measures the rate of change of displacement and is
traditionally an electromechanical device.
Proximity Probe (Eddy Current Probe)
A proximity probe is a non-contact electromagnetic sensor which
converts displacement (distance) to voltage. The DC component
of the signal measures the average distance from the shaft
whereas the AC component measures the dynamic fluctuation in
displacement i.e. the vibration.

Typical Sensitivities are 3.94 mV/mm (100 mV/mil) and most commonly 7.87
mV/mm (200 mV/mil). [Note: 1 mil = 25.4 mm].
Proximity Probe - Mounting
Pairs of radial probes are orientated 90o apart and referred to as ‘X’ and ‘Y’.
Processing these two signals together produces a shaft orbit.
Proximity Probe – Gap Voltage
The DC component of proximity probe is known as its gap voltage and accurately
measures the distance of the probe tip from the shaft. A probe has a typical
range of 0 to -18 Vdc (~ 2.3 mm). The probes response is linear across a large
proportion of this range.
 Proximity probe gap voltages are
normally set up at – 9 Vdc.
 As a rough guide if the gap voltage is
between -6 Vdc and – 12Vdc it is well
within its linear operating range. If
the gap voltage is not between -3
Vdc to -15 Vdc it is potentially
outside its linear range.
 If a gap voltage is close to zero it is Note: Gap voltages by convention are
short circuited or to close to the always negative. Thus the more
‘positive’ the gap voltage the closer
shaft. If a gap voltage is -18 Vdc it is
you are to the shaft.
open circuit or pointing into space.
Proximity Probe – Run out
Shaft surface imperfections (e.g. scratches, dents, irregular conductivity or
permeability) are indistinguishable from vibration to a proximity probe. This
additional ‘signal’ is known as run out.

 On high speed machines (>2,500 RPM) run out is measured as part of machine
commissioning during slow roll tests (typically 300 to 600 RPM).
 API standards set limits for acceptable levels run out.
 As a guide 6 mm (microns) or 10% of the overall vibration signal is acceptable.
Proximity Probe – Used as a Keyphasor

Shaft Shaft

Probe Probe

Out Out

-18V or -24V Proximitor -18V or -24V Proximitor

-Volts Keyway -Volts

Projection
0 0

(5) (5)

(10) (10)

(15) (15)
(20) (20)

A Keyphasor provides a once-per-revolution pulse used as a reference to

measure absolute phase. [Note: Keyphasor is a Bently Nevada trade name]
Proximity Probe – Used as a Speed Reference
A Keyphasor provides a once
per revolution pulse which
will give an indication of
machine running speed.

To use a proximity probe for

precise speed control requires
a pulse multiple times per
revolution.

For precision speed control a toothed wheel (also known as a Phonic Wheel) is
targeted by a proximity probe. The signal is processed to give a highly accurate
measurement of shaft speed which is updated multiple times per revolution.
Vibration Transducer Comparison - Advantages

Advantages
Accelerometer Velocity Transducer Proximity Probe
1. Surface Mounted. 1. Surface mounted and 1. Direct measurement of
portable. shaft of shaft
2. Small, Portable and motion/position within
Robust. 2. Self-generating no journal bearing.
complex signal
3. Large Dynamic Frequency conditioning. 2. Very sensitive to low
Range. frequencies down to DC.

4. Relatively Inexpensive.

5. Signal can be integrated

to measure velocity or
displacement.
Vibration Transducer Comparison - Disadvantages

Disadvantages
Accelerometer Velocity Transducer Proximity Probe
1. Requires amplifier 1. Bulky. 1. Limited frequency range
electronics. (0 to 10 kHz). Practical
2. Limited Frequency Range range 0 to 2 kHz.
(<1.5 kHz).
2. Permanently mounted
3. Moving parts potentially (not portable) often
wear over time. difficult to replace.

3. Conditioning electronics
required and interface
panel must be housed in
non-hazardous area.
Vibration Transducer Comparison - Applications

Accelerometer
1. Machines with rolling
element bearings.
2. Gearbox Fault Diagnosis.
3. Heavy rigid rotors with light
casing/foundations.
4. Highly utilised with
portable handheld data
collectors.
Applications
Velocity Transducer Proximity Probe
1. Portable transducer for 1. Machines with journal
measurement of low speed bearings.
machines.
2. Machines with lightweight
high speed rotors in heavy
casing/foundations.
3. Measurement of radial
shaft vibration and axial
shaft position.
4. Keyphasor (phase reference
device).

5. Speed reference.
 Section 3:

ISO & API Standards

Alarm Setting Guidelines
American Petroleum Institute (API) and International Standards Organisation
(ISO) have produced guidelines as to acceptable levels of vibration based upon
generic machine types.

These guidelines are used in acceptance testing, e.g. during commissioning, to

ascertain if a new machine is fit for purpose. They are also useful for
reference purposes but it should be noted that alarm settings for condition
monitoring purposes should be set on a machine by machine basis taking into
account historical data.

The rate at which vibration levels and characteristics deteriorate over time is
as important as the magnitude of vibration i.e. a high level of vibration that
remains stable over time may be less cause for concern than a lower level of
vibration which shows a deteriorating trend or changing characteristics.

In routine condition monitoring we are looking for deterioration in vibrations

levels not just absolute values.
ISO-10816-1 Vibration Standard
ISO-10816-1 – Mechanical Vibration – Evaluation of Machinery Vibration by
Measurements on Non-Rotating Parts is the most commonly referenced standard
in routine condition monitoring for the evaluation of overall vibration levels taken
on machine casings.
 The standard provides general guidelines for the severity of overall casing (i.e.
bearing cap) vibration levels based upon Machine Classes I, II, III and IV. These
classes are define by power rating and stiffness of the mounting arrangement
(i.e. rigid or relatively soft mounts).
 The severity of vibration is classified into Evaluation Zones A, B, C and D to
quantify if the level of vibration is acceptable for long term operation of the
machine.
Extract from ISO-10816-1 – Machine Classifications
The machine classifications are as follows:
Class I: Individual parts of engines and machines, integrally connected to the
complete machine in its normal operating condition. (Production electrical
motors up to 15 kW are typical examples of machines in this category).
Class II: Medium sized machines (typically electrical motors with 15 kW to 75 kW
output) without special foundations, rigidly mounted engines or machines (up to
300 kW) on special foundations.
Class III: Large prime-movers and other large machines with rotating masses
mounted on rigid and heavy foundations which are relatively stiff in the direction
of vibration measurements.
Class IV: Large prime-movers and other large machines with rotating masses
mounted on foundations which are relatively soft in the direction of vibration
measurements (for example, turbo generator sets and gas turbines with outputs
greater than 10 MW).
[Note: soft or special foundations refer to anti-vibration mounts]
Extract from ISO-10816-1 - Evaluation Zones
The following typical evaluation zones are defined to permit a qualitative
assessment of the vibration on a given machine and to provide guidelines on
possible actions:
Zone A: The vibration of newly commissioned machines would fall within this
zone.
Zone B: Machines with vibration within this zone are normally considered
acceptable for unrestricted long-term operation.
Zone C: Machines with vibration within this zone are normally considered
unsatisfactory for long-term continuous operation. Generally, the machine may
be operated for a limited period in this condition until a suitable opportunity
arises for remedial action.
Zone D: Vibration values within this zone are normally considered to be of
sufficient severity to cause damage to the machine.
Extract from ISO-10816-1 – Typical Zone Boundary Limits
ISO-10816-1 - Typical Zone Boundary Limits
Vibration Velocity
Class I Class II Class III Class IV
mm/s (RMS)
0.28
0.45 A
A
0.71 A
A
1.12
B
1.8
B
2.8
C B
4.5
C B
7.1
C
11.2
C
18 D
D
28 D
D
45
API Standards
The American Petroleum Institute (API) has produced numerous standards to
satisfy the specific needs of the petroleum, chemical and gas industries. These
standards closely specify the detailed design, inspection and testing of generic
machine types, for example:
 API 610 – Centrifugal Pumps
 API 611 – General Purpose Steam Turbines
 API 612 – Special Purpose Steam Turbines
 API 613 – Special Purpose Gear Units
 API 616 – Gas Turbines
 API 617 – Axial and Centrifugal and Expander Compressors
 API 618 – Reciprocating Compressors
 API 619 – Rotary-Type Positive Displacement Compressors
 API 670 - Vibration, Axial Position and Bearing Temperature Monitoring
Systems
 API 674 – Positive Displacement Pumps – Reciprocating
 API 676 – Positive Displacement Pumps – Rotary
 API 677 – General Purpose Gear Units
 API 681 – Liquid Ring Pumps
API Standards – Guidelines for Vibration
API 670 ‘Vibration, Axial Position and Bearing Temperature Monitoring Systems’
specifies requirements for the supply, installation and calibration of radial shaft
vibration and axial-position transducers and bearing temperature sensors for
online machinery protection systems. The various machine specific standards give
guidance as to acceptable levels of vibration.
The following equation is common to several of the API standards and defines the
maximum allowable level of relative shaft vibration (i.e. vibration measured using
a proximity probe):

e.g. Gearbox vibration shall not exceed 50 micrometers (peak-to-peak) or that

defined by the above equation, whichever is less [API 613].
e.g. Centrifugal Compressor vibration shall not exceed 25 micrometers (peak-to-
peak) or that defined by the above equation, whichever is less [API 617].
 Section 4:

Machinery Fault Diagnosis

 Section 4.1:

Imbalance
Imbalance
Imbalance (also known as unbalance) occurs when there is a deviation between
the geometric centre of a rotor and its centre of mass. Or put
more simply – when there is a heavy spot on the shaft.

 Imbalance is one of the most common causes of excessive vibration in

rotating machinery.
 It is always characterised as radial vibration at the 1X running speed
(rotational speed) of the shaft.
 Dependent upon the relative support stiffness, radial vibration may be
more prominent in the horizontal or vertical axis.
 It is often misdiagnosed as many faults exhibit 1X running speed
characteristics - other vibration symptoms should be investigated before
balancing is attempted.
Static (Force) Imbalance

 Is characterised by a dominant
1X running speed component.

 Is measured in the radial

direction and is in-phase.

 Is corrected by one balance

weight in one plane at Rotor
centre of gravity (single plane
balance).
Couple Imbalance

 Is again characterised by a
dominant 1X running speed
component.

 Will be 180o out of phase on

same shaft and can exhibit both
high axial and radial vibration.

 Is corrected by applying balance

weights in more than one plane
(multi-plane balance).
Dynamic Imbalance
A rotor with static imbalance can be diagnosed when the machine is not running.
This is carried out by placing the rotor in frictionless bearings. If the rotor has a
heavy spot it will rotate within the bearings until the heavy spot is at the
bottom.

Conversely a rotor with pure coupled imbalanced will not rotate when placed in
frictionless bearings and will only manifest itself when the machine is running.

In practice a rotor is likely to have a combination of static and couple imbalance

which is collectively referred to as Dynamic Imbalance.
Overhung Rotor Imbalance

 Is again characterised by a
dominant 1X running speed
component.

 Will show a high axial 1X

running speed component.

 Axial vibration readings tend

to be in-phase whereas radial
readings may be unsteady.
Vertical Rotor Imbalance

 Characterised by a dominant 1X running

speed radial vibration.

 Will show highest vibration at the motor

non-drive end irrespective of source.

Note: Vertically mounted pumps will often show large 1X running speed vibration at the
motor non-drive end for a number of faults (e.g. pump bush wear, flow turbulence). Try to
isolate the problem by uncoupling the motor/pump. Carry out measurements on motor
whilst uncoupled. If 1X vibration is still relatively high the motor is at fault if not it is the
pump.
Fan and Overhung Imbalance
Imbalance can be very common in fans but should not be mistaken for belt drive
problems which can also reveal 1X running speed vibration.

 Fan imbalance is characterised by a

dominant 1X running speed component
in the radial direction.
 Overhung fan imbalance is
characterised by a dominant 1X
running speed component in the axial
direction.
 Ensure the fan blades are clean and
show no signs of damage. Often the
cause of imbalance can be a build up
of deposits on the fan blades.
 Fan balancing can often be carried out
in situ. The impeller can normally be
balanced by applying a single weight.
Imbalance Severity
Manufacturer’s and API Standards may impose acceptable levels of
imbalance for specific machine types. The following is a rough guideline to the
severity of imbalance relating to the 1X running speed vibration component of
vibration (for machines running between 1800 and 3600 RPM).

Severity of Imbalance Guidelines for Machines Running at 1800 to 3600 RPM

1X Vibration Level
VdB (US) re 10E-8 m/s Equivalent mm/s Diagnosis Repair Priority
(rms) (rms)
No
Recommendatio
<108 0 to 2.5 Slight Imbalance n
Moderate
Very high speed machines will have lower tolerance levels as the forces
108 to 114 2.5 to 5.0 Imbalance Desirable
generated by imbalance, increase with machine speed.
Serious
114 to 124 5.0 to 15.8 Imbalance Important
Effects of Imbalance
 All machines inherently have some form of residual imbalance.
 Some slight imbalance will have little effect on the operating lifespan of a
machine or its components.
 An unacceptable level of imbalance can severely reduce the lifespan of
bearings and seals.
 A high level of imbalance can have catastrophic effects for large machinery
with flexible rotors (running above shaft critical speeds).
 Section 4.2:

Misalignment
Misalignment
Misalignment occurs when the axis's of coupled machine components are not
collinear.

 Misalignment can be parallel (offset) or angular which will be diagnosed by

whether the vibration characteristics are dominant in the radial or axial
planes respectively. It can, however, be a combination of both.

 Misalignment introduces a static preload force into the coupled shafts.

 It is typically characterised by 1X, 2X and 3X running speed vibration

components.

 Misalignment characteristics may also indicate coupling problems.

 Misalignment will be effected by thermal and dynamic growth and may

manifest itself more prominently once the machine reaches its steady state
operating condition.
Parallel (Offset) Misalignment

 Parallel misalignment is characterised by dominant 1X and 2X and to a lesser

extent 3X running speed components of vibration in the radial direction.
 Approaches 180o out of phase across the coupling in the radial direction.
Angular Misalignment
 Cocked Bearing
A cocked bearing is a form of misalignment which can generate high axial
vibration, characterised by high 1X, 2X and 3X running speed characteristics.

 Will show 180o phase shifts (in

the axial plane) top/bottom
and/or left/right on the same
bearing housing.
 Realignment or balancing will
not cure the problem - the
bearing will need to be removed
and reinstalled correctly.
 Section 4.3:

Mechanical Looseness and

Rotor Rub
Mechanical Looseness
Mechanical looseness is a common cause of high vibration and is by far one of the
easiest problems to check.

 Mechanical looseness should ALWAYS be checked before more intrusive

maintenance activities are considered.

 Mechanical looseness often reveals high 1X vibration.

 Mechanical looseness can exhibit high 1X, 2X and 3X times running speed
components of vibration which can be misinterpreted as misalignment.

 A raised noise floor or ½ x multiples of running speed harmonics of

vibration can sometimes be associated with looseness.

 The Technical Associates of Charlotte define mechanical looseness in three

categories; A: Structural, B: Fasteners and C: Component Fits.
Mechanical Looseness – Type A: Structural
“Type A is caused by Structural looseness/weakness of machine feet, base plate or
foundation; also by deteriorated grouting, loose hold-down bolts at the base; and
distortion of the frame or base (i.e., soft foot).

Phase analysis may reveal approximately 90° - 180° phase difference between
vertical measurements on bolt, machine foot, base plate, or base itself.” -
Technical Associates of Charlotte, Inc.
Mechanical Looseness – Type B: Fasteners
“Type B is generally caused by loose pillowblock bolts, cracks in frame structure or
in bearing pedestal.” - Technical Associates of Charlotte, Inc.
Mechanical Looseness – Type C: Component Fits
“Type C is normally generated by improper fit between component parts. Causes a
truncation of time waveform and a raised noise floor in the spectrum. Type C is often
caused by a bearing liner loose in its cap, a bearing loose turning on its shaft, excessive
clearance in either a sleeve or rolling element bearing, or a loose impeller on a shaft, etc.

Type C Phase is often unstable and may vary widely from one measurement to next,
particularly if rotor shifts position on shaft from one startup to next. Mechanical Looseness
is often highly directional and may cause noticeably different readings comparing levels at
30° increments in radial direction all the way around one bearing housing. Also, note that
looseness will often cause subharmonic multiples at exactly 1/2 or 1/3 RPM (.5X, 1.5X, 2.5X,
etc.). .” - Technical Associates of Charlotte, Inc.
Rotor Rub
“Rotor Rub produces similar spectra to Mechanical Looseness when rotating parts contact
stationary components. Rub may be either partial or throughout the entire shaft revolution.
Usually generates a series of frequencies, often exciting one or more resonances. Often
excites integer fraction subharmonics of running speed (1/2, 1/3, 1/4, 1/5,...1/n),
depending on location of rotor natural frequencies. Rotor rub can excite many high
frequencies (similar to wide-band noise when chalk is drug along a blackboard).

It can be very serious and of short duration if caused by shaft contacting bearing babbitt. A
full annular rub throughout an entire shaft revolution can induce "reverse precession" with
the rotor whirling at critical speed in a direction opposite shaft rotation (inherently unstable
which can lead to catastrophic failure).
- Technical Associates of Charlotte, Inc.
 Section 4.4:

Rolling Element Bearings

Rolling Element (Frictionless) Bearing Faults
Rolling element bearing faults are characterised by discrete (non-synchronous)
frequency vibration components which are dependent upon the bearing
construction.

Prism4 includes a database of the most common bearing tags and can compute
these frequencies through it’s Frequency Analysis Module (FAM).
Deep Groove Ball Bearing Components

Seal Rolling elements Inner ring

Outer ring Cage Seal

Rolling Element Bearing Stages of Failure
The very early stages of bearing faults are detected at very high frequencies using
spike energy, shock pulse or HFD (High Frequency Detection) techniques.
Rolling Element Bearing Stages of Failure
The development of bearing faults can be tracked using acceleration enveloping
techniques.

Final stages of bearing faults will become evident in the vibration velocity spectra.
Acceleration Enveloping – How it Works?
The principles behind acceleration enveloping:

 All low frequency vibration components which are attributed to mechanical

faults such as imbalance, misalignment, mechanical looseness etc. are filtered
out.
 Bearing fault frequencies are non-synchronous components of vibration i.e.
they are not exactly 1X, 2X etc. running speed multiples. They may however
be very close to a multiple of running speed (e.g. 4.9X running speed) and thus
difficult to separate out in an unfiltered signal. Enveloping inherently
provides this filtering.
 Very sensitive to the onset of bearing faults. By the time a fault is visible in a
vibration spectra, the bearing is already significantly worn.
Acceleration Enveloping – How it Works?
Time waveform illustrating dominant 1X running speed vibration with frequent
impacts (the transient superimposed on the waveform is like the ringing of a bell):
Acceleration Enveloping – How it Works?
Firstly the low frequency components of vibration are removed using a high pass
filter:
Acceleration Enveloping – How it Works?
The resulting filtered signal contains only the high frequencies with the lower
frequencies removed:
Acceleration Enveloping – How it Works?
The time waveform would now only show the bearing transient impacts (note the
1X running speed waveform has been filtered out):
Acceleration Enveloping – How it Works?
The signal is now demodulated so that the high frequencies are flipped over into
the baseband of the frequency scale:
Acceleration Enveloping – How it Works?
The negative components of the signal waveform are flipped over to the positive
portion of the signal:
Acceleration Enveloping – How it Works?
A low pass filter is now applied to remove any unwanted signals from other sources
of modulation:
Acceleration Enveloping – How it Works?
The high frequency components are now removed. This is the enveloped signal:
Acceleration Enveloping - Analysis
The analysis on the remaining spectra is based upon trending of the frequency
peaks against the noise floor:
Acceleration Enveloping - Considerations
Acceleration enveloping has become the favoured technique for assessing the
onset and deterioration of rolling element bearing faults. It is both highly
sensitive and easy to trend, however, great care should be taken in interpreting
the results as:
 The measurement location/technique needs to be highly repeatable. It is
advised to stud mount the accelerometer. Swapping between stud and
magnetic mounted or handheld measurements will produce highly spurious
results/trends.
 The measured characteristics are analysed as deterioration relative to the
signature of a new bearing. Starting measurements well into a bearings
lifespan provides a less reliable reference point.
 Acceleration enveloping can give an indication of insufficient lubrication which
could be misinterpreted as poor bearing condition.
Acceleration Enveloping - Guidelines

 There are basic guidelines to unacceptable levels of acceleration enveloping

(gE); based upon machine speed and shaft diameter.
 These should be used with caution as it is the relationship between the
discrete bearing fault frequencies and the noise floor (often termed ‘carpet
level’) which provides the best indication of bearing condition.
 This relationship will differ dependent on the stage of bearing wear which is
why it is always important to gather baseline data when a new bearing is
fitted.
Acceleration Enveloping Guidance Levels:
 Section 4.5:

Journal Bearings
Journal Bearing Faults
Journal Bearings also known as Sleeve, Plain, Fluid Film or White Metal Bearings
show very different fault characteristics to rolling element bearings.
The most common Journal bearing faults are:
 Wear and Clearance Problems
Due to improved tolerances in Journal Bearing
design the following faults are nowadays less
common but can still cause catastrophic effects:
 Oil Whirl Instability
 Oil Whip Instability

NOTES:
1.Wear, Clearance and Oil Whirl problems can be detected in steady state vibration
spectra, whereas Oil Whip is more likely to occur during machine start-up, which requires
more advanced transient analysis.
2. Acceleration Envelope and HFD/Spike Energy/Shock Pulse measurement techniques are
NOT applicable to Journal Bearings.
Journal Bearing Wear and Clearance Problems
Journal bearing wear and clearance problems show very similar symptoms to
mechanical looseness and are identified by strong running speed harmonics:

Wiped journal bearings will often show high vertical vibration compared with the
horizontal measurement.
Oil Whirl Instability
Oil whirl can be characterised by vibration frequencies just below, but never equal
to, half times running speed:

Changes in lube oil viscosity, lube oil pressure and external preloads can all
effect oil whirl.
Oil Whip Instability
“Oil Whip may occur if machine operated at or above 2X rotor critical frequency.
When rotor brought up to twice critical speed, whirl will be very close to rotor
critical and may cause excessive vibration that oil film may no longer be capable
of supporting. Whirl speed will actually "lock onto" rotor critical and this peak will
not pass through it even if machine is brought to higher and higher speeds.
Produces a lateral forward processional subharmonic vibration at rotor critical
frequency. Inherently unstable which can lead to catastrophic failure.” – Technical
Associates of Charlotte, Inc.
Journal Bearings
Journal bearings are often fitted to large machinery with online protection
systems.

Alert and trip setting will be set for vibration, axial displacement and bearing
temperatures.

When a journal bearing wipes both the vibration and temperature will
increase instantaneously, most likely tripping the machine.
 Section 4.6:

Gear Analysis
Gear Faults
When analysing gears (e.g. helical, spur, worm, bevel, epicyclic) gear mesh
frequencies can be calculated from:
 Input & Output Shaft Speed
 Number of Teeth on Pinion & Wheel
(Note: For a two stage gearbox the shaft speed
and teeth of the intermediate gears would also
be required).
When collecting vibration data on gearboxes,
where possible, time waveform data should
be captured along with the FFT Spectra, in
at least one axis.
Theoretically Acceleration Envelope techniques can be applied to gear analysis,
they should be utilised with caution however, as the transmission path between
the meshing gears and vibration measurement location is often indirect
(especially on large gearboxes).
Gear Mesh Frequency
Gear mesh frequency is defined as the number of teeth on a gear multiplied by
its shaft rotating frequency:
Gear mesh frequency = (Low Speed Shaft RPM / 60) x number of teeth on wheel
Or (High Speed Shaft RPM / 60) x number of teeth on pinion
[NOTE: The RPM has been divided by 60 to convert it into Hertz]
When analysing gearboxes it is essential that data is captured using a suitable
frequency range to capture up to 3.25 x gear mesh frequency in order to assess
gear misalignment issues.
We would normally capture a vibration velocity measurement with a frequency
range up to 5 kHz to capture gear natural frequencies and an acceleration
reading up to 20 kHz to capture harmonics of gear mesh frequency.
Gear Mesh Characteristics
A typical gearbox vibration spectrum will show low speed and high speed shaft
running speed components accompanied by low amplitude gear mesh frequency
with shaft running speed sidebands.
 The highest level of vibration will be
either radial or axial dependent upon
the type of gear e.g. spur or helical
gear.
 The time waveform should show
evenly spaced impulses of similar
amplitude for a healthy gearbox. A
pulse is produced as each tooth
meshes.
 The time waveform is often easier to
analyse than the vibration spectra in
the diagnosis of gear faults.
Gear Tooth Wear
When gear teeth start to wear the sidebands of gear mesh frequency become more
pronounced – the amplitude and number of sidebands will increase. Gear natural
frequencies will also be excited.
 The gear mesh frequency sidebands
will correspond to the gear with the
wear e.g. if the sidebands are equal
to the high speed shaft running speed
it will be the pinion gear teeth which
exhibit wear.
 The gear natural frequencies are
lower than gear mesh frequency and
will also exhibit sidebands relating to
the bad gear.
Gear Tooth Load
Gear mesh frequencies can be very sensitive to load. High gear mesh frequencies
do not necessarily indicate a problem provided sideband frequencies remain at low
amplitudes and gear natural frequencies are not excited.
 In order to trend gear mesh activity,
vibration measurements should be
recorded with the machine operating
at the same load each survey
(wherever possible).
 Machine load should be recorded
each survey as a manual entry
reading.
Gear Eccentricity and Backlash
Relatively high sidebands around gear mesh frequency can indicate eccentricity,
backlash or non-parallel shafts. The bad gear will be indicated by the spacing of
the sideband frequencies.
 Eccentricity will normally show a high
1X running speed component of
vibration.
 Improper backlash often excites gear
mesh harmonics and gear natural
frequencies.
 Gear mesh frequency amplitudes will
often reduce with increasing load if
backlash is the problem.
Gear Tooth Misalignment
Gear tooth misalignment is diagnosed in a similar manner to angular or parallel
misalignment except it is 1X, 2X and 3X gear mesh frequency which reveals the
symptoms.
 In order to assess for gear tooth
misalignment vibration
measurements must be taken with a
frequency range > 3.25 x gear mesh
frequency.
 Gear tooth misalignment will cause
uneven tooth wear.

NOTE: A loose fit journal bearing can also exhibit high 1X, 2X and 3X times gear
mesh frequency vibration.
Gear Cracked or Broken Tooth
A cracked or broken tooth is best diagnosed in the time waveform which will show
a large impulse every time the problem tooth tries to mesh with the teeth on the
mating gear.
 The frequency spectra will reveal
gear natural frequencies.
 The time waveform will reveal high
amplitude 1X running speed
component of the problem gear.
These spikes will reveal themselves in
the time waveform up to 10 to 20
times higher than in the frequency
spectrum.
 In the example opposite (gear with 12
teeth) the time waveform reveals a
very large impulse for the
cracked/broken tooth.
Gear Hunting Tooth
Hunting tooth problems occur due to faults on both the gear and pinion created
during manufacture or improper handling. A hunting tooth problem can often be
overlooked as it is revealed at very low frequencies, often less than 10 Hz.
 A gearbox with a hunting tooth
problem may emit a ‘growling’ sound.
 The effect is at its worst when the
faulty gear and pinion teeth try to
mesh at the same time. This may
only occur once every 10 to 20
revolutions.
 The number of teeth on a gear are
often a prime number to avoid
hunting tooth problems i.e. two
imperfect teeth will not repeatedly
mesh.
 Section 4.7:

Hydraulic and Aerodynamic

Faults
Hydraulic and Aerodynamic Faults
Hydraulic and Aerodynamic forces in pumps, fans and compressors will produce
vibration at Blade Pass or Vane Pass frequency. Blade pass frequency can be
calculated by multiplying the number of blades or vanes by the shaft rotational
speed:

Blade Pass Frequency = Number of Blades x (RPM / 60) Hz

Vane Pass Frequency = Number of Vanes x (RPM / 60) Hz

Blade pass frequency is an inherent characteristic of the machine which will vary
with process conditions and does not normally cause a problem.
Blade Vane Faults
Large blade pass frequencies are generated if the gap between the rotating vanes
and stationary diffusers is not equal all the way round i.e. it is eccentric.

 High blade pass frequencies, with 1X

running speed sidebands, can be
generated if an impeller wear ring seizes
on the shaft or if welds which fasten
impeller vanes fail.
 High blade pass frequencies can also be
generated by flow disturbance or the
eccentric positioning of the pump or fan
rotor within its housing.

NOTE: Blade pass frequency fluctuates significantly with process condition so a

deteriorating trend must be established before intrusive maintenance is
considered. Closely monitor for increases in the blade pass frequency sidebands.
Flow Turbulence
Flow turbulence is caused by variations in the pressure or velocity of air passing
through a fan or blower.

 Flow turbulence will normally exhibit

sub-synchronous (below 1X running
speed) random noise in the vibration
spectra.
 Excessive turbulence can also exhibit
broadband high frequency vibration i.e.
an increase in noise floor above the
blade pass frequency.
Cavitation
Cavitation is normally caused by insufficient suction pressure (starvation) or inlet
flow and normally generates random high frequency broadband vibration which
appears as a raised noise floor above the blade pass frequency.

 Cavitation can be audible – sounding like

gravel is passing through the pump.
 Cavitation can cause erosion of pump
internals and impellers if left
uncorrected.
 Cavitation may vary from one survey to
the next but can often be overcome by
increasing the pump suction pressure.
 Section 4.8:

Electrical Faults
Electrical Faults
Mechanical faults such as imbalance, misalignment and bearing problems are
typically more common in electric motors than electrical faults.

Electrical fault frequencies may be present in a vibration spectra at low levels,

which could be a characteristic of the machine and not likely to cause long term
detrimental effects.

One of the simplest ways to identify whether a vibration component is mechanical

or electrical in origin is to shutdown the power to the machine whilst recording
real time vibration. If the fault frequency immediately disappears when the
power is switched of the problem is electrical but if the faults frequency reduces
with running speed the fault is mechanical in origin.
AC Induction Motors
3-Phase AC Induction Motors are the most common motors utilised in industrial
applications due to their relatively high efficiencies.
 A motor with a 60 Hz line frequency and 2 stator poles will run at a speed
of 3600 RPM (or 3000 RPM with a 50 Hz line frequency). A motor with a
60 Hz line frequency and 4 stator poles will run at a speed of 1800 RPM
(or 1500 RPM with a 50 Hz line frequency).

 In an induction motor the motor speed is always slightly less than

synchronous speed. The difference between the actual speed and the
synchronous speed is known as Slip. The difference between the
running speed frequency and synchronous speed frequency is known as
the Slip Frequency.

 The greater the slip, the greater the induced current in the rotor bars
and the greater the output torque. This is why the actual speed of an
induction motor will vary slightly with load.
AC Induction Motors – Stator Eccentricity
Stator problems generate high 2X line frequency components of vibration i.e. 100
Hz or 120 Hz dependent upon whether the line frequency is 50 Hz or 60 Hz
respectively. Stator eccentricity produces a uneven stationary air gap between
the rotor and stator which produces very directional vibration.
 Differential air gap should not exceed 5%
for induction motors and 10% for
synchronous motors.
 Soft foot and warped bases can produce
eccentric stators.
 Shorted stator windings can produce
thermally-induced vibration which can
significantly increase with operating time
causing stator distortion and air gap
problems.
NOTE: Electric motors will have a low level of 2X line frequency vibration as a
normal characteristic.
AC Induction Motors – Eccentric Rotor
An eccentric rotor will produce a variable air gap between the rotor and stator
producing pulsating vibration. Eccentric rotors produce 2X line frequency
components with pole passing frequency sidebands.
 Pole passing frequency = slip frequency X
numbers of poles.
 Not to be confused with soft foot or
misalignment which can produce variable
air gaps due to distortion (mechanical
problem not electrical).
 Zoom analysis may be required to
separate 2X line frequency from 2X
running speed harmonics.
 On high voltage motors; Motor Phase
Current Analysis can be used to assess
rotor eccentricity.
AC Induction Motors – Rotor Bow
Uneven heating of a rotor due to unbalanced rotor bar current distribution can
cause a rotor to warp or bow. Rotor bow can be misdiagnosed as mechanical
imbalance as it has similar 1X running speed characteristics.
 Rotor bow can be distinguished from
imbalance as it will worsen when the
motor is hot and the symptoms will
subside when the motor cools down.
 If the local heating effect is very severe
it can cause the offending rotor bar to
melt and lodge within the air gap.
AC Induction Motors – Broken or Cracked Rotor Bars
Broken/Cracked rotor bars or shorting rings, bad joints between shorting rings and
rotor bars or shorted rotor laminations will produce high levels of 1X running speed
vibration harmonics with pole pass frequency sidebands.
 High resolution measurements are
required typically using a 3200 line FFT
spectra.
 Running speed harmonics may be notable
to 5X running speed and above.
 On high voltage motors; Motor Phase
Current Analysis is often employed to
assess for broken rotor bars.
AC Induction Motors – Loose Rotor Bars
Loose rotor bars will exhibit a peak at rotor bar passing frequency (the number of
rotor bars times the motor RPM) with 2X line frequency sidebands.

 A relatively high frequency measurement

is required to detect loose rotor bars as
the rotor passing frequency is often over
2000 Hz.
 Even if the number of rotor bars is
unknown a high frequency vibration
component exhibiting 2X line frequency
sidebands is most likely caused by loose
rotors.
AC Induction Motors – Phasing Problems
Phasing problems due to loose or broken connectors can result is excessive 2X line
frequency vibration with 1/3 line frequency sidebands.

 Very high 2X line frequency vibration

levels in excess of 25 mm/s can result if
the problem is left uncorrected.
 The problem is accentuated if the
defective connector makes intermittent
contact.
 The loose or broken connector must be
repaired to avoid catastrophic failure.
 Section 4.9:

Belt Drive Faults

Belt Drive Faults
Belt drives are an inexpensive means of power transmission. They can however be
prone to a number of faults including:

 Belt Wear
 Misaligned Sheaves (Pulleys)
 Eccentric Sheaves (Pulleys)
 Belt Resonance
Belt Drive Equations
The following equations are useful in determining operating speeds and fault
frequencies for belt drives:

Driving _ RPM  Driving _ Sheave _ Diameter

Driven _ RPM 
Driven _ Sheave _ Diameter

PI  Sheave _ RPM  Sheave _ Diameter

Belt _ Frequency 
Belt _ Length
Ti min g _ Belt _ Frequency  Belt _ Frequency  Number _ of _ Belt _ Teeth
or
Ti min g _ Belt _ Frequency  Sheave _ RPM  Number _ of _ Teeth _ on _ Sheave

Where PI = 3.1416
Worn, Loose or Mismatched Belts
Belt frequencies are below both the driving and driven units running speeds. A
worn, loose or mismatched belt will exhibit up to 3X or 4X harmonics of belt
frequency.
 A high amplitude of 2X belt frequency is
normally present.
 Amplitudes are unsteady and will
sometimes fluctuate between the driving
and driven unit running speed.
 Timing belts will exhibit high amplitudes
of timing belt frequency.
Sheave (Pulley) Misalignment
Misaligned sheaves will produce high vibration at 1X RPM predominantly in the
axial direction. Often with pulley misalignment, the highest axial vibration on the
motor will be at fan RPM, or vice versa.
 A high amplitude of 1X belt frequency
will be present in the axial direction.
 Harmonics of belt frequency may
sometimes be present in the axial
direction.
Eccentric Sheaves (Pulleys)
Eccentric sheaves will generate high 1X running speed vibration, especially in the
axis parallel to the direction of the belts.

 This condition is very common and can

be misdiagnosed for imbalance.
 The 1X running speed of an eccentric
sheave will be evident on both the
driving and driven unit bearings.
 Pulley eccentricity can be confirmed by
phase analysis which should reveal
horizontal and vertical phase differences
close to either 0o or 180o.
Belt Resonance
Belt resonance can reach high amplitudes if the belt natural frequency coincides
with either the driving or driven unit running speed.

 This condition can be checked by

tensioning and then releasing the belt
whilst taking vibration readings on the
pulleys or bearings.
 The natural frequency of the belt can be
altered, by either changing the belt
length or tension, to correct this
problem.
 Section 4.10:

Machinery Fault Diagnosis

Guidelines
Machinery Fault Diagnosis Guidelines
When carrying out machinery fault diagnosis there is often more than one type of
fault that can be associated with the vibration characteristics measured. The
following guidelines should be taken into account before making
recommendations involving intrusive maintenance:
 Vibration levels change with load and process conditions. If a step increase in
1X running speed vibration is evident look for evidence of changes in process
conditions. Process parameters should be recorded wherever possible (as
manual entry readings) for comparison with historical data.
 Pump 1X running speed and blade pass frequency vibration can often fluctuate
spuriously due to altering process conditions.
 If vibration levels show a marked increase or deteriorating trend; survey more
often. By surveying weekly instead of monthly the rate of deterioration can be
more easily established. If the vibration levels appear stable revert to
monthly monitoring.
 Always check for mechanical looseness if 1X, 2X or 3X vibration is present as it
is one of the most inexpensive and easiest maintenance actions to carry out.
Machinery Fault Diagnosis Guidelines Continued
 High acceleration envelope levels on rolling element bearings can be associated
with lack of lubrication. To check this, grease the bearings and then repeat
vibration measurements. The acceleration envelope readings should notably
reduce. Repeat the vibration measurements after the machine has been
allowed to run for a further 24 hours. If the acceleration envelope readings
remain stable, lack of lubrication is confirmed. If the acceleration envelope
readings start to increase closely monitor for bearing fault frequencies.
 Ensure that the vibration spectra reveals good quality data. A ski-slope in the
vibration spectra is an indication of a poorly taken measurement and should be
repeated:
Machinery Fault Diagnosis Guidelines Continued
 Ensure measurements are always taken at the same locations. Stud and mark-
up the measurement locations wherever possible to ensure the best possible
repeatability of measurements.
 If a stud should fall off a measurement location immediately replace it.
Acceleration envelope measurements are highly dependent upon a repeatable
mounting arrangement. If the measurement is changed from stud to magnetic
mounting the envelope levels will be inconsistent and the baseline/trend will be
lost.
 Look for changes in vibration characteristics i.e. frequency component changes
as well as changes in overall vibration levels. A fault may not manifest itself as
an increase in overall vibration levels.
 Section 6:

Supplementary Condition
Monitoring Techniques
Supplementary Condition Monitoring Techniques
Vibration analysis is a very powerful technique in the diagnosis of many common
rotating machinery problems. Additional condition monitoring techniques can be
employed in support of vibration analysis and to provide supplementary
information regarding both machinery and plant condition. Such techniques
include:

 Lube Oil Analysis

 Thermography

 Watch Keeping

 Motor Phase Current Analysis

 Rogowski Coil Analysis

 Performance Monitoring
Lube Oil Analysis
Lube oil analysis serves two main purposes:

 To assess oil quality to ensure the oil is fit for further

use i.e. meets the lubricating requirements for the
machine
 To assess machine condition by
examining the oil for signs of
mechanical wear
Lube Oil Sampling Guidelines
Care should be taken when collecting lube oil samples to ensure
the oil sample is representative of the oil circulating in the
machine and to ensure that the sample is not contaminated. As
a guide:
 Oil samples should only be taken from running machines. Cold samples taken
from non-running machines will allow any particles contained within the oil to
separate and settle. Similarly, when taking samples from piping systems ensure
the sample is taken from a location where oil is circulating e.g. a bend where
flow is more turbulent as opposed to straight pipe where flow is laminar.
 The sample must be taken from the same location each survey. Note: placing a
sampling tube too far into oil reservoir is likely to collect the debris at the
bottom of the tank which must be avoided.
 Samples must be taken upstream of any filters.
 Always use fresh tubing and bottles for each sample and ensure they are
properly stored to avoid contamination.
 Reference oil samples should be taken whenever a new batch of oil is utilised.
Lube Oil Analysis – Laboratory Analysis
Lube oil analysis is carried out on a periodic basis to compare the chemical and
elemental properties of a used oil to a baseline of unused oil. Over time a trend is
built up to determine the rate of deterioration of oil quality or abnormalities
which could be indicative of mechanical wear.
Standard tests include:

 Viscosity – too low a viscosity reduces oil film strength, weakening its ability to
prevent metal-to-metal contact.
 Spectrographic Analysis – measures the concentration (normally in parts per
million (ppm)) of elements (e.g. lead, copper, sodium etc.) entrained in the oil
to determine wear metals, contaminants and additives.
 Total Acid Number (TAN) – is used to measure the acidic content of the oil.
 Total Base Number (TBN) – indicates the ability of an oil to neutralise acid.
TBN is an important test for diesel engines. A low TBN can indicate overdue oil
changes and overheating.
Lube Oil Analysis – Laboratory Analysis Continued
 Water Content – is a standard test used to assess the percentage of water
within the oil sample. Water content should not normally exceed 0.1%.
 PQ Index (Particle Quotient) – gives an indication of ferrous wear debris in the
oil sample. The PQ Index can be easily trended over time and is normally
carried out as part of a standard analysis. A high PQ Index can indicate that
wear is present but ferrographic analysis is required to identify the type of
wear.
 Ferrographic Analysis (also known as Wear Debris Analysis) – is a relatively
expensive test in comparison to standard analysis. It is used to assess the size,
shape and number of wear particles suspended in the oil sample. The size and
shape of the wear particles can be associated to specific wear modes indicative
of mechanical faults.
Lube Oil Analysis – Example Laboratory Report Part 1
Lube Oil Analysis – Example Laboratory Report Part 2
Lube Oil Analysis – Example Laboratory Report Part 3
Thermography
Thermography is a highly utilised technique in the condition monitoring of
electrical switch gear but can also be used as an indicator of mechanical wear. It
works on the principle that temperature changes occur as the condition of
components alter e.g. electrical arcing and bearing wear.
Thermography measures infrared radiation emitted from different materials to
allow the remote (non-contact) measurement of temperature and temperature
differences.
Special viewing panels are often fitted to electrical switch gear enclosures to allow
thermographic survey; as opening the enclosure would let the heat escape.
Thermography - Applications
Watch Keeping
Watch keeping is normally carried on a daily basis to record operating and process
parameters around the plant. This is an ideal opportunity to walk around any
machinery and use your senses to look and listen for any abnormalities in machine
condition, for example:
 Seal Leaks - vibration analysis cannot diagnose seal leaks however a very quick
visual inspection can.
 Bearing wear and lack of lubrication – if a machines’ bearings are worn or
inadequately lubricated they can emit audible noise. This can be confirmed
by vibration analysis.
 Cavitation sounds like gravel is passing through the pump and will emit an
audible noise. This can be confirmed by vibration analysis.
 Low pump discharge pressure or high motor current readings can indicate a
pump is not running efficiently.

Use Your Senses!

Motor Phase Current Analysis
Motor Phase Current analysis is used for rotor fault diagnosis of AC induction
motors to detect broken rotor bars and air gap eccentricity. The technique is
based upon frequency analysis of the phase current supplying the motor.

 Motor Phase Current Analysis is

generally applicable to large AC
induction motors.

 Measurements can be carried

out using portable equipment
by connecting a current clamp
to the low voltage side of the
power transformer.
Rogowski Coil Analysis
Rogowski coil analysis is used to assess the condition of stator winding insulation in
high voltage electrical machines such as power generators. Assessment of
condition is based upon the examination of high frequency signals caused by
partial discharge activity measured using Rogowski coils.

 Rogowski coil analysis is generally undertaken on larger, high voltage machines

such as electrical generators and motors of 6.6 kV or greater.

 Rogowski coils can be fitted to machines during manufacture on the client’s

request.

 Specialist data collection equipment is required to interface to the fitted

instrumentation to carry out diagnostics.
Performance Monitoring
Performance monitoring can give an indication of machine condition and running
efficiency based upon calculations of measured operating and process parameters.
Performance indicators can be utilised to:
 Ensure the efficient running of power turbines and compressors; where
reduced efficiency can result is substantial financial loss.
 Indicate when a gas turbine should be water washed.
 Indicate deterioration in machine condition based upon a reducing trend in
performance.
Performance Indicator Example – MOL Pump
Calculating the ratio of the energy out of a system to the energy in, will give and
indication of the system’s efficiency. In this example the differential pressure
multiplied by the flow rate of a pump was divided by the motor current. Over
time it was seen that the pump performance indicator showed a deteriorating
trend. The pump was overhauled and found to have badly worn impellers. After
overhaul the performance indicator improved.
MOL Pump

10.00

P  Q
Pump _ Performance _ Indicator 
I
7.50
Performance Indicator

5.00

2.50

0.00
19/03/1999

19/05/1999

19/07/1999

19/09/1999

19/11/1999

19/01/2000

19/03/2000

19/05/2000

19/07/2000

19/09/2000

19/11/2000

19/01/2001

19/03/2001

19/05/2001

19/07/2001

19/09/2001

19/11/2001

19/01/2002

19/03/2002

19/05/2002

19/07/2002

19/09/2002

19/11/2002
Date