5.5K views

Uploaded by ronfrend

- Vibration analysis lectures
- Vibrations in Rotating Machinery (7th Int'l Conf.) (IME, 2000) WW
- Vibration Diagnostics
- Rotating Machinery Vibration - From Analysis to Troublesh Ooting
- 36199644-Vibration-Intermediate.pdf
- Basics of Vibration Analysis
- Section II - Basic Vibration Theory
- Vibration Training Quick Reference
- Vibration Adv 0402
- Vibration Analysis
- VA II Training
- Machine Vibration ( a Beginners Guide)
- Vibration Analysis Handbook - James Taylor
- Level II Metric
- Vibration Basic 0400
- Basic Vibration Analysis
- Vibration Analysis Manual Rev0
- Vibration Analysis Level- 1 Updated [Compatibility Mode]
- M. Rades - Dynamics of Machinery 3
- Vibration

You are on page 1of 161

rotating equipment

1

Vibration Analysis on Rotating Equipment

CONTENTS

Vibration Examples------------------------------------------------------------------------------------------------- 9

Imbalance--------------------------------------------------------------------------------------------------------------- 9

Misalignment ---------------------------------------------------------------------------------------------------------10

Looseness--------------------------------------------------------------------------------------------------------------11

Rolling Element Bearing Defects----------------------------------------------------------------------------------13

VIBRATION THEORY--------------------------------------------------------------------------- 15

Simple Harmonic Motion-------------------------------------------------------------------------------------------15

RMS vs. PEAK-------------------------------------------------------------------------------------------------------18

Time Domain ---------------------------------------------------------------------------------------------------------19

The Frequency Domain ---------------------------------------------------------------------------------------------19

What is an FFT?------------------------------------------------------------------------------------------------------20

The FFT Analyzer ---------------------------------------------------------------------------------------------------20

Advantages of FFT Analyzers -------------------------------------------------------------------------------------21

Frequency Spans -----------------------------------------------------------------------------------------------------21

Measurement Basics -------------------------------------------------------------------------------------------------21

Spectrum---------------------------------------------------------------------------------------------------------------21

Parameter Selection------------------------------------------------------------------------------------------------ 22

Selecting displacement, velocity or acceleration----------------------------------------------------------------22

Accelerometers -------------------------------------------------------------------------------------------------------24

Theory -----------------------------------------------------------------------------------------------------------------25

The Demodulation Process -----------------------------------------------------------------------------------------26

Resonance Sources---------------------------------------------------------------------------------------------------28

A.C. Motor Example.------------------------------------------------------------------------------------------------29

FAILURE MODES-------------------------------------------------------------------------------- 32

Induction Motors--------------------------------------------------------------------------------------------------- 32

Mechanical or Electrical Effects-----------------------------------------------------------------------------------32

Armature Related Problems ----------------------------------------------------------------------------------------32

Stator Related Problems --------------------------------------------------------------------------------------------33

Broken Rotor Bars ---------------------------------------------------------------------------------------------------34

DC Motors----------------------------------------------------------------------------------------------------------- 34

How DC Power Is Created. ------------------------------------------------------------------------------------34

DC Systems and Controls-------------------------------------------------------------------------------------------36

DC Control Firing Cards--------------------------------------------------------------------------------------------37

S.C.R. problems ------------------------------------------------------------------------------------------------------38

Example of a Firing Card Fault ------------------------------------------------------------------------------------38

2

Vibration Analysis on Rotating Equipment

Example of a Comparitor Card Defect ---------------------------------------------------------------------------41

Importance of Exact RPM ------------------------------------------------------------------------------------------42

Example of Mechanical -v- Electrical Frequencies ------------------------------------------------------------43

Imbalance--------------------------------------------------------------------------------------------------------------45

Vibration due to imbalance-----------------------------------------------------------------------------------------46

Misalignment ---------------------------------------------------------------------------------------------------------51

Looseness--------------------------------------------------------------------------------------------------------------54

Aerodynamic cross coupling ---------------------------------------------------------------------------------------60

Surging-----------------------------------------------------------------------------------------------------------------60

Choking or Stone Walling ------------------------------------------------------------------------------------------61

Elasto Hydrodynamic Lubrication --------------------------------------------------------------------------------62

First Stage of Bearing Failure--------------------------------------------------------------------------------------63

Second Stage of Bearing Failure ----------------------------------------------------------------------------------64

Third Stage of Bearing Failure-------------------------------------------------------------------------------------65

Fourth Stage of Bearing Failure -----------------------------------------------------------------------------------66

Bearing Defect Frequency Calculation ---------------------------------------------------------------------------67

Analysis of bearing defects -----------------------------------------------------------------------------------------70

Balancing ------------------------------------------------------------------------------------------------------------ 75

In-place Balancing ---------------------------------------------------------------------------------------------------75

Vibration Related to Imbalance------------------------------------------------------------------------------------78

How to Balance - Single Plane-------------------------------------------------------------------------------------79

Single Plane Vector Method of Balancing-----------------------------------------------------------------------80

Four-step Method of Balancing Single Plane -------------------------------------------------------------------82

Balancing in One Run -----------------------------------------------------------------------------------------------85

Introduction --------------------------------------------------------------------------------------------------------- 91

Base cause-------------------------------------------------------------------------------------------------------------91

Failure type------------------------------------------------------------------------------------------------------------91

External manifestation-----------------------------------------------------------------------------------------------92

Technology------------------------------------------------------------------------------------------------------------92

Parameter --------------------------------------------------------------------------------------------------------------92

Analysis----------------------------------------------------------------------------------------------------------------92

Interval-----------------------------------------------------------------------------------------------------------------92

Setup -------------------------------------------------------------------------------------------------------------------92

3

Vibration Analysis on Rotating Equipment

Stage 1 -----------------------------------------------------------------------------------------------------------------93

Stage 2 -----------------------------------------------------------------------------------------------------------------94

Stage 3 -----------------------------------------------------------------------------------------------------------------94

Stage 4 -----------------------------------------------------------------------------------------------------------------94

Conclusion ----------------------------------------------------------------------------------------------------------- 96

Measurement Windows ---------------------------------------------------------------------------------------------99

Averaging ----------------------------------------------------------------------------------------------------------- 100

Real Time Bandwidth and Overlap Processing --------------------------------------------------------------- 101

Octave Analysis ---------------------------------------------------------------------------------------------------- 102

Severity charts ------------------------------------------------------------------------------------------------------ 105

Two channel functions -------------------------------------------------------------------------------------------- 106

Representation by complex numbers --------------------------------------------------------------------------- 108

Cascade & waterfall plots----------------------------------------------------------------------------------------- 109

Triggering ----------------------------------------------------------------------------------------------------------- 109

Bodé plots ----------------------------------------------------------------------------------------------------------- 110

Orbits----------------------------------------------------------------------------------------------------------------- 110

Mechanical ---------------------------------------------------------------------------------------------------------- 115

Liquids & pumping systems-------------------------------------------------------------------------------------- 118

Air & gases---------------------------------------------------------------------------------------------------------- 120

Karman Vortices --------------------------------------------------------------------------------------------------- 122

Phase---------------------------------------------------------------------------------------------------------------- 127

Running machine Bump Test ------------------------------------------------------------------------------------ 130

Reverse Bump or “Plucking” the Suspect Part. --------------------------------------------------------------- 130

Impact hammer ----------------------------------------------------------------------------------------------------- 131

4

Vibration Analysis on Rotating Equipment

Plates---------------------------------------------------------------------------------------------------------------- 137

Springs --------------------------------------------------------------------------------------------------------------- 140

Rubber --------------------------------------------------------------------------------------------------------------- 142

Damping ------------------------------------------------------------------------------------------------------------- 145

Changing the Mass------------------------------------------------------------------------------------------------- 146

Changing the stiffness --------------------------------------------------------------------------------------------- 146

5

Vibration Analysis on Rotating Equipment

List of illustrations

Figure 1 Velocity spectrum showing imbalance ______________________________________________ 9

Figure 2 - Velocity spectrum showing fan imbalance _________________________________________ 10

Figure 3 Velocity spectrum of misaligned fan - radial ________________________________________ 11

Figure 4 Velocity spectrum of misaligned fan - axial _________________________________________ 11

Figure 5 Velocity spectrum from a loose fan drive motor______________________________________ 12

Figure 6 Envelope spectrum of a fan drive motor with loose bearing ____________________________ 13

Figure 7 Enveloped acceleration spectrum of bearing - inner race defect _________________________ 14

Figure 8 Inner race spall_______________________________________________________________ 14

Figure 9 Simple Harmonic Vibration _____________________________________________________ 16

Figure 10 Integration from acceleration to velocity __________________________________________ 17

Figure 11 Integrating to displacement ____________________________________________________ 18

Figure 12 Peak -v- RMS _______________________________________________________________ 19

Figure 13 Compression mode accelerometer _______________________________________________ 24

Figure 14 Shear mode accelerometer _____________________________________________________ 24

Figure 15 Simple modulation example ____________________________________________________ 25

Figure 16 Bearing modulation example ___________________________________________________ 26

Figure 17 Demodulation process ________________________________________________________ 27

Figure 18 Enveloping process___________________________________________________________ 27

Figure 19 Fast Fourier Transform _______________________________________________________ 28

Figure 20 FFT - 3D view_______________________________________________________________ 28

Figure 21 Two channel time waveform - bearing defect _______________________________________ 29

Figure 22 High frequency waterfall ______________________________________________________ 30

Figure 23 Enveloped acceleration spectrum________________________________________________ 30

Figure 24 Comparison - velocity to envelope _______________________________________________ 31

Figure 25 The creation of DC power _____________________________________________________ 35

Figure 26 FFT spectrum of half wave rectification___________________________________________ 36

Figure 27 FFT spectrum of full wave rectification ___________________________________________ 36

Figure 28 Basic DC system circuit _______________________________________________________ 37

Figure 29 FFT spectrum full wave DC firing card frequencies _________________________________ 38

Figure 30 FFT spectrum after repair _____________________________________________________ 38

Figure 31 FFT spectrum showing half wave firing card frequencies _____________________________ 39

Figure 32 FFT spectrum of same motor (no load) ___________________________________________ 40

Figure 33 FFT spectrum showing comparitor card defect._____________________________________ 41

Figure 34 FFT after the comparitor card was replaced _______________________________________ 42

Figure 35 DC motor components ________________________________________________________ 42

Figure 36 FFT from a 5 HP motor - full wave rectified _______________________________________ 43

Figure 37 Same motor - speed lowered by 25% _____________________________________________ 44

Figure 38 Imbalance slide 1 ____________________________________________________________ 46

Figure 39 Imbalance slide 2 ____________________________________________________________ 46

Figure 40 Imbalance slide 3 ____________________________________________________________ 47

Figure 41 Imbalance slide 4 ____________________________________________________________ 47

Figure 42 Imbalance slide 5 ____________________________________________________________ 48

Figure 43 Imbalance slide 6 ____________________________________________________________ 48

Figure 44 Imbalance slide 7 ____________________________________________________________ 49

Figure 45 Imbalance slide 8 ____________________________________________________________ 49

Figure 46 Imbalance slide 9 ____________________________________________________________ 50

Figure 47 Misalignment slide 1__________________________________________________________ 51

Figure 48 Misalignment slide 2__________________________________________________________ 51

Figure 49 Misalignment slide 3__________________________________________________________ 52

Figure 50 Misalignment slide 4__________________________________________________________ 52

Figure 51 Misalignment slide 5__________________________________________________________ 53

Figure 52 Looseness slide 1 ____________________________________________________________ 54

Figure 53 Looseness slide 2 ____________________________________________________________ 54

6

Vibration Analysis on Rotating Equipment

Figure 55 Looseness slide 4 ____________________________________________________________ 55

Figure 56 Looseness slide 5 ____________________________________________________________ 56

Figure 57 Looseness slide 6 ____________________________________________________________ 56

Figure 58 Looseness slide 7 ____________________________________________________________ 57

Figure 59 Looseness slide 8 ____________________________________________________________ 57

Figure 60 Looseness slide 9 ____________________________________________________________ 58

Figure 61 Aerodynamic forces __________________________________________________________ 60

Figure 62 Elasto-hydrodynamic lubrication ________________________________________________ 62

Figure 63 Loss of Lubricant - Ball Bearing Inner Race Courtesy of the Barden Corporation__________ 63

Figure 64 Loss of Lubricant - Roller Bearing Courtesy of the Torrington Company _________________ 64

Figure 65 Waterfall plot from a damaged motor bearing ______________________________________ 65

Figure 66 Early Fatigue - Ball Bearing Courtesy of the Barden Corporation ______________________ 66

Figure 67 Developed Fatigue on Roller Bearing Courtesy of the Torrington Company ______________ 66

Figure 68 Ball Bearing Terminology _____________________________________________________ 68

Figure 69 Waterfall of early damage to a motor bearing collected every 1.5 hrs over 14 days _________ 69

Figure 70 Bearing damage severity assessment chart ________________________________________ 70

Figure 71 Demodulated acceleration spectrum from a dry bearing ______________________________ 70

Figure 72 Demodulated acceleration spectrum of a marked bearing_____________________________ 71

Figure 73 Demodulated acceleration spectrum from a slightly more heavily marked bearing _________ 72

Figure 74 Time waveform from a marked bearing.___________________________________________ 72

Figure 75 Time waveform from a heavily marked bearing _____________________________________ 73

Figure 76 Velocity spectrum from a spalled bearing _________________________________________ 73

Figure 77 Sources of imbalance _________________________________________________________ 75

Figure 78 Assembly tolerance stack up ____________________________________________________ 76

Figure 79 Heavy spot _________________________________________________________________ 76

Figure 80 Units of measure of imbalance __________________________________________________ 77

Figure 81 Mass centre displacement______________________________________________________ 77

Figure 82 Force due to imbalance _______________________________________________________ 78

Figure 83 The vector diagram___________________________________________________________ 80

Figure 84 Simplified vector diagram______________________________________________________ 81

Figure 85 Additional corrections ________________________________________________________ 82

Figure 86 Direction to shift the weight ____________________________________________________ 83

Figure 87 Sample problem vector diagram_________________________________________________ 84

Figure 88 Determining the flash angle ____________________________________________________ 86

Figure 89 Typical tap block for mounting an accelerometer ___________________________________ 87

Figure 90 Accelerometer mounting techniques a-d___________________________________________ 89

Figure 91 Accelerometer mounting techniques e-g___________________________________________ 90

Figure 92 Overview of accelerometer mounting techniques ____________________________________ 90

Figure 93 PFA development for rolling element bearings _____________________________________ 93

Figure 94 PFA for a main motor_________________________________________________________ 95

Figure 95 General severity chart for vibration _____________________________________________ 105

Figure 96 Vector addition of 2 vibrations _________________________________________________ 108

Figure 97 Cascade of fan over 20mS ____________________________________________________ 109

Figure 98 Bode plots _________________________________________________________________ 110

Figure 99 Orbit showing misalignment___________________________________________________ 111

Figure 102 Harmonic series for the tone C. _______________________________________________ 120

Figure 103 Sonic vibration in a tube_____________________________________________________ 122

Figure 104 Karman vortices ___________________________________________________________ 123

Figure 107 Campbell Diagram _________________________________________________________ 125

Figure 109 Mode shape readings _______________________________________________________ 127

Figure 110 Phase relationships_________________________________________________________ 128

Figure 111 Phase / frequency relationships @ resonance ____________________________________ 129

Figure 112 Impact hammer response ____________________________________________________ 132

Figure 113 Impact hammer specification sheet_____________________________________________ 132

7

Vibration Analysis on Rotating Equipment

Figure 115 Stress/strain diagram for rubber ______________________________________________ 140

Figure 116 Resonance Curve __________________________________________________________ 141

List of Tables

Table 1 Speed of sound in liquids _______________________________________________________ 118

Table 3 Natural frequency calculation of uniform beams_____________________________________ 135

Table 4 Standard values for uniform beams _______________________________________________ 136

Table 5 Damping ranges of vibration isolators_____________________________________________ 139

Table 6 Whole body vibration (frequency ranges) __________________________________________ 148

List of Equations

Equation 3 Newton's 2nd law __________________________________________________________ 117

Equation 5 Differential equation of motion of a single-degree-of-freedom system _________________ 118

Equation 6 Velocity of sound in materials_________________________________________________ 119

Equation 7 Speed of sound in the ocean __________________________________________________ 119

Equation 8 General formula relating speed, wavelength & frequency ___________________________ 119

Equation 9 Newton-Laplace eq. for the speed of sound in a gas _______________________________ 121

Equation 10 Ratio of specific heats (gamma) ______________________________________________ 121

Equation 11 Speed of sound in a gas ____________________________________________________ 121

Equation 12 Karman vortices __________________________________________________________ 123

Equation 14 Amplitude magnification due to springs ________________________________________ 141

Equation 15 Natural frequency of a spring ________________________________________________ 142

Equation 16 Modulus of elasticity for rubber ______________________________________________ 144

8

Vibration Analysis on Rotating Equipment

The study of noise and vibration phenomena dates back centuries. The first recorded incidence of such

study was by Pythagoras in the sixth century B.C. who studied the origin of musical sounds and the

vibration of strings. In 1638 Galileo described the vibrations of pendulums, the phenomenon of resonance

and the factors influencing the vibration of strings. Euler in 1744 and Bernoulli in 1751 developed the

equation for the vibrations of beams and developed the normal modes for various boundary conditions. In

1882 Hertz developed the first successful theory for impact. So we can see that vibration analysis itself is

not new but some of the ways that we take the measurements and apply those measurements as machine

health diagnoses are very new.

In this section we will briefly take a look at some vibration examples of typical defects suffered by fans and

fan drives without delving too deeply into why!

Vibration Examples

When the novice analyst first carries out vibration analysis he will usually rush out and take a vibration

spectrum using the default parameters set up in the analyzer. We will carry on that noble tradition and look

at some spectra that have been collected from real machines and show typical examples of common defects.

Imbalance

Figure 1 shows a vibration spectrum that was taken at the sheave end of a centrifugal fan in the vertical

direction. The fan was driven from the AC motor via a V-belt and rotated at about 720 rpm. The AC drive

motor rotates at just under 1200 rpm.

The spectrum is a simply a graph of the vibration frequency on the bottom axis with the amplitude at that

frequency on the vertical axis. This spectrum is of velocity vibration so the amplitude units could be in

9

Vibration Analysis on Rotating Equipment

mm/s or ips (inches/second). The frequency is in cpm (cycles/minute) but it could have been displayed in

Hz (Hertz or cycles/second) or in orders (multiples of run speed). The spectrum was recorded from a

vibration transducer which was mounted radial to the shaft (vertical in this case).

Notice in the spectrum that there is one big spike which is labeled at 716.59 cpm and there are two much

smaller spikes just to the right. The first spike to the right is at 1187 cpm which equates to the run speed of

the motor and the second spike is at 1433.18 cpm which is exactly twice fan speed. Because the one spike

is so dominant that is the one that we are concerned about. A check with a stroboscope confirmed that the

fan was actually running at 717 rpm so the big spike of vibration is at exactly (within the precision of the

strobe) run speed. At this stage we are not concerned about the physics of why a vibration at run speed is

usually indicative of imbalance but we will look at our spectral explanation charts (see appendix 1) and

have a fair degree of confidence that the fan needs balancing.

Figure 2 shows a similar problem on a different fan but we see that the spectrum looks very similar with one

dominant spike at the run speed of the fan. Figure 1 amplitude was displayed with metric units and figure 2

with inch units but the shape of the spectrum is the same in both cases.

Misalignment

Probably 40% of all bearing and shaft failures are caused by misalignment of the components creating an

extra axial thrust on the bearings.

10

Vibration Analysis on Rotating Equipment

In figure 3 we see what initially looks like an imbalance condition of the fan, although the amplitudes are

relatively low.

However, in figure 4 we are now looking at the vibration taken axial to the shaft. If the problem was simple

imbalance of the fan we would expect all of the forces to be caused by centrifugal force and therefore acting

in a direction which was radial to the shaft. Again, looking at our spectrum explanation charts we see that,

on a belt driven train, a high axial velocity vibration relative to the radial vibration is almost always

indicative of component misalignment to the belt.

Looseness

Looseness exists when the component is not directly attached to the structure or rotating element and has a

relatively large clearance, allowing the component to rattle.

11

Vibration Analysis on Rotating Equipment

The above spectrum was recorded at the sheave end of the drive motor of an underground colliery main

ventilation fan. The motor was running at 590 rpm and immediately we see the large family of harmonics

of run speed. The amplitudes do not seem too high but the machine was massive and any vibratory forces

have to move the mass before we see a vibration. In this case the structure of the bedplate was cracked

causing parts of the structure to vibrate freely at the excitation frequency of the motor (speed).

Whenever we see multiples or sub-multiples of run speed vibration frequencies we immediately consider

the possibility of loose components.

12

Vibration Analysis on Rotating Equipment

The early stages of looseness can be detected in a similar manner, be looking for harmonics of run speed,

and using demodulated or enveloped acceleration readings. Figure 6 shows the early stages of looseness of

a bearing inside the fan drive motor. As the looseness deteriorates the envelope readings will decrease but

then the velocity readings will start to increase.

The primary tool in assessing bearing condition is the use of enveloped acceleration readings.

Figure 7 shows the envelope spectrum from a bearing with a severe spall in the inner race. Notice that the

spike at about 8,772 cpm is marked BPIR which stands for Ball Pass Outer Race. We will study bearing

defects in detail later but notice that the main defects are not multiples of run speed. In other words they are

non-synchronous.

13

Vibration Analysis on Rotating Equipment

14

Vibration Analysis on Rotating Equipment

Vibration Theory

The following section is meant as a primer to help the newcomer to vibration analysis understand some of

the terms used and to develop an understanding of the concepts.

To understand the concept of vibration analysis, it is important to realize that the motion of the measured

surface varies with time. The transducer converts the movement into an electrical signal which is passed to

the spectrum analyzer which in turn converts that signal from the time domain into the frequency domain.

The time domain waveform is composed of a machine’s response to many individual forces such as

imbalance, misalignment, gear meshing forces, rotating electrical fields, and many other factors. When

viewing the time domain data it can be quite difficult to separate these components of vibration. However,

in the frequency domain it is much easier to separate these elements to determine the importance of each.

Vibration amplitude is measured using three different parameters, acceleration, velocity and displacement.

The purpose of this section is to describe the relationship between each of these and how they are used on

rotating machinery.

Simple harmonic motion can be visualized by many common examples such as a pendulum, a mass and

spring combination, a rotating mechanism or a diving board- Figure 9 uses a pendulum. If the pendulum

swings back and forth 100 times in one minute, then the frequency is 100 cycles per minute. Similarly if a

machine is rotating 100 times in one minute, its speed is 100 revolutions per minute or 100 RPM.

The frequency of vibration is often expressed in terms of cycles per second or HERTZ after the German

physicist Heinrich Hertz. However for predictive maintenance techniques where rotational speed is often

the key to vibration peaks, cycles per minute are used in preference to Hertz.

In addition to frequency the amplitude is the other necessary quantity that must be known in order to

characterize vibration. In figure 9 the points B and C represent the extreme position of the pendulum and

the distance between them is the peak to peak displacement-

Amplitude meters are often calibrated to give the peak to peak value because it is the displacement extremes

that are of interest. In vibration work, the displacement is often expressed in terms of mils or micron. One

Mil is equal to 0.001 inch and one micron (µm) is 0.001 mm.

Since the pendulum is continuously moving, it has a velocity associated with each position and, like

displacement, the velocity also varies between a positive peak and a negative peak.

15

Vibration Analysis on Rotating Equipment

(48)

* FREQUENCY

* AMPLITUDE

... Displacement

... Velocity

... Acceleration

The amplitude is expressed in units of either displacement,

velocity or acceleration.

one cycle

C B

cycles per unit time, called the frequency. Frequency is usually

expressed in terms of cycles per second, or Hertz.

Figure 10 shows that at position B and C, the velocity is zero, and at position A the velocity is maximized,

first to the right, then to the left. Since the peak positive velocity occurs 1/4 cycle before the peak positive

displacement, velocity is said to lead displacement by 90°. The 90 degree phase lead is shown in the

diagram on figure 10. Velocity amplitude is expressed only in terms of zero to peak or zero to RMS.

The negative peak velocity differs only in direction, not magnitude. The rate of change of displacement is

the velocity, therefore if D is expressed in terms of inches, instead of the usual mils, then the product 2πfD

will be the velocity in inches per second which are the units used for velocity in vibration work.

16

Vibration Analysis on Rotating Equipment

(49)

A B A C A

Disp

Peak

Peak

to

Peak

C B

peak-to-peak displacement. Displacement meters are often calibrated

in peak-to-peak units. The amplitude is one half of the peak-to-peak

value for a sine wave.

A B A C A

Disp

Vel

C B

Highest

Velocity

o

where displacement is maximum. Therefore a 90 phase shift exists

between displacement and velocity. The velocity amplitude is directly

proportional to frequency for a given displacement.

As velocity is continuously changing, an acceleration is also associated with the velocity, this acceleration is

also associated with the motion. Acceleration is the third way to express vibration amplitude. Figure 11

shows that at position B and C the acceleration is maximum. Just prior to point B, velocity is to the right

and just after it is to the left. At B therefore the rate of change of velocity, the acceleration, is maximum.

Conversely just prior to point A velocity is increasing and just after, it is decreasing. Therefore the rate of

change of velocity (the acceleration) must be zero at A. Note that acceleration reaches its maximum at

Points B and C just as displacement does, but at B acceleration is to the left whereas the displacement is to

the right. The maximum acceleration to the right occurs 1/2 a cycle before the maximum velocity to the

right and acceleration is said to lead displacement by 180°. Acceleration leads velocity by 90°.

The diagram in figure 11 shows these phase leads and also the acceleration amplitude relationship, A =

(2πf)2D. This says that for any given value of displacement, the acceleration is proportional to the square of

the frequency.

The unit of acceleration is the “g” which is equal to 9.81 m/sec2 and is derived from the acceleration due to

earth’s gravity.

17

Vibration Analysis on Rotating Equipment

(50)

A B A C A

Disp

Vel

C B

Accel

A

Acceleration Acceleration

o

maximum to the right, a 180 phase shift. Acceleration amplitude varies

as the square of frequency for a given value of displacement.

Peak

Avg RMS

Peak

Amplitude

to

Peak

Time

RMS = 0.707 x Peak

and Peak amplitude values for sine waves are not valid for

combination or random waveforms.

The rms or root mean square value is calculated by breaking the waveform down into a number of points,

squaring the amplitude value of each point, calculating the mean of the squared values and then finding the

square root of the mean.

Using rms values can be compared to the use of rms in electrical circles i.e. stereo speakers power values

are measured in rms values. Electrical (AC) voltage is also measured in rms. This, like vibration signals, is

a continuously varying quantity, ranging from zero to a peak value. To measure only the peak value may be

misleading since the voltage is actually at a peak for only a small portion of the cycle. During most of the

cycle the value of the instantaneous voltage is somewhere between zero and peak.

RMS, then, is an attempt to apply a single quantitative value -which reflects the effective value of this

varying function. This same logic applies to vibration. Velocity is a quantitative measure of the effective

velocity and reflects the power or energy being used to vibrate the machine mass.

Peak value is the maximum amplitude seen during the measurement. When using FFT analyzers care

should be taken when evaluating peak or rms severity as the peak amplitude in the spectrum is derived from

a sine wave. True peak can be seen in the time waveform.

18

Vibration Analysis on Rotating Equipment

(51)

Peak

Avg RMS

Peak

Amplitude

to

Peak

Time

Peak

Peak

to

Amplitude

RMS Peak

Time

Complex Waveform

Time Domain

The traditional way of observing signals is to view them in what is called the time domain. The time

domain is a record of what happened to a parameter compared to time- Typically the signal would be

displayed on an oscilloscope. With respect to machinery vibration, analysis of signals in the time domain

can be very difficult and is far easier in the frequency domain

If we now convert a time waveform to the frequency domain we will get a totally different picture. We now

have axes of amplitude v frequency instead of amplitude -v- time. Every sine wave separated out by the

FFT appears as a separate line. Its height represents its amplitude, its position represents its frequency.

The method most analyzers use to transform signals from the time domain to the frequency domain is called

:-

19

Vibration Analysis on Rotating Equipment

What is an FFT?

The fast Fourier transform (FFT) is an algorithm for transforming data in the time domain to the frequency

domain. Most analyzers have an FFT processor, which performs this transformation automatically and then

stores the computed spectra into memory.

We cannot transform to the frequency domain in a continuous manner. We therefore must sample and

digitize the time domain input. The number of samples determines the resolution (number of lines) of

frequency.

Most analyzers offer resolutions of 100,200,400,800,1600,3200 or even 6400 Lines.

FFT Spectrum Analyzers take a time varying input signal, like you would see on an oscilloscope trace, and

compute its frequency spectrum.

Fourier's theorem states that any waveform in the time domain can be represented by the weighted sum of

sines and cosines. The FFT spectrum analyzer samples the input signal, computes the magnitude of its sine

and cosine components, and displays the spectrum of these measured frequency components.

Many of these measurements were once done using analog spectrum analyzers. In simple terms, an analog

filter was used to isolate frequencies of interest. The signal power, which passed through the filter, was

measured to determine the signal strength in certain frequency bands. By tuning the filters and repeating the

measurements, a spectrum could be obtained.

An FFT spectrum analyzer works in an entirely different way. The input signal is digitized at a high

sampling rate, (2.56 x Fmax usually). Nyquist's theorem says that as long as the sampling rate is greater

than twice the highest frequency component of the signal, then the sampled data will accurately represent

the input signal. Certain analyzers pass the input signal through an analog filter, which attenuates all

frequency components above Fmax by 90 dB to make sure that Nyquist's theorem is satisfied. This is the

anti-aliasing filter. The resulting digital time record is then mathematically transformed into a frequency

spectrum using an algorithm known as the Fast Fourier Transform or FFT. The FFT is simply a clever set

of operations which implements Fourier's theorem. The resulting spectrum shows the frequency

components of the input signal.

Now here's the interesting part. The original digital time record comes from discrete samples taken at the

sampling rate. The corresponding FFT yields a spectrum with discrete frequency samples. In fact, the

spectrum has less than half as many frequency points as there are time points (remember Nyquist's

theorem). Suppose that you take 1024 samples at 2560 Hz. It takes 0.4 Seconds to take this time record.

The FFT of this record yields 400 frequency points or lines, but over what frequency range? The highest

frequency will be determined by the in-built ratio of Fmax to data sampling rate - 2.56. The lowest

frequency is just the Fmax divided by the number of lines:

Fmax = data sampling rate / 2.56

No. Of Lines = No samples / 2.56

Bin resolution = Fmax / No. of lines

= (2560 / 2.56) / (1024 / 2.56)

= 2.5 Hz (the same as the lowest measurable frequency)

20

Vibration Analysis on Rotating Equipment

Everything below 2.5 Hz (for this example) is considered to be DC. The output spectrum thus represents

the frequency range from DC to 1000 Hz with points every 2.5 Hz.

The advantage of this technique is its speed. Because FFT spectrum analyzers measure all frequency

components at the same time the technique offers the possibility of being hundreds of times faster than

traditional analog spectrum analyzers. In the case of a 1000 Hz span and 400 resolvable frequency bins, the

entire spectrum takes only 400 mS to measure. To measure the signal with higher resolution the time record

is increased, but again, all frequencies are examined simultaneously, providing an enormous speed

advantage.

Frequency Spans

Before we continue, let's clarify a couple of points about our frequency span. We just described how we

arrived at a DC to 1000 Hz frequency span using a 400 mS time record. Because the signal passes through

an anti-aliasing filter at the input, the entire frequency span is not useable. A typical filter has a flat

response from DC to 1000 Hz and then rolls off steeply from 1000 Hz to 2.56 kHz. The range between

1000 Hz and 2.56 kHz is therefore not useable and the actual displayed frequency span stops at 1000 Hz.

There is also a frequency bin labeled 0 Hz (or DC). This bin actually covers the range from 0 Hz to 2.5 Hz

(the lowest measurable frequency) and contains the signal components whose period is longer than the time

record (not only DC). So our final displayed spectrum contains 400 frequency bins. The first covers 0 - 2.5

Hz, the second 2.5 - 5 Hz, and the 400th covers 997.5 - 1000 Hz.

The length of the time record determines the frequency span and resolution of our spectrum. What happens

if we make the time record 800 mS or twice as long? Well, we ought to get 2048 time points (sampling at

2560 Hz) yielding a spectrum from DC to 1000 Hz with 1.25 Hz resolution containing 800 points. But the

analyzer places some limitations on this. One is memory. If we keep increasing the time record, then we

would need to store more and more points. (0.00125 Hz resolution would require 2,048,000 values.)

Another limitation is processing time. The more points you take, the longer the processing time.

Measurement Basics

An FFT spectrum is a complex quantity, This is because each frequency component has a phase relative to

the start of the time record. (Alternately, you may wish to think of the input signal being composed of sines

and cosines.) If there is no triggering, then the phase is random and we generally look at the magnitude of

the spectrum. If we use a synchronous trigger then each frequency component has a well-defined phase.

Spectrum

The spectrum is the basic measurement of an FFT analyzer. It is simply the complex FFT. Normally, the

magnitude of the spectrum is displayed. The magnitude is the square root of the FFT times its complex

conjugate. (Square root of the sum of the real (sine) part squared and the imaginary (cosine) part squared).

The magnitude is a real quantity and represents the total signal amplitude in each frequency bin,

independent of phase.

If there is phase information in the spectrum, i.e. the time record is triggered in phase with some component

of the signal, then the real (cosine) or imaginary (sine) part or the phase may be displayed. The phase is

simply the arc tangent of the ratio of the imaginary and real parts of each frequency component. For

vibration measurements phase is usually considered to be relative to the trigger pulse.

21

Vibration Analysis on Rotating Equipment

Parameter Selection

As previously discussed, displacement amplitude is higher at lower frequencies. Therefore when motions

are being measured a displacement measurement is in order because frequencies of interest on the shaft are

limited to 20 or so orders of rotation. For a 3600 rpm machine, 20 orders is a frequency of 1200 Hz. At

that frequency 0.5 in/sec is 0.13 mils pk-pk, very small but certainly a measurable value.

For higher frequencies however, significant vibration has a displacement value which -is too small to

conveniently measure and velocity or acceleration is more appropriate. Velocity measurements are

especially good for a number of reasons.

The most prominent advantage of a velocity measurement is that the value of rms velocity is related to the

potential for mechanical damage, regardless of the frequency. The many published vibration severity charts

are based on this principle. As an example suppose a displacement of 0.l mils is observed, is this severe?

At 6 Hz this is not severe at all; at 60 Hz this is rough but at 200 Hz this is very rough and should not be

permitted for machines up to the 100HP class. Now suppose a velocity of 0.6 in/sec (15mm/s) is observed.

Is this severe? The answer is “Yes, this is severe regardless of the frequency”.

Newton’s second law (F=ma) tells us that the acceleration of a body is directly proportional to the force

applied to the body. In other words the acceleration vibration gives a good indication of impactive forces

inside the machine such as bad bearings.

In summary, displacement measurements are good from 0 Hz to 500 Hz, velocity up to 1 kHz and

acceleration from 2 Hz to 20 kHz depending on the design of the accelerometer. In applying this to rotating

machinery displacement measurements are relative readings of the displacement of the shaft to a reference,

usually the bearing. Velocity and acceleration measurements are usually made on the bearing cap or on the

machine casing in way of a structural web to enhance the transmission of vibration to the pick-up point.

Consider a rotating machine (a motor) which has, for example, an out of balance condition on the rotor so

that for every revolution of that rotor the out of balance mass generates a centripetal (opp. to centrifugal)

force. We place our transducer on the drive end of the motor in the vertical direction, as near as possible to

the bearing and couple the transducer to a spectrum analyzer. The transducer sees the force once per rev. of

the rotor as a “simple harmonic motion”. That is to say that the machine surface will cause the transducer to

move in a downwards direction with the machine as the force itself is acting downwards and will cause the

transducer to move upwards when the machine is moving up etc. The output of the transducer will depend

on what type of transducer we are using.

• Displacement transducers will give an output proportional to the linear displacement

of the transducer in thousandths of a inch or micron.

• Velocity transducers will give an output proportional to the linear speed (velocity) of

the transducer in inches/second or millimeters/second.

• An accelerometer will give an output which is proportional to the acceleration of the

transducer in G’s or inches/second/second or meters/second/second.

For predictive maintenance purposes we use accelerometers almost exclusively so we will concentrate on

them for now. According to Newton’s Second Law

F=m⋅⋅x a

22

Vibration Analysis on Rotating Equipment

m = the mass

a = the acceleration

So immediately we see that the output from the accelerometer is directly proportional to the internal forces

acting on the machine. Newton also says that for a rotating body

F=m⋅⋅ϖ2⋅r

where ϖ = the rotational speed in radians/second

r = the radius at which the force is acting.

As we know that the acceleration is proportional to the force and we assume that the mass and radius of

force of the machine stay constant, then we may safely say that the acceleration is also proportional to the

square of the speed.

ϖ 2⋅ r

a=ϖ

The important point here is that the faster the machine goes, or the higher up the frequency range we go, the

acceleration amplitudes must increase for a given force even if there is nothing wrong with our machine.

However, we know that acceleration is simply the rate of change of velocity. So if we integrate our

acceleration reading with respect to time we will get a velocity reading. Integrating acceleration will

change our value from:

inches/second2

to

inches/second

effectively finding the square root of the acceleration (for time). We have already said that we have a

concern that the acceleration increases with frequency, so if we need a value that is independent of

frequency for severity analysis purposes we can use the velocity reading.

Back to our motor. If we plot the acceleration against time (time domain) we would see a sine wave which

is the result of simple harmonic motion. This is the signal that is passed along to the analyzer. The analyzer

will then convert this time domain signal into a frequency domain signal either as acceleration or as an

integration from acceleration into velocity. Either way, the out of balance condition will show itself in the

frequency domain as a single spike at a frequency which corresponds to the run speed of the machine. For

example, if the motor is running at 1,200 rev/min the spike will have a frequency at 1,200 cycles/minute

(cpm) or 20 Hertz (Hz).

23

Vibration Analysis on Rotating Equipment

Accelerometers

Looking at the figure above we see a schematic of an accelerometer. Modern accelerometers are available

as compression mode or shear mode. Generally speaking the shear mode accelerometer offers better axial

sensitivity with much better mechanical integrity. In other words the shear mode accelerometer is not as

affected by thermal transients and gives better accuracy for the axis in which it is mounted.

Many low cost industrial accelerometers are now shear mode. For off-line measurements the accelerometer

will probably be connected to a magnet and the magnet positioned at a pre-determined point every time a

reading is taken. However, the response from the accelerometer is better if it is permanently mounted.

Permanently mounting an accelerometer should be done with care. The way the accelerometer is mounted

will affect the resonant frequency and, hence, the useable frequency range. By far the best way to mount an

accelerometer is to spot face the subject surface and drill and tap it to accept the stud for the accelerometer.

However, on a motor it is usually not practical to drill into the motor frame for obvious reasons. The best

24

Vibration Analysis on Rotating Equipment

alternative to stud mounting is to have tap blocks made with a tapped hole that will accept the accelerometer

stud.

Theory

But before we look at any case histories using DEMODULATION we should be clear in our mind about

exactly what is MODULATION.

A signal may be said to be amplitude modulated if the amplitude of that signal is changing over a period of

time because of the influence of another signal. The example above was taken from a large steam turbine

running at 3600 rpm. The run speed signal is being MODULATED by a signal at 4 Hz which is probably a

foundation resonance. This type of modulation is commonly found in maintenance applications but

consider the example below.

25

Vibration Analysis on Rotating Equipment

Here we see a vibration at 2 kHz which has been modulated slightly more than three times within the time

period (50 mS which equates to 1 revolution of the inner race). The 2 kHz vibration is the resonance of the

bearing which is being excited by the bearing outer race frequency (3.07 x run speed). The excitation of the

2 kHz frequency by the bearing defect on the outer race causes the 2 kHz amplitude to be changed like the

“roller coaster” example above. In other words the bearing outer race frequency is modulating the bearing

resonance frequency. The demodulation process extracts the modulating frequency to produce a time

waveform which can be handled by the F.F.T. process.

When we DEMODULATE the above reading we are not interested in the 2 kHz frequency but we are

interested in the outer race defect frequency which is:

(1000/50*3.07) Hz = 61.4 Hz.

As can be seen from Figure 2, the modulation is at this frequency. In vibration terms, demodulation is a

way of extracting the rate of occurrence of high frequency resonances.

The time waveform of a machine with a bearing in the early stages of deterioration will look like the top

plot below. The bearing excitation resonance is shown as small, high frequency pulses sitting on top of the

high amplitude, low frequency vibration.

26

Vibration Analysis on Rotating Equipment

The demodulator circuit now passes the signal through a high pass filter to give the time waveform shown in

the lower section of the plot.

With the time domain signal in this format the F.F.T. conversion would give a single spike in the frequency

domain at the resonant frequency which we have earlier said is not what we want. To modify the signal so

as to be suitable for F.F.T. we must “envelope” (figure above) each parcel of energy by first rectifying and

then passing the signal through a smoothing R-C (resistance-capacitive) circuit.

27

Vibration Analysis on Rotating Equipment

The signal is NOW passed through the F.F.T. and we get a spike in the frequency domain at the bearing

defect frequency (figures above and below).

Resonance Sources

When taking a demodulated reading we must first decide on which filter setting to use that will allow the

carrier signal to pass without allowing the low frequency, high amplitude noise to pass. Conventional

thinking will tell you that the resonance frequency which we are using as the carrier wave is always the

resonant frequency of the bearing; while this is often the case it is not always so. For vibration readings, the

accelerometer which we will use to detect the signal will probably be sitting on top of a magnet which will

give a structural resonance in the 1.5 to 4 kHz range (typically). The bearing housing will have its own

28

Vibration Analysis on Rotating Equipment

resonance, the machine structure will have its own resonance. In short, the carrier wave signal resonance

could be coming from any part of the mechanical structure.

If we are taking a reading with a non-vibration parameter we will probably be utilizing a different carrier

signal so we may have to use a different high pass or band pass filter. Ultra-sound data are heterodyned to

the audible range so demodulating at 5 to 8 kHz gives acceptable results while A.C. electric current should

be demodulated from the A.C. frequency of 60 Hz or 50 Hz.

This plot shows the signal from the inboard bearing of a 35 H.P. A.C. motor operating a belt-driven fan.

The 2 upper plots are the time domain signal in two planes over a period of 640 mS. The lower plots show

the time domain (left) and frequency domain (right) over a 50 mS period of the lower 640 mS plot. Note

that the frequency spectrum shows spikes at 2 kHz and 3 kHz while the time domain plots show an “angel

fish” pattern which is classic of a bearing defect. Note also that the lower left portion of the plot is a zoom

of the windowed part of the long time record. This shows a detail of the one “angel fish” and the amplitude

can be seen to be passing from positive to negative and back again many times during the life of a single

angel fish - i.e. a high frequency oscillation. This leads us to the conclusion that this is the frequency of 2

and/or 3 kHz seen in the spectrum and one or both of these frequencies are the result of impacts and

subsequent ring down and they are occurring at the resonant frequency of part of the mechanical structure.

29

Vibration Analysis on Rotating Equipment

The plot above shows a time/frequency cascade of the same time interval cropped below 0.001G. This

clearly shows the modulation of the 2 kHz frequency while the 3 kHz frequency is static. The modulation

has been calculated to be equal to the bearing outer race defect frequency of the motor inboard bearing.

Every time one of the bearing balls passes a defect on the outer race, the ball impacts on the defect causing

the 2 kHz vibration to suddenly rise and then ring down. The 2 kHz is the resonant frequency and the

bearing defect frequency (outer race) is the modulating frequency.

The figure below shows the demodulated spectrum on the left with waterfall plot on the right above a trend

of the defect frequency.

Note that the demodulated spectrum is clean and extremely easy to analyze. The spikes occur at the bearing

defect frequency (outer race) with multiple harmonics but there is no sign of the resonant frequency because

this high frequency has been removed during its use in the demodulation process. The frequency range of

30

Vibration Analysis on Rotating Equipment

the spectrum is such that the frequency of the impacts is clearly visible but we do not need to see the

resonant frequency. The last spectrum in the waterfall is lower than the previous spectrum due to greasing

of the motor bearings which lowered the amplitude at which the impacts caused the bearing to vibrate at

resonance.

This figure shows a similar defect on another machine but here the velocity spectrum (left) is displayed

alongside the demodulated spectrum (right). Note that the demodulated spectrum is much cleaner and

easier to analyze.

31

Vibration Analysis on Rotating Equipment

Failure Modes

Induction Motors

Vibration of electrical motors can be either mechanical or electrical in origin. Mechanical problems may

include imbalance, misalignment, defective bearings and looseness. Vibration caused by electrical

problems are normally the result of unequal magnetic forces acting on the rotor or stator. The unequal

magnetic forces may be due to open or shorted windings, broken rotor bars, unbalanced phases, unequal air

gap and other similar problems. Generally, the largest component frequency of vibration resulting from

these electrical problems will be 1 x RPM and, this will appear similar to imbalance. A common way to

check for electrical vibration is to observe the change in vibration amplitude the instant electrical power is

disconnected from the unit.

If the vibration disappears the instant the power is shut off, the vibration is likely to be due to electrical

problems. If this is the case conventional electrical testing procedures can be carried out to pinpoint the

true cause of vibration. On the other hand, if the vibration amplitude decreases only gradually after power

is disconnected, the problem is more likely to be mechanical in nature. Perhaps an even better indication of

the contribution of electrical problems is by observing the time waveform of the vibration as power is

disconnected.

Electrical problems with induction motors will often cause the motor load current ammeter to swing or

pulsate in a cyclic manner. If phase readings are taken, or if the motor has a strobe light flashed at run

speed, it will be seen that the phase is erratic and instead of the strobe “freezing” the rotor, the rotor will

appear to swing back and forth. This pulsating vibration common with induction motors will either be a

single frequency whose amplitude is being modulated or it will be a beat between two frequencies of

vibration which are very close together. If the nature of the pulsating vibration can be determined, this can

help significantly to identify the specific problem as discussed in the following paragraphs.

Typical problems associated with the rotor or armature of an induction motor which cause electrical

vibration include:

• Broken rotor bars This type of defect is best discovered by the use of motor

current analysis and comparing the height of the slip * No. of poles sidebands around line

frequency to previous levels.

• Wound rotor windings Defects in the rotor will cause a modulation of rotor bar pass

or stator slot passing frequencies at run speed. In other words these frequencies will have

sidebands of 1x. If the machine has a “beat” at slip frequency this is usually due to a defect in

one rotor winding phase such as a broken conductor or bad brush.

• A bowed rotor This defect usually occurs on very large, horizontal motors

where the motor has sat idle for an extended period and the weight of the rotor causes a sag in

the middle of the rotor. In vibration readings this will look very like imbalance with a small

axial component. If bowing of the rotor is suspected then the condition can usually be

corrected by “slow rolling” the motor for up to two days to reset the sag.

• An eccentric rotor The variable air gap this produces between the rotor and stator

give a vibration at 2 x line frequency with sidebands at pole pass frequency as well as

sidebands around run speed of pole passing frequency.

32

Vibration Analysis on Rotating Equipment

• Out of magnetic center This is almost always caused by improper fitting of the

bearings. This causes to motor to run out of magnetic center which gives a vibration at run

speed. There will be a phase difference in the vibration between energized and de-energized

Electrical problems in the stator of an induction motor can also result in vibration with a pulsating

amplitude. However, in this case the pulsation is the result of a beat between two separate frequencies of

vibration which are very close together. Common stator related problems which can be expected include:

• Stator windings Phasing problems such as a loose connector in the stator

windings or supply will cause a modulation or sideband of the reciprocal of the number of

pairs of poles x line frequency (e.g. 60 x 1/3 for a 1200 rpm motor) around 2 x line frequency.

A defect in the windings themselves will also cause an unequal magnetic flux around the

motor although is usually very small except in severe cases.

• Windings insulation Gradual degradation of the insulation of the windings will result in a

fall off of the insulation to ground resistance which should be at least 1.5 MΩ for a main drive

motor.

• Imbalanced phases A difference in the supply voltage or power factor of the three

phases will cause a vibration of 2 x line frequency (120 Hz) around rotor bar passing

frequency. If demodulated spectra are used this will show up as very high spikes at 120 Hz

(100 Hz in Europe) in the frequency domain.

In the case of a stator related problem, to produce a vibration whose amplitude pulsates in a cyclic fashion it

is necessary that two frequencies of vibration be present. One of these vibration frequencies may be the

result of some imbalance or misalignment occurring at the running speed of the armature. The other

vibration will probably be an electrical vibration which occurs at the rotating speed of the magnetic field

powering the motor. If any of the above stator problems are encountered a mechanical vibration will occur

at the rotating speed of the magnetic field. Since the mechanical and electrical vibrations are relatively

close in frequency their amplitudes will alternately add together and subtract at a rate equal to the difference

between their frequencies. The result will be noticeable steady pulsation or “beat” of the vibration

amplitude.

Observing the pulsating vibration in time waveform on an oscilloscope or spectrum analyzer can be useful

in identifying the beat frequency characteristics of stator related induction motor problems. The phase

relationship between the two individual vibration frequencies is constantly changing producing a resultant

vibration whose amplitude increases and decreases in a periodic fashion.

Electric motors have inherent vibration due to “torque pulses”. Torque pulses are generated as the rotating

magnetic field of the motor energizes the stator poles. Since each motor pole is essentially energized twice

for each cycle of AC current, the vibration resulting from torque pulses will be two times the line frequency

powering the motor. Thus, if an AC line frequency is 60 Hz or 3600 cpm, torque pulse frequency will be

120 Hz or 7200 cpm. This vibration is rarely troublesome except where extremely low vibration levels are

required, or if the torque pulses should happen to excite a resonance condition in the machine or structure.

Torque pulses have also been known to excite loose rotor bars and loose stator windings at frequencies of

2x, 3x, and even 4x torque pulse frequency. Care should be taken if a significant 7,200 cpm vibration is

seen in velocity vibration readings as this can also be caused by stator distortion brought on by a severe

misalignment or a “soft foot” condition.

Demodulated readings will demodulate the frequencies above the high pass filter or inside the band pass

filter. This range will usually also include the rotor bar pass and slot pass frequencies. A modulation of the

rotor bar or slot passing frequencies by 2 x line frequency is not uncommon and does not necessarily mean

that there is a defect in the motor. It is the author’s experience that inequalities in the power factor at each

of the three phases will cause very large changes in this modulation. If frequencies at 120 Hz (7,200 cpm)

are apparent in your demodulated spectrum, the first thing to check is the individual phase power factor.

33

Vibration Analysis on Rotating Equipment

AC induction motors experience a wide range of mechanical problems common to most machinery such as

misalignment, looseness, bad bearings etc. However, they also have their own unique set of problems that

are often linked to the electro-magnetically generated fields in the stator and the rotor. Frequency analysis

of the motor load current has been consistently proven to be able to detect the presence of broken rotor bars,

end ring resistances and cast rotor blow holes. Each of these three problems give the same effect to the

motor - variations in current draw, as the defective bar cuts the lines of flux. In the 1960’s Aberdeen

University in Scotland carried out a joint research project with Shell Exploration & Production on several

off-shore oil and gas production facilities in the North Sea. The object of the project was to determine a

reliable and repeatable method for the evaluation of broken rotor bars on A.C. induction motors. The

driving force for this project was the large number of motor failures suffered at these production facilities.

The project concluded that the presence of sidebands at the motor slip multiplied by the number of poles

around the line supply frequency indicated not only the presence of broken rotor bars (or equivalent) but

also how many bars were affected. The quantitative analysis of the number of broken rotor bars relies

heavily on the height of the sideband compared to the height of the line frequency spike (generally

expressed in dB Amps) at a fixed speed and a steady load of at least 50% during the measurement. To

identify the separate spikes clearly the spectrum must have a fine line resolution. These measurements are

usually taken with 1600 or 3200 lines with a bandwidth of 80 Hz (for 60 Hz line supplies) or 65 Hz (for 50

Hz line supplies). A 3200 line spectrum with a bandwidth of 80 Hz will take 40 seconds to collect during

which time the load and speed of the motor under test must not vary significantly.

DC Motors

Direct current is created by taking three phase alternating current and converting it through a bank of silicon

controlled rectifiers (SCRs) into direct current. AC is supplied in the United States at 60 cycles per seconds

or 60 Hz (Hertz). Industrial power in the United States is supplied in a three phase 60.Hz format. This

means there are three individual alternating current waveforms being supplied simultaneously at 60 Hz but

120 degrees out of phase. The three waveforms are identical in amplitude and duration; thus, a one second

snapshot of three phase AC will reveal 180 positive and 180 negative amplitude peaks

When AC is rectified to DC these peaks are electronically processed to allow only positive peaks to remain

in the wave form. These peaks, although no longer alternating, create a pulsing which is detectable through

vibration analysis.

When three SCRs (half-wave rectified) are used to convert AC to DC then a pulsing or frequency equal to

the 180 Hz is created in the DC drive System. When six SCRs are used to convert the AC to DC, (full-

wave rectified or High Efficiency System) then a pulsing or frequency of 360 Hz (6 x 60 Hz) is created in

the DC drive system.

Those who are more comfortable using cpm (Cycles Per Minute), rather than Hz, need only multiply the

frequency in Hz by 60 seconds.

Thus 180 Hz x 60 sec. = 10,800 cpm and 360 Hz x 60 sec. = 21,600 cpm.

34

Vibration Analysis on Rotating Equipment

cpm and also has 180 positive and 180 negative amplitude peaks per second.

• Half-wave rectified direct current (3 SCRs) is 3 x 60 Hz @ 180 Hz or 180 Hz x 60

sec. @ 10,800 cpm

• Full-wave rectified direct current ( 6 SCRs ) is 6 x 60 Hz = 360 Hz or 360 Hz x 60

sec. = 21,600 cpm

A half-wave rectified DC drive system will then have a dominant electrically related frequency of 10,800

cpm and a full-wave rectified system will have a dominant electrically related frequency of 21,600 cpm

When these frequencies exist within their respective systems they should be considered normal unless

amplitudes greater than 0.1 in./sec. peak are detected. This usually means that there is an electrical control

problem.

35

Vibration Analysis on Rotating Equipment

Direct current drive systems use rectified alternating current to power an electric motor. This DC source

can be varied through system controls to change the running speed of the motor. This can be controlled

manually or be adjusted automatically by allowing the control system to monitor the motor speed through

the use of a tachometer, self adjusting the power source to achieve the desired speed. The self controlled

systems or closed loop systems use low voltage control components to fire or open the pathways which

allow the full DC power to be supplied to the motor. This allows the motor to run at the desired speed or

36

Vibration Analysis on Rotating Equipment

desired rate of speed change to meet the needs of the driven system. The electrical problems associated

with these systems are:

• AC power supply

• AC to DC rectification components

• DC control components and

• DC motor component failure

The system operates by DC power being supplied to the motor which creates a magnetic field and causes

the motor armature to rotate at a speed proportional to the DC power supplied. As the motor turns, the

tachometer (tach.) also turns which creates a low voltage proportionate to its speed. The tach. low voltage is

compared to a constant or predetermined voltage variance by the comparitor card. The comparitor card,

based on the voltage differential between the constant voltage pot and the tach. low voltage, signals the

firing cards. There is a firing card or order for each SCR which controls the power flow through the SCR.

The SCRs fire or open to create and supply DC power to the motor which either speeds or slows the motor

to the appropriate speed determined by the pot. This is a very basic explanation of a DC motor circuit but

should be sufficient to give an understanding of the concept of how the various components interact.

The firing cards, which control the opening and closing of the SCRs, must perform correctly to allow the

system to function normally. Because there is one firing order per SCR there is a potential for one or more

of the SCRs to perform erratically or not at all if the firing card malfunctions.

Vibration analysis has been used to determine malfunction on this card or SCR’s based on frequencies

which have sub multiples of the DC dominant frequency present. DC power frequencies will always be

constant at 21,600 cpm on a full-wave system or 10,800 cpm on a half-wave system. If one sees a

frequency of 3,600 cpm and frequency separation of 3,600 cpm between existing spectrum peaks or a

frequency of 7,200 cpm and frequency differences of 7,200 cpm then faulty firing cards or SCRs may be the

cause.

37

Vibration Analysis on Rotating Equipment

1

S.C.R. problems

Vibration analysis was requested on a 300 HP DC motor which was critical to a production system. One

day prior to the request for analysis the system would not maintain the expected speed and the motor was

pulsing. A mechanical problem within the gearbox or drive train was initially suspected because the

electrical system appeared to be functioning normally. The motor was uncoupled from the drive train and a

visual inspection performed on the motor when run under a no load condition. The pulsing was still

apparent even under a no load condition, however, the motor did start and run. A motor bearing then

seemed the next logical failure point. The production system was critical to the plant operation, as

previously mentioned, and a decision to recouple the motor and operate until a vibration analysis could be

performed to determine bearing wear or failure seemed the best alternative. The analysis was performed

1

Thanks to Bill Rinehart for his permission to use this data

38

Vibration Analysis on Rotating Equipment

revealing the FFT spectrum in figure 26. The dominant frequencies are 7,200 cpm, 14,400 cpm and 21,600

cpm. These frequencies are related to DC ELECTRICAL problems; not bearings or mechanical defects

When an FFT appears with dominant 1/3 frequencies of the DC full pulse frequency (21,600 for a full

wave, six SCR system), then suspect firing card or SCR problems in the DC control.

This DC control system uses three firing cards to control six SCRs which is typical of many DC control

systems. When one firing card is not functioning then 1/3 of the power is lost. Dominant frequencies of

7,200 cpm and 14,400 cpm or frequencies equivalent to multiples of 1/3 of 21,600 cpm are representative

of this situation. The firing cards were examined and a loose connection on one card was found and

repaired. The FFT spectrum in figure 27 was taken after the repair. The 7,200 cpm and 14,400cpm

frequencies are now gone. The 21,600 cpm frequency is the normal frequency of a full wave system and

should be present.

Half-wave rectified AC power sources will tend to have 1/3 multiples of 10,800 cpm or frequency

separations of 3,600 cpm. Full-wave rectified AC power sources can also have frequency separation of

3,600 cpm if the system has:

• One firing card for each SCR, and one card is out

• A three card system and one card is partially disabled

• One SCR is not functioning

The FFT spectra directly above and below also show the difference between the firing card frequency

amplitudes of a motor that is uncoupled and running under no-load (below) and the same motor coupled and

running under a load (above). Although the frequencies are the same in each spectrum the amplitudes are

considerably load dependent. The amplitude at 7,200 cpm on the spectrum below is only 0.00847 in/sec

peak while the amplitude of the spectrum above at 7,200 cpm is 0.3037 in/sec peak, yet each spectrum

represents the same firing card malfunction.

Another possible reason for seeing these 1/3 multiples would be if one phase of the AC power source was

not present. This would affect one-third of the system power and virtually render one bank of SCRs

inactive. A simple voltage test of the three incoming AC phases should confirm this situation if present.

39

Vibration Analysis on Rotating Equipment

The comparitor card is another low voltage control component which is responsible for determining the

difference between the system actual speed and the set or predetermined speed performance. When this

component malfunctions it has been observed that there are side-bands present around DC frequencies.

These side-bands are not of a particular set frequency but are always equally spaced from the DC

frequencies. it has also been observed that these side-bands will grow or diminish as the motor RPM is

varied, however, they will remain equally spaced. It has not yet been determined if these side-bands are

related to the RPM fluctuation or hunting which often accompanies comparitor card problems or if the

constant collapsing and regenerating of the magnetic field of a system that is hunting is the cause. The side-

bands do, however, exist regardless of the cause and should be considered a warning of this component

failure. These side-bands may occur at small cpm increments as shown below and may require a high

resolution FFT to differentiate them from the dominant frequencies.

To resolve side bands related to comparitor card malfunctions a FFT spectrum of 3200 lines of resolution at

Bandwidth of 24,000 cpm is suggested.

Another possible reason to see these FFT characteristics could be a faulty or malfunctioning tachometer

which would corrupt the voltage going to the card. Testing the voltage output of the tach. should confirm

this situation.

40

Vibration Analysis on Rotating Equipment

0.0872

5000 16,000

Vibration analysis was requested on a 125 HP DC motor which was thought to be vibrating heavily. The

motor had been uncoupled from the belt to see if the motor vibrated when running solo.

The first set of spectra was collected under these conditions and revealed the above spectrum. A RPM

check using a digital tach. also revealed that the motor was fluctuating or hunting approximately 30 cpm at

an RPM of 1440. The tach. voltage was then checked but seemed to be consistent with the operating speed

and fluctuations. Based on the side-bands a recommendation was made to change the comparitor card (also

called the control card). There was not a spare card available at this time but there were spare firing cards.

The firing cards were changed but did not solve the problem. A spare comparitor card was eventually

located and the replacement accomplished. The speed fluctuations stopped and another set of spectra

collected revealing the spectrum below.

DC motors are different from AC motors because of their power supply which requires different

components. The most obvious of these is the tachometer which extends, usually, from the back of the

motor. These units usually have small bearings which can be monitored in the same manner as any bearing.

Caution: Never Place A Magnetic Accelerometer Mount On A Tachometer

These devices use magnets to generate the voltage which the control system monitors to determine the

motor speed. Placing a powerful magnet on or near the tach may alter or destroy the voltage output causing

the motor to literally speed up until it destroys itself.

41

Vibration Analysis on Rotating Equipment

The commutator is the device which transfers the DC power to the motor armature. Brushes, usually made

of a carbon alloy, ride against the commutator and supply the DC power to the commutator. It has been

observed that as these brushes wear, readings at one times the motor RPM will rise in amplitude, When the

brushes arc it has been observed that these one times RPM readings will increase dramatically, sometimes

reaching 0.3 ins./sec. peak or even higher in extreme cases. Another frequency associated with the

commutator is the number of slots on the commutator times the motor RPM. Although the significance of

this frequency has not been specifically related to a problem, the brush wear would, once again, be suspect.

Knowing the exact RPM of the variable speed system which is being analyzed is critical to determining

which system component is creating the vibration frequency of interest. Never assume that the RPM of the

system is what the panel display reads, these displays are seldom accurate. The use of either a good strobe

light, or a digital tachometer will give an accurate RPM. Ensure that your data always has the RPM stored

with it to prevent mistakes when looking at your data for analysis.

42

Vibration Analysis on Rotating Equipment

When determining a frequency relationship to a machine component, be sure that the resolution of the

spectrum is enough to see the difference between closely positioned frequencies. If necessary take multiple

spectra at different resolutions to capture and differentiate all possible frequencies. Remember,

mechanically related problems will follow the speed increases or decreases proportionately and electrical

frequencies will remain constant. This is a powerful analysis tool when determining a mechanical or

electrical frequency.

The following spectra will verify the importance of correct RPM and the ability to determine mechanical

frequencies by noticing how the frequencies change related to the RPM.

Vibration analysis was requested on a 5 HP DC motor which was causing problems on a production line.

The motor is one of four motors which must work in sequence for the line to run efficiently. The spectra

was collected revealing the FFT in figure 33, which had what was initially suspected to be DC electrical

frequencies with a 3,600 cpm separation. A quick reference reading of G/SE (high frequency pulse

detection) was taken and registered a higher than normal value, this is not characteristic of an electrical

problem.

The motor RPM was reduced by 25% from 1004 RPM to 756 RPM and new spectra collected. The

frequency separation was now 2,700 cpm, a 25% difference from 3,600 cpm. The frequency separation

followed the RPM proportionately indicating a mechanical problem.

The motor had a SKF 6208 bearing which has a 3.606 BPFO multiplier (1004 RPM x 3.606 = 3620 cpm).

What was assumed to be a 3600 cpm separation was due to a bearing outer race defect.

43

Vibration Analysis on Rotating Equipment

44

Vibration Analysis on Rotating Equipment

Rotating Equipment

Imbalance

Let us consider a fan impeller of 50 kg weight which is running at 1000 rpm. Let us imagine that this

impeller has an out of balance corresponding to 0.5 kg at 0.3 meter from the center (we will use S.I. units

here to make the math easier).

Newton says that:

F = (m ω2 r)/9.81

using S.I. units F is force in kg.F

m is the mass in kg

ω is the rotational speed in radians per second

and r is the radius at which the force (the out of balance) is acting

= [0.5 x (1000/60⋅2π)2 x 0.3]/9.81

= [0.5 x 104.722 x 0.3]/9.81

= [0.5 x 10,966x 0.3]/9.81

= 167.68 kgF (369 lb.F)

In other words we have added over an eighth of a ton to the apparent weight of the impeller. Let us see

what happens if the impeller is running at 2000 rpm.

= [0.5 x (2000/60.2π)2 x 0.3]/9.81

= 670.71 kgF. (1476 lb.F)

By doubling the speed to 2000 rpm we have quadrupled our out of balance forces to almost three-quarters

of a ton force.

A deep groove Conrad bearing for a shaft of about 3 inches diameter could be a MRC 215-S, which is in the

MRC mid range of bearings. This bearing has a static load rating of 23,680 lb. at 1,000 rpm as calculated

using the 3AFBMA method of evaluating load ratings. The speed factor for calculating radial load at 2,000

rpm is 0.7937 so our load is now

3,680 x 0.7937 = 2,920 lb.

By increasing the speed to 2,000 rpm and allowing an imbalance force of about 1 lb at about 1 foot away

from center we have halved the effective load carrying capacity of the bearing.

2

Data taken from TRW service catalog Form 382-14

3

Anti Friction Bearing Manufacturers Association

45

Vibration Analysis on Rotating Equipment

46

Vibration Analysis on Rotating Equipment

47

Vibration Analysis on Rotating Equipment

48

Vibration Analysis on Rotating Equipment

49

Vibration Analysis on Rotating Equipment

To confirm a suspected imbalance check the time waveform. An rotor imbalance will give a sine wave in

the velocity time signal with a period equal to the time it takes for one revolution of the rotor. Phase

readings will be steady .

Out of balance may occur in more than one plane. Most spectrum analyzers have a two plane balancing

function built into the software. Multiple plane balancing (more than two planes) is usually only necessary

on complex multiple disk rotors such as turbines which operate above their critical speed.

50

Vibration Analysis on Rotating Equipment

Misalignment

51

Vibration Analysis on Rotating Equipment

52

Vibration Analysis on Rotating Equipment

53

Vibration Analysis on Rotating Equipment

Looseness

54

Vibration Analysis on Rotating Equipment

55

Vibration Analysis on Rotating Equipment

56

Vibration Analysis on Rotating Equipment

57

Vibration Analysis on Rotating Equipment

58

Vibration Analysis on Rotating Equipment

Fans and blowers will inherently have some vibration due to aerodynamic forces. This vibration results

from the fan blades striking the air and will occur at a frequency equal to the number of fan blades times fan

R.PM. Normally, the amplitudes of vibration resulting from aerodynamic forces will be low and no cause

for concern. When excessive vibration at the aerodynamic frequency is encountered a common cause is

resonance of some part of the machine or structure and the checks for resonance, described near the end of

the manual should be carried out to determine the resonant part. If it is confirmed that a condition of

excessive aerodynamic vibration is not due to resonance, the fan should be checked carefully for

obstructions that may disturb the smooth flow of air through the fan. For example, high amplitudes of

aerodynamic vibration are sometimes encountered on cooling tower fans. Many of these fans consist of a

drive motor, mounted outside the fan venturi, coupled to the fan gearbox by means of a long torque tube or

drive shaft. The torque tube can act as an obstruction to the smooth flow of air through the fan and an

aerodynamic pulsation is generated each time a fan blade passes over the torque tube. The result is often

excessive vibration at the aerodynamic frequency and may require that the distance between the blade path

and torque tube be increased to minimize these aerodynamic pulsations.

On centrifugal fans, excessive vibration at the aerodynamic frequency can sometimes result if the fan rotor

is positioned eccentrically in -the f an housing. Therefore this should be checked in the event that the

problem cannot be traced to resonance.

Vibration due to aerodynamic forces can also occur at a frequency equal to 1 x fan R.PM and will appear

similar to normal imbalance. This aerodynamic imbalance will result if the fan blades do not have the same

track or pitch. If the fan operates under a constant aerodynamic load, the force of aerodynamic imbalance

can be compensated by following normal balancing procedures. However, it often occurs that changing the

fan load will produce a corresponding change in the vibration at 1 x R.PM.

For example, a centrifugal fan was balanced with the access doors in the fan housing removed to simplify

the addition and movements of trial weights. After the fan had been satisfactorily balanced, the access

doors were replaced. Unfortunately, however, this produced a significant change in aerodynamic conditions

and the result was a significant increase in vibration. In this case, it was necessary to balance the fan

operating under its normal aerodynamic conditions. If a fan must operate smoothly over a broad range of

aerodynamic loads it may be necessary to check and correct for significant variations in blade track or pitch

before this -can be achieved.

59

Vibration Analysis on Rotating Equipment

Aerodynamic cross coupling is a problem occasionally encountered on turbines, centrifugal compressors

and fans operating above the first rotor critical and generally results from eccentric rotation of the rotor

caused by rotor bow or deflection. In the case of a centrifugal compressor the layer of air or other gas being

compressed will have a rotating speed less than that of the rotor, similar to the rotating oil whirl in a plane

bearing. If the rotor is bowed slightly, the layer of rotating gas between the rotor and the machine housing

will produce a torque reaction on the rotor causing the rotor to whirl at the rotating speed of the gas layer.

The frequency at which this whirl occurs can vary from one machine to the next-

The vibration may have the same frequency characteristics of oil whirl and hysteresis whirl where the

lowest natural frequency of the rotor bearing system is excited. In most cases, the vibration frequency will

be less than the rotating frequency. Cases have been reported where sub-multiples of 0.5, 0.33 or perhaps

0.25 times the rpm of the rotor have been excited.

Since rotor whirl generated by aerodynamic cross coupling is excited by -the compressed air or gas, it seems

logical that the condition would be affected by machinery load. In general, the machine will be more likely

to experience this condition under heavily loaded conditions and changing the load of the machine to

determine its effect on the vibration can be useful in -diagnosing this problem.

Surging

Surging is a rather common problem encountered on high speed centrifugal and axial flow compressors and

occurs when the compressor is operated outside designed limits. Typically a compressor is designed to

deliver air or other gases over a specified mass flow range and at a specified pressure ratio. These

requirements are met by selecting rotor speed, number of compressor stages, blade configurations and other

factors. The manufacturer of the compressor can supply performance characteristics curves showing the

range of stable operation in terms of pressure ratio, mass flow and rpm. Attempting to operate the unit

outside the design range can result in excessive vibration and damage to the machine.

The problem of surge occurs when, for a particular operating speed, the delivery pressure to inlet pressure

ratio is too high or if mass flow is too low relative to design conditions. When this occurs, a reversal of gas

60

Vibration Analysis on Rotating Equipment

flow in the compressor will result. In the initial stages of surge, the flow reversal may only occur in the

boundary layers of the rotor blades (a rotating stall) however, at full surge the gas flow reverses its direction

and flows from the discharge to the inlet. Rotating stall shows itself as a vibration at approximately 35% of

run speed but is dependent on the physical configuration of the compressor.

The vibration characteristics resulting from compressor surging can vary depending on the extent of the

problem. In cases of mild surge, a noticeable increase in the vibration at blade passing frequency can

usually be detected. This frequency is the product of the number of rotor blades times the rpm of the rotor.

In other cases multiples of blade passing frequency may also be detected. When a full surge condition is

encountered the result may be a high amplitude of random, erratic vibration usually covering a rather broad

frequency range. This is caused by the turbulent flow within the compressor exciting the various natural

frequencies of the rotor wheels, rotor blades, diffuser blades, casing, shaft and other components. Of

course, if this condition is allowed to continue, extensive damage to the compressor c result.

The problem of 'choking or 'stone-walling' in a compressor is essentially the opposite of surging but again is

the result of attempting to operate the unit outside design parameters. Choking occurs when discharge

pressures are too low. When discharge pressures are low, velocities are high and when flow velocity in the

diffuser section approaches Mach 1 a turbulent or circulating flow between the blades will occur which has

the effect of blocking the flow of gas. When this occurs a noticeable drop in efficiency and pressure ratio

can be seen, along with an increase in vibration due to the turbulent flow within the compressor. The

vibration characteristics of choking are essentially the same as those encountered during surging. A check

of other operating parameters such as pressure, temperature, flow etc. should be undertaken to distinguish

between the two.

61

Vibration Analysis on Rotating Equipment

Bearing Failures

Arguably the most common type of failure on rotating equipment are failures arising from the collapse of

the bearings.

Let us consider the lubrication of rolling element bearings.

Let us consider the lubrication of a rolling element bearing. As the roller rotates the pressure point is very

small so the pressure loading is very high. So high, in fact, that the contact point of the roller and the race

(the contact ellipse) becomes elastically deformed, trapping a very small amount of the lubricant into a

“wedge”. The lubricant wedge itself, is very small but is large enough to keep the roller physically

separated from the race by a small distance (in the micron range). THERE IS NO METAL TO METAL

CONTACT IN A PROPERLY LUBRICATED BEARING RUNNING AT NORMAL LOAD.

As the lubrication mechanism starts to break down the lubricant loses its ability to separate the roller from

the race. This may be due to a deterioration of the lubricity, an increase in bearing load, overheating of the

bearing or a fatigue failure of an old bearing at the area most prone to elastic deformation.

• Stage 1 Lubrication problems

• Stage 2 Marking of the raceways and / or rolling elements

• Stage 3 Spalling of the raceways and / or rolling elements

• Stage 4 Collapse of the bearing

We will discuss each stage of the failure and discover how each stage may be identified with predictive

maintenance techniques.

62

Vibration Analysis on Rotating Equipment

Lubrication problems in the case of grease or oil systems is not necessarily the physical loss of lubricant but

the loss of the oil or grease’s properties. Since the life of a grease lubricant is strongly temperature

dependent and since grease loses half of its life for every 20°F rise in temperature, this can be seen to have a

very significant effect on bearing health. In order to avoid the temperature effect due to over packing,

bearing manufacturers usually suggest that bearings are packed with grease to between 15% to 20% of the

bearing's free volume. We know that over packing is the most common cause of raised bearing

temperature, which leads to reduction in grease life and eventual failure. Under-packed bearings, or

bearings which have lost grease due to physical migration, may generate high bearing temperatures when

running at high speeds.

Low speed, starved bearings usually wear into a condition of excessive looseness and fail without

appreciable temperature increases. While this is fairly common on lightly loaded motor bearings, press

main drive motors will usually fail catastrophically.

Figure 63 Loss of Lubricant - Ball Bearing Inner Race Courtesy of the Barden

Corporation

At this stage of the bearing failure the rolling elements have metal to metal contact onto the raceways

because the lubricant is no longer supporting the rolling element via a lubricant wedge. Because of this, the

metal to metal impacts excite the resonant frequencies of the bearing - just like hitting a bell with a hammer.

For a rolling element bearing these frequencies will be in the range of 1 to 4 kHz. The metal to metal

contacts also generate 4ultra sonic frequencies at between 30 and 50 kHz. Both of these frequency ranges

do not experience any modulation at this stage as the metal to metal contacts are irregular.

4

Ultra sonic - above the human audible range. Set at above 20 kHz for industrial applications.

63

Vibration Analysis on Rotating Equipment

This the early stage of fatigue. This fatigue may be initiated on the surface or beneath the surface. Surface

fatigue is usually caused by scratches on races, balls, or rollers, abrasive contamination, or brinelling.

These marks produce “stress raisers”, a point on the bearing surface that experiences abnormally high stress

due to the physical conditions at that spot. Simply, a given load over a given area produces stress. If a

crack or contaminant is found at that location, the load is distributed over a different (often smaller) area

and therefore greatly increases stress at that point. This phenomenon limits the number of cycles a bearing

can survive. These raised stress areas provide a start point for micro-crack formation that leads eventually

to pitting, spalling, and wear.

Subsurface fatigue is usually caused by voids, foreign matter or coarse carbides introduced into the material

at the time of formation. These material anomalies again provide for a point of crack formation if they fall

within a high stress area. and once a crack is formed beneath the surface, it works its way outward and

eventually develops into a spall.

As the crack reaches the surface it creates a small void into which the lubricant wedge collapses.

Remember that the lubricant wedge is microscopically small so even a tiny crack in the material can cause

the roller to impact heavily onto the race. As each roller passes the void it impacts onto the race. As with

the first stage, the impacts generate vibration at resonant and ultra sonic frequencies, with the difference,

however, that these frequencies are now modulated by the rate at which the rollers hit the defect. In other

words the resonant frequencies are excited every impact, giving a vibration like the pattern below.

Notice the “roller coaster” in Figure 64 at just below 2 kHz. This is a bearing resonant frequency and the

rise and fall in amplitude coincides with the impacts from the rollers on the outer race defect (in this case).

At this stage of the failure there is no appreciable rise in temperature and the velocity vibration at the defect

frequencies is insignificant.

64

Vibration Analysis on Rotating Equipment

Fatigue failure or spalling results from mechanical materialogical failure of the bearing. Literally a stress

related failure of the material which results from cyclic stresses due to operation at high loads. As the

rollers repeatedly impact onto the small crack, the very small area produces very high point loading so the

material starts to flake off around the crack. As the crack propagates, more and more material is removed

until the crack becomes a visible cavity or spall.

At this stage of the bearing failure, the velocity vibration becomes apparent at the defect frequencies and

harmonics, possible also with sidebands of run speed and / or cage frequency. The resonant vibration has

also increased in amplitude at the defect frequencies along with a general rise in floor level. Temperature

will be elevated above normal but not significantly, particularly not if there is any air movement around the

surface of the bearing or bearing housing.

65

Vibration Analysis on Rotating Equipment

With time both surface and subsurface fatigue flaws spread over the active bearing surfaces causing bearing

wear, growth in spalls, and eventual machine failure. The metal contaminants or wear particles removed

from the bearing during spalling are either washed out with the oil, in oil-lubricated bearings, or are trapped

in the bearing, as is common in sealed and grease-packed bearings. In these latter bearings, continuous

recirculation of the particles causes progressively higher wear to the point where either the bearing becomes

excessively loose and fails to support the load suitably or the induced damage leads to failure.

Company

66

Vibration Analysis on Rotating Equipment

It is important to understand that as a bearing with a damaged surface rotates, the regularity, or frequency

with which the roller or ball impacts on the defect indicates potential failure and allows us to determine the

type of damage that exists. A number of characteristic frequencies are generated by a damaged bearing ,

and are known as:-

• Cage or fault train frequency = FTF

• Ball pass with respect to the outer race = BPFO

• Ball pass with respect to the inner race = BPFI

• Ball rolling about its own axes = BSF

• Shaft frequency of rotation = RPM

FTF = RPM/2 [1 - (BD/PD) cos θ] (rotating inner race)

approximated by RPM * 0.45 (rotating inner race)

or RPM * 0.55 (rotating outer race)

BPFO = n * RPM/2 [1 - (BD/PD) cos θ]

approximated by RPM * O.4* n

BPFI = n * RPM/2 [1 + (BD/PD) cos θ]

approximated by RPM * O.6 * n

BSF = (RPM * PD)/2BD * [1 - (BD/PD)2 COS2 θ]

(no valid approximation)

PD = pitch diameter

θ = contact angle

n = number of balls (rollers).

When bearing geometry is not known but the number of balls or rollers can be counted or estimated, it is

suggested that the approximate equations be used to establish the bearing frequencies of interest. All the

equations listed above show a direct dependence of the calculated frequency on the frequency of rotation..

The following figure illustrates the bearing geometry used in the above equations

67

Vibration Analysis on Rotating Equipment

68

Vibration Analysis on Rotating Equipment

Figure 69 Waterfall of early damage to a motor bearing collected every 1.5 hrs over

14 days

The plot above shows a waterfall display of acceleration vibration up to 2 kHz over a period of almost two

weeks on the drive end bearing of a drive motor. Notice that the vibration at about 1.5 kHz has risen

steadily. This frequency is typical of the bearing resonance. Notice also that there appear to be small

sidebands around the resonance frequency which are at the bearing outer race defect frequency. The

presence of a bearing defect frequency excited resonance does not give justification by itself to change the

motor bearings. We must also wait until the defect frequencies with sidebands and/or harmonics show up in

the velocity spectrum.

The severity chart in figure 69 is applicable for motors and fans running between 500 to 3600 rpm. The

envelope (or demodulation) amplitudes are quoted in dBG re 0.001G. As with any severity chart it is

important that the vibration analysts use their own judgment and experience when deciding whether or not

to change the bearing. For machines that are outside the speed ranges quoted the amplitudes will be lower

for slower machines and higher for faster machines. Velocity readings are heavily affected by the mass of

the machine so care should be taken when assessing very small or very large motors.

Don’t forget that roller bearings will stand more impactive forces than ball bearings. A roller bearing

should not be allowed to operate at above about 12 G (5true peak) and a ball bearing should not be allowed

to operate at above about 7 G (true peak).

5

True peak is defined as the peak seen in the waveform - not the derived peak seen in the frequency

spectrum.

69

Vibration Analysis on Rotating Equipment

The figure above was taken (as was most of these examples) from the drive end bearing of a drive motor.

Notice in the spectrum that there are no significant spikes but the spectrum is raised up from the floor - this

70

Vibration Analysis on Rotating Equipment

is a raised carpet level. A certain amount of rise from the floor is normal but when you find that your alarm

limits (based on the baseline) have been exceeded in the demodulated spectrum and there are no significant

spikes, the chances are that the bearing is starting to suffer lubrication problems.

In the case of a main drive motor the bearing is almost certainly grease lubricated. The grease lubricated

bearing may be of shielded (or sealed) construction or non-shielded construction. If the bearing is fitted

with seals then it is sometimes possible to force some grease pass the seal with a grease gun, but usually you

will just have to watch the bearing deteriorate and change the bearings before the damage gets so bad as to

cause secondary damage.

If the bearing is non-shielded then the bearing should be lubricated following the bearing and motor

manufacturers instructions. If a plug is fitted opposite the grease fitting make sure that you remove it.

Injecting too much grease into the bearing cavity will cause pressurization of the cavity and the grease will

force its way past the bearing into the motor windings

As you inject the grease into the bearing have a spectrum analyzer attached to an accelerometer on the

bearing housing and watch the vibration levels. A simple rule of thumb for bearing condition is that if the

vibration goes down and stays down, the bearing only had a lubrication problem and you have just fixed it.

If the vibration level goes down but rises again then the bearing is damaged - the sooner the level rises again

the worse condition the bearing is in (from a few of days to several minutes for a very bad bearing).

71

Vibration Analysis on Rotating Equipment

0.872

bearing

Figures 71 and 72 show two examples of rolling element bearings which have suffered some marking of the

races (both of these examples show marking of the stationary outer race). Note that the fundamental

frequency has several harmonics but it is the fundamental frequency which will coincide with the generated

bearing defect frequency and help us in our analysis. At this stage of the bearing deterioration it is

sometimes still possible to save the bearing with additional lubrication. At the very least you will extend

the life of the bearing.

1.000

1.000

72

Vibration Analysis on Rotating Equipment

2.500

2.500

The two time waveform spectra above show the individual impacts caused by the rollers impacting onto the

damaged raceways. Generally, a ball bearing should not be allowed to run with impacts of more than 0-7 G

peak in the time domain and a cylindrical roller bearing should be running less than 12 G. If these figures

are exceeded then the bearing is almost certainly severely damaged.

Looking at the time interval between the peaks in the time domain, we can correlate this time difference

with the frequencies seen in the frequency spectrum.

0.0463

73

Vibration Analysis on Rotating Equipment

As the bearing deteriorates the bearing defect frequencies start to show up in the velocity spectrum. Figure

75 shows a velocity spectrum from a spalled bearing with multiple harmonics of the outer race defect

frequency. As a general rule of thumb, if you see the same defect frequencies in the demodulated

acceleration and the velocity spectra, irrespective of amplitude, then that bearing is spalled.

74

Vibration Analysis on Rotating Equipment

Balancing

Imbalance has been named the most common cause of vibration in our machines. Since this is

true it would be to our advantage to be able to correct this condition as easily as possible. Before

we can balance a part with the Vibration Analyzer, certain conditions must be satisfied. The

vibration must be due to imbalance, we must be able to make the weight corrections on the rotor,

and we must have a shaft angle reference (such as being able to observe the rotor with the strobe

light as used in this paper, or other type of key phasor) for phase measurements.

In-place Balancing

The conditions which must be satisfied so you can balance can almost always be met with the

imbalanced part mounted in its own bearings, operating as it normally does. The process of

balancing a part without taking it out of the machine is called In-place Balancing. The following

pages deal directly with this application. In-place Balancing eliminates the need to disassemble

the machine, transport the part to a balancing machine, balance under artificial conditions, and

assures smooth operation of the machine when you are done.

In-place Balancing is a straight forward process which involves following a few simple rules.

However, before we discuss balancing we should first understand imbalance, where it comes

from and what must be done to correct it.

Imbalance is the unequal distribution of the weight of a part about its rotating centreline. Illustrated

are some reasons for an unequal weight distribution. The blow hole in the casting, the thicker

web, the eccentric hole location and eccentric machining of the pulley groove all add up to more

weight on one side of the rotating centreline than the other. The faults shown are exaggerated but

could easily exist in almost every rotating part manufactured.

75

Vibration Analysis on Rotating Equipment

More important as a source of imbalance is the stack up of tolerances possible when assembling

rotating parts. The example illustrated is a very common one. The hole in the pulley is necessarily

larger than the shaft diameter and when a key or set screw is attached the take-up in clearance

shifts the weight of the pulley to one side of the rotating centreline of the shaft. Thus the assembly

is out of balance.

All of the combinations of unequal weight distribution can be considered to be concentrated at one

point called the heavy spot. This heavy spot represents the accumulated results of all of the

imbalance of the pulley. An equal amount of weight at the same radius but opposite the heavy

spot will balance the rotor.

76

Vibration Analysis on Rotating Equipment

The units for measuring imbalance are generally ounce-inches, gram-inches or, more commonly

now, gram-centimetre. One ounce-inch of imbalance would be a heavy spot of one ounce in

weight one inch from the rotating centreline. Three ounces located three inches from the centre

would be nine ounce- inches.

The imbalance of a part may also be given in terms of the distance from the rotating centreline to

the mass centre of the part. The mass centre of a part is more commonly known as the centre of

gravity or C.G. The centre of gravity or mass centre is the point about which all of the weight of a

body is equally distributed.

The example illustrated shows a condition of mass centre displacement of two thousandths of an

inch. Such a condition could very easily be duplicated by installing a 100- ounce pulley on a shaft

with a two-thousandths clearance between the hole and the shaft. The resulting imbalance is two-

tenths of an ounce-inch.

77

Vibration Analysis on Rotating Equipment

The effect of imbalance is that it generates a force which causes the part and the supporting

bearing and structure to vibrate. The size of the force generated depends on the speed of rotation

and the amount of imbalance. In the illustration the part has an imbalance represented by a heavy

spot located some distance from the centre of rotation. If the heavy spot and distance are in

ounces and inches, the force generated can be computed by the formula

RPM, the rotating speed of the part and

ounce-inch the amount of imbalance.

1.77 is a constant required to make the formula dimensionally correct.

F = 1/16(RPM/1000)2 gram-inches.

The amount of force generated by one ounce-inch of imbalance at high speeds is surprising and

explains why good balance becomes absolutely necessary for high speed machines. For

instance, at 3600 RPM nearly 23 pounds of force is generated for each ounce-inch of imbalance.

Balancing is the process by which we learn the amount and position of the heavy spot so we can

either add an equal amount to the opposite side of the rotor or remove weight at the heavy spot.

We have learned that the more imbalance we have the greater the force, therefore, the greater

the vibration. So we will use the amount of vibration to indicate how much imbalance we have. We

will use the position of the reference mark as seen by the strobe light to tell us the location of the

imbalance.

78

Vibration Analysis on Rotating Equipment

If an imbalance weight is added to a perfectly balanced part it will have a vibration at a frequency

equal to the rotating speed of the part. The part will vibrate a certain number of mils and the

reference mark will appear at some definite position when the strobe flashes. The amount of

vibration is 5 mils and the reference mark appears at 12: 00 o'clock. If we double the amount of

the imbalance weight without changing its position the vibration will increase to 10 mils and the

reference mark will still appear at 12: 00 o'clock when the strobe light flashes. This experiment

shows .it is correct to use the amount of vibration to tell us how much imbalance is in the part.

Note that the reference mark appeared at the same place each time.

What happens when the position of the heavy spot or imbalance weight is changed? If we move

the heavy spot 30 degrees clockwise from where it was before we will see the same amount of

vibration, 10 mils, but the reference mark will now appear at 11: 00 o'clock or 30 degrees

counterclockwise from where it was before. If we move the weight 30 degrees counterclockwise

from its first position we will still see the same amount of vibration, 10 mils, but the reference mark

will appear at 1: 00 o'clock or 30 degrees clockwise from it’s first position of 12: 00 o'clock.

The experiment tells us two things:

1. The amount of vibration is proportional to the amount of the imbalance.

2. The reference mark shifts in the opposite direction to a shift of the heavy spot. The angle the

reference mark shifts is equal to the angle of shift of the heavy spot.

How can we use this information to balance?

How to Balance - Single Plane

We said earlier that balancing is the process by which we learn the size and position of the heavy

spot so we can either add an equal amount to the opposite side of the rotor or remove weight at

the heavy spot. At the start of a balancing operation we do not have the least idea how large the

heavy spot is nor do we know where on the part it is located. The imbalance in the part at the start

is called the original imbalance and the vibration amount and phase which represents that

imbalance is called our original reading. We change the original imbalance by adding a trial weight

to the part. The new total imbalance in the part will be represented by a new amount and phase of

vibration. The change caused by the trial weight can be used to learn the size and location of the

original imbalance, or where the trial weight must be placed to be opposite the original imbalance

and how large the trial weight must be to be equal but opposite the original heavy spot.

Starting with the part out of balance, we see 1.8 mils of vibration and a reference mark position at

12:40 o'clock. What we must do is find a weight and position for that weight which would oppose

the original imbalance. Let's start by adding a trial weight. Three things can happen.

First, if we are lucky we might add the trial weight right on the heavy spot. If we do, the vibration

will in- crease and the reference mark will appear in the same position it did on the original run. To

balance the part all we have to do is move the trial weight directly opposite its first position and

adjust the amount until we reach a satisfactory balance.

The second thing that could happen is we could add the trial weight in exactly the right place

opposite the heavy spot. If the trial weight were smaller than the imbalance, then the heavy spot

would still be the heavy spot but we would see a decrease in vibration and the reference mark

would be in the same place. To balance all we would have to do would be to add more weight until

we reached a satisfactory vibration level. If the trial weight were larger than the imbalance then it’s

position would now be the heavy spot and the reference mark would shift 180° or directly opposite

where it was originally. To balance we would reduce the trial weight amount until we reached a

satisfactory vibration level.

The third thing that can happen is the usual one where the trial weight is added neither at the

heavy spot or opposite it. When this happens the reference mark shifts to a new position and the

vibration displacement changes. By what angle should the trial weight be shifted, in which

direction, and should its amount be in- creased or decreased to be equal but opposite the original

imbalance?

79

Vibration Analysis on Rotating Equipment

This question can best be answered by making what we call a vector diagram. A vector is simply

a line whose length represents the size of the imbalance and whose direction represents the

angle of the imbalance. When a trial weight is added to the part we actually add to the original

imbalance in the part and change the position of the total imbalance to some new position

between the two. We see this as a new vibration displacement and reference mark position or

phase. Our original imbalance was represented by 1.8 mils and a phase of 12:40 o'clock. After

adding a trial weight the imbalance due to both the original plus the trial weight is represented by

2.30 mils and a phase of 2.30 o'clock.

Using polar coordinate graph paper we plot our original imbalance vector. We draw a line from the

centre at the same angle we saw our reference mark, 12: 40 o'clock. Each hour is equal to 30°.

We select a convenient scale for the length of the line such as 1 mil equals 2 divisions. We'll call

this vector “O” for the original imbalance. Next we draw the original plus the trial weight imbalance

vector “O + T'' to the same scale. It is drawn at the new angle for the original plus the trial weight

imbalance phase 2: 30 o'clock. This is all the information needed to solve for the effect of the trial

weight alone.

To solve for the trial weight vector we must complete the vector parallelogram. To complete the

vector parallelogram we first connect the end of the original imbalance vector to the end of the

original plus trial weight imbalance vector, line A. Parallel to line A we draw a line outward from the

centre of the graph, line B. Then parallel to the original imbalance vector and from the end of the

original plus the trial weight vector, O + T, we draw another, line C, until it crosses line B. Line B

then represents the vector for the trial weight. The length of line B represents the amount of the

trial weight. The direction in which the trial weight acts with respect to the original imbalance is

represented by the direction of line B. In order to balance the part we want the trial weight to be

equal and opposite to the original imbalance as shown by the dotted line. We can see that the trial

weight must be moved by an angle equal to the angle between the trial weight vector and the

place where we want it to be. The direction we must move the weight, however, is not clockwise

as shown here but is counterclockwise. Remember our experiment where we learned that the

reference mark moves in the opposite direction to that of the heavy spot. Therefore, we should

adopt this rule. When the reference mark after adding a trial weight moves clockwise from its

80

Vibration Analysis on Rotating Equipment

original position then 'we must more the trial weight counterclockwise or shift the trial weight

opposite the shift of the reference mark.

In the example shown the trial weight is larger than it needs to be. We know this because the trial

vector, line B, is longer than the “O'' vector. To determine the correct amount of the trial weight we

simply multiply the trial weight by the length of the original imbalance vector and then divide by the

length of the trial weight vector.

By inspecting our vector parallelogram we can see that the line A connecting the ends of the

original imbalance vector and the original plus trial weight imbalance vector is exactly the same

length as the trial weight vector line B. Furthermore, the angle between the original imbalance

vector and line A, is the same as the angle between the trial weight vector, line B, and its position

to get the desired effect to balance the part, the dotted line. Therefore, to solve the vector problem

we actually need only the three lines, O, O + T, and A.

Our vector method of balancing then becomes quite simple. We draw the “O'' vector then the O +

T vector, connect the ends of the two, and this becomes the “T'' vector or Trial Weight Vector, or

line A in the more complex vector diagram shown above. To solve this problem measure the

lengths of line “O'' and line ”T'', measure the angle between “O'' and “T''. The correct balance

weight is equal to the trial weight x O/T. The balance weight should be added at an angle equal to

the angle between “0'' and “T'' away from where the trial weight was added. The direction of this

angle is opposite from the shift of the reference mark on the trial run. By following these

instructions carefully the part should now be balanced by moving the trial weight 75 degrees

counterclockwise and reduce its size. However, very small errors in measuring the phase angle, in

shifting the weight, and adjusting the weight to the proper amount can result in some remaining

vibration still due to imbalance.

81

Vibration Analysis on Rotating Equipment

To make further corrections we simply observe the new amount and phase of vibration, .5 mils at

9:00 o'clock. We must plot this reading as a new “O + T'' (shown above). We then connect the

ends of “O'' and the new “O'' + “T'' vector to find the new “T'' We can see that we shifted the trial

weight too far and must move it back by the angle between “O'' and the new “T'' This may be

repeated as many times as necessary but always using the original “O'' line. This then is the

vector method for single plane balancing. It is simple to use and provides accurate information to

balance a part in the fewest number of starts and stops of the machine as possible.

Four-step Method of Balancing Single Plane

Another method for balancing which is not as precise as the vector method follows the same

basic procedure except we do not construct a vector diagram. The four- step method follows a

few simple rules to find the proper location for the correction weight after which the amount is

adjusted to balance the part. Since it generally requires many starts and stops of the machine it is

not too popular except when the number of balancing runs is not important.

First we observe the amount of vibration and the phase or position of the reference mark when

the strobe light flashes. In the illustration these readings are 2.4 mils at 5: 00 o'clock. This is the

original run. AII future data will be referred to these readings just as we did in the vector method.

Next we add a trial weight to the part at any convenient location and again observe the amount

and phase; 1.8 mils at 7:00 o'clock.

82

Vibration Analysis on Rotating Equipment

Our first goal is to shift the trial weight to a position where the reference mark returns to its original

position or 180° away. We do this by shifting the trial weight in a direction opposite the shift of the

reference mark. Do not be afraid to move the trial weight by a large angle because if you move it

too far, the phase will direct you to move it back. The reference mark shifted from 5: 00 o'clock to

7: 00 o'clock after we added the trial. This is a clockwise shift of the reference mark. Therefore,

we move the trial weight in the counterclockwise direction.

A new reading might be 2.4 mils at 3:û0 o'clock (not illustrated) . This would indicate we moved

the weight too far because the mark is counterclockwise from the first position of 5: 00 o'clock.

Shifting the weight again we see the reference mark has been returned to its original position, and

the vibration has been reduced to 1 mil. This means the weight is in the proper position and we

need only to increase the size of the trial weight. If the reference mark had appeared 180° away

from its original position the weight would have been in the proper place but too large. If the

reference mark had returned to its original position but with an increased vibration then the trial

weight would have been on the heavy spot.

It should be apparent that the vector method and the four-step method both follow the same basic

rules. Know and understand both of these methods well.

Sample Problems

Following are some examples of balancing problems.

You are given the original readings and the readings after a trial weight has been added. See if

you can solve each problem correctly. Your answer should indicate first the direction to move the

trial weight, second the angle, and third, the adjustment required to get the correct size of the trial

weight

Amount Phase

1. Original Reading 5.0 mils @ 3: 00 o'clock

First Trial Reading 3.0 mils @ 4: 00 o'clock

Answer.

The reference mark moved from 3: 00 to 4: 00, a clock- wise shift. The trial weight must be moved

counterclockwise from its position for the trial run.

83

Vibration Analysis on Rotating Equipment

Discussion:

The vector diagram was constructed as shown above.

The “O'' vector was scaled off to five units long representing five mils vibration.

The “O'' + “T'' vector was scaled off to the same scale as the “O'' vector. The ends of the “O” and

“O” + ''T'' vector were connected. This is the ''T'' vector.

The angle between the “O'' vector and the “T'' vector is the angle the trial weight must be moved

from its position for the trial run.

The length of the “T'' vector was measured using the same scale used for the “O'' and “O'' + ''T''

vector The effect of the trial weight is smaller than the original imbalance because the “T'' vector is

shorter than the “O'' vector. The amount the trial weight must be increased was calculated using

formula

Corr. Wt. = T.W. x O/T.

The solution is to move the trial weight 32° counterclockwise from its position on the first trial run.

At the same time increase the size of the trial weight 1.75 times its original size.

Now, you work the rest.

Amount Phase

2. Original Reading 3.0 mils @ 7: 00 o'clock

First Trial Reading 4.0 mils @ 7: 00 o'clock

First Trial Reading 4.5 mils @ 4: 30 o'clock

84

Vibration Analysis on Rotating Equipment

First Trial Reading 1.5 mils @ 9: 00 o'clock

First Trial Reading 9.0 mils @ 12: 00 o'clock

After the corrections indicated by the readings above have been made a new reading was taken

indicating additional correction is required.

Second Trial Reading 3.0 mils @ 4: 45 o'clock

What additional correction is required? Check your answers to these problems with the answers

supplied further in the manual. Then try some real problems on a simple system like the one

pictured here in the text.

Balancing in One Run

Do not let the title of this section fool you. To date no method has been devised to permit

balancing in only one run the first time a unit is balanced. However, once a unit has been

balanced by either of the methods described before it is possible to learn how much and where

weight must be added to balance the unit in one run. This information can be learned for any

system but applies only to the individual system. If the system is changed in any way, how much

and where weight must be added will also change. To show how this is done we will use the motor

and pulley system illustrated before.

We will assume that it is the pulley that is out of balance. With the pickup mounted in the vertical

direction and the filter tuned to rotating speed of the pulley we measure the vibration caused by

the imbalance. This is the original reading which, for the example is 5.1 mils with the reference

mark stopped at 2: 00 o'clock. After the part is balanced to a vibration of 0.1 mil we can learn how

much weight was added to correct for 5.0 mils of vibration. We will assume that 20 grams of

weight added at a radius of 2.0 inches was required. The imbalance corrected was therefore 40

gram-inches. Therefore, 40 gram-inches of balance correction was required for g 5.0 mils of

vibration or 8 gram-inches per mil.

The next time it is necessary to balance this particular unit or one just like it we will know 8 gram-

inches of correction is required for each mil of vibration. For example if we measure 3 mils, 24

gram-inches is required.

lf the weight is added at two inches 12 grams will be needed. This is true of course only for the

machine mounted the same way, operating at the same speed, and with the vibration pickup in

the same place.

Knowing how much weight to add is important, of course, but where is even more important.

Before a trial weight was added to the pulley the phase or position of the reference mark was 2:00

o'clock. The pulley's position with the reference mark at 2: 00 o'clock is the position of the pulley

when the strobe light flashed. With the reference mark at 2:00 o'clock we can see in the

illustration that the weight was added at 5:00 o'clock. The heavy spot or point of imbalance is 180°

away or at 11: 00 o'clock when the strobe light flashed.

85

Vibration Analysis on Rotating Equipment

The angle between the point where the pickup is applied and the position of the heavy spot when

the strobe light flashes is called the Flash angle for the system.

Therefore, for this system the heavy spot will always appear at 11: 00 o'clock or 30 degrees

counterclockwise from the pickup when the strobe light flashes. The heavy spot can be at any

angle around the pulley but it will always appear at 11: 00 o'clock when the strobe light flashes.

The reference mark has nothing to do with the relationship since it can be placed anywhere on the

pulley. The reference mark is simply a way to see the position of the pulley when the strobe light

flashes.

After these relationships have been learned about a part then many parts may be balanced on a

production basis. These principles are also useful in two plane balancing. So that you can be sure

you understand we suggest the following experiment.

Balance a pulley mounted on a motor similar to the one we have used here. Remove the

reference mark from the pulley to avoid any confusion regarding it’s position and the heavy spot.

Next put a heavy spot on the pulley and observe its position when the strobe light flashes. Do the

same thing with the heavy spot in other locations. The location of the heavy spot with respect to

the position of the pickup when the strobe light flashes should be apparent. Now put the pickup in

a new location and repeat the experiment.

86

Vibration Analysis on Rotating Equipment

Taking measurements

The figure below shows a typical tap block for mounting an accelerometer to a motor. This block should be

machined with a “gramophone” finish on the reverse side to give good adhesion to the mounting surface.

The material used for the tap block should be capable of being magnetized so the tap blocks can also be

used for off line measurements with magnet and accelerometer prior to an accelerometer being permanently

fitted; for this reason the material quoted is bright steel or austenitic stainless.

The block is designed for mounting to the surface with an epoxy adhesive such as Loctite DEPEND or

similar. Whichever adhesive is used, it must set hard in order to transmit the vibration. The motor

mounting surface should be prepared with a hand grinder or similar to remove all paint and get down to

bare metal. The tap block should be mounted without the accelerometer and allowed to set firmly, holding

the tap block in place while the adhesive is setting with duct tape or similar.

Before mounting the accelerometer it is worth considering the purchase of accelerometers with local

connectors rather than those with integral cables so that if the motor has to be dismantled it is easier to

disconnect the accelerometer.

just wider than the base of

the accelerometer.

1/4” 28

Material:

Bright Steel or

Austenitic Stainless

Note:

One side to be machined smooth and

the other side to be machined to a

rough “gramaphone” finish.

1/4”

When it comes time to mount the accelerometer, put a thin layer of grease on the accelerometer and a thin

layer of epoxy adhesive on the top of the tap block. This will help the transmission of vibration through the

tap block and allow for easy disassembly. Be very careful not to over-tighten the accelerometer onto the tap

block - follow the manufacturers torque settings. Notice that the tap block is made from hex. stock so that if

it is necessary to remove the block then it can be wrenched off with a suitable spanner. Do not mount the

accelerometer directly onto the surface of the motor with the adhesive unless you want to lose your

accelerometers every time your motor goes away for repair.

87

Vibration Analysis on Rotating Equipment

Particularly on motors, care should be taken to ground the shield wire at one end only - the end furthest

away from the motor - on to a good electrical ground. Failure to do this will result in high amplitude

multiples of electrical line frequency (60 Hz). Do not connect both ends of the shield to ground as you will

get beautiful ground loops that will really dominate your signal. Case isolated accelerometers are usually

less susceptible to picking up electrical frequencies. When running your accelerometer cable back to the

junction box, keep the cable as far away as possible from the motor power cable to avoid “cross-talk”.

A speed output should be installed for variable speed motors. This could be in the form of a T.T.L. pulse,

once per rev from an installed proximity switch, triggered by a key or from the installed tachometer channel

on the P.L.C. which gives a 4-20 mA output. To use this output the 4-20 mA must be dropped across a

resistor of about 6250 Ω to give a 2.5 volt drop for a 10 mA signal and 5 volt at full speed (20 mA). Before

installing this resistor check with the engineers responsible for the drive system to make sure the drive

system or the P.L.C. is not adversely affected by the resistor. Once installed the speed input will have to be

calibrated in volts / rpm and fed to a channel input as opposed to a trigger input as would be required for the

T.T.L. pulse..

6

V = I.R where V is volts, I is amps, R is ohms.

88

Vibration Analysis on Rotating Equipment

Thin film of

silicone grease

Steel stud

Max temp 1000°C (1800°F)

Thin film of

silicone grease

Mica washer

Steel stud

Thin layer of

bees wax

Methyl cyanoacrylate

cement (super glue)

Methyl cyanoacrylate

cement (super glue)

soft glue

Steel stud

89

Vibration Analysis on Rotating Equipment

adhesive disk

Thick double sided

adhesive disk

disk

H a n d D u a l R a il F la t M o u n t in g A d h e s iv e S tu d

P ro b e M a g n e t M a g n e t P a d M o u n t M o u n t

Sensitivity Deviation

+ 4 0

+ 3 0

+ 2 0

+ 1 0

0

-1 0

-2 0

1 .0 1 0 1 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0

L o g F re q u e n c y (H z )

90

Vibration Analysis on Rotating Equipment

A methodology for objective set up

by

Ron Frend

Introduction

The purpose of monitoring equipment in predictive maintenance (PdM) is to be able to assess the health and

condition of the machine relative to any potential failures. In order for us to carry out this assessment we

must be absolutely certain that we are taking the correct measurements at the right place and that we are

taking them often enough so that we do not miss a developing failure. The methodology that we will

employ to ensure that we are taking the right measurements with the correct parameters is the potential

failure analysis (PFA) tree.

The tree is structured in the following way:

Setup

Interval

Analysis

Parameter

Technology

External Manifestation

Failure Type

Base Cause

Let us take each of these components and look at them in detail from the roots up.

Base cause

As the name suggests this is the root cause of any potential failure. Examples could be: lubrication

problems, misalignment, manufacturing defects and so on. The base cause often branches out to more than

one failure type, for example misalignment could cause a bearing failure or a shaft breakage.

Failure type

This is the failure that we could expect from the base cause. In other words this is a very short description

of the actual failure that the machine would suffer should the base cause carry on without remedial work

being carried out.

91

Vibration Analysis on Rotating Equipment

External manifestation

This is where we start getting technical. How will the problem show itself at the various stages of failure?

For illustration we will discuss the four stages of bearing failures and how they show themselves. An

example would be that a misalignment would show itself as an increase in vibration at the machine, whereas

a plugged heat exchanger would show itself as a rise in differential pressure and with a change in

temperature from optimum of the cooled fluid.

Technology

So now we know how the problem will show it’s ugly head but what is the best technology to detect it. We

have already said that misalignment shows itself as a rise in vibration so the technology to use would be a

vibration based technology. The heat exchanger temperature change may be detected by a mercury-in-glass

thermometer or alternatively by the use of infra-red technology. At this stage you need to have detailed

understanding of the failure modes and predictive maintenance applications.

Parameter

So the technology of choice for detection misalignment is vibration, but what type of vibration is best? In

this case the best vibration parameter is velocity. For the heat exchanger we may decide that the most

suitable parameter is a radiometric thermal image or maybe we will decide to use a simple point and shoot

infra-red thermometer.

Analysis

At this point we are giving the analyst an idea of what detail he should expect to see in the chosen parameter

when the subject machine has a developing defect. Our misalignment will show as an increase in velocity

vibration in the axial direction for a belt driven machine at run speed with two or three harmonics and our

thermal image will show a high temperature gradient across the cooler.

Interval

How often do we need to take the reading to ensure that we do not miss a developing problem. For example

if we are looking for rolling element bearing defects we should take the readings at least once a month but if

we are looking for a misalignment then once every three months would be ample.

Setup

To detect the misalignment we have said that we will use vibration technology with a velocity parameter

and that we are looking for two or three multiples of run speed in the frequency spectrum. If that is the only

defect that we are looking for then we can safely set the maximum frequency in the spectrum at about ten

time run speed with 400 lines of resolution as we are not looking for small changes in frequency such as slip

sidebands.

92

Vibration Analysis on Rotating Equipment

Bearings

Consider figure 1, above. It shows the four stages of a rolling element bearing failure. Before you can

develop the PFA tree for the rolling element bearing you must have a thorough understanding of the

possible failure modes and extrapolate these modes to their logical conclusions.

Stage 1

Notice that the first stage of the defect is linked to lubrication problems. At this stage we can detect any

problems using high frequency vibration parameters such as demodulation or ultra sonics. To detect a

defect at this stage with vibration we need to use a high frequency technique such as envelope signal

processing (ESP) with the following parameters:

Fmax = BPFI x 8 (or thereabouts)

Envelope filter = 2.5 - 5 kHz (for electric motor bearings up to about 250 HP)

No lines = 400 or 800

Window = Hanning, Hamming or Kaiser

Averages = typically 4 with maybe 50% overlap

If your data collector does not support a high frequency function such as ESP or HFB then take an overall

acceleration reading.

93

Vibration Analysis on Rotating Equipment

Stage 2

The next stage of failure involves light marking of the bearing. Again the parameter of choice is a high

frequency technique but experience tells us that a spectral parameter is most suited to detecting marked

races. So the choice here would almost certainly be ESP. The setup for the ESP reading would be the same

as in stage 1. The onset of stage 2 can be also be detected by searching for a rise in amplitude in resonance

frequencies so if you do not have ESP you can take a velocity or acceleration spectrum with an Fmax above

resonance. If we expect resonance to occur at about 2 kHz then the Fmax should be about 3 kHz. Note that

if you are using seismic velocity transducers you will not be able to look this high in the frequency range.

Stage 3

The third stage of the bearing failure is when the bearing starts to spall. At this stage we start to see the

defect in velocity usually at the third or fifth harmonic of the bearing defect frequency. We may also see

sidebands of run speed or cage frequency around the defect frequencies or their harmonics. We should be

taking an ESP and a velocity reading. The ESP reading should be the same as that set up in stage 1. The

velocity reading should be set up as follows:

Fmax = BPFI x 8 (if also using ESP or similar reading)

or = 2.5 kHz (do not confuse AC motor frequencies with a brg defect)

No lines = 400 or 800

Window = Hanning, Hamming or Kaiser

Averages = typically 4 with maybe 50% overlap

We should also take a time waveform reading in G acceleration to look for the bearing impacts at all stages

of the failure. The time duration should be calculated to be enough to show about 3 revs of the shaft. We

also need enough resolution in the time domain to be able to differentiate the impacts and calculate the

frequencies of occurrence of the impacts. Usually 1,024 samples is adequate resolution. To calculate the

Fmax in the point setup screen we use the following formula:

time for 3 revs x 2.56

Example. For a bearing on a shaft running at 1195 rpm the time for on rev is:

60/1195

= 0.0502 seconds

or 50.2 milliseconds (mS).

If we require a resolution of 1,024 points, the appropriate Fmax would be:

Fmax = 1024

0.0502 x 3 x 2.56

= 2,656 Hz

So we would set the Fmax to 2.5 kHz or the nearest available frequency.

Stage 4

The final stage of failure is when the bearing physically collapses. At this stage the bearing overheats

dramatically and very quickly. Temperature measurement using embedded thermo-couples have been used

94

Vibration Analysis on Rotating Equipment

successfully for protection systems but usually only for thrust bearings which develop the raised

temperature before standard radial support bearings.

We now include all of the information from our failure study of the rolling element bearings into our

potential failure analysis tree (figure 2). Notice that the bearing failure is only one of the potential failures

that we now know of.

We have carried out potential failure analyses on all of the possible defects that we think a motor, which is

operating under these particular conditions, may eventually suffer. Not only have we considered the

possibility of bearing failure but we have included insulation breakdown, rotor bar defects, loose

components and even drive problems on D.C. motors.

The procedure for building the PFA tree is the same for any equipment or component:

• Identify the possible failure mechanisms

• Determine how the failure mechanism will show itself

• Quantify the best contemporary method for detecting the failure mechanism

• Define set up parameters to best use the detection method for this failure mechanism

External

Base Cause Failure Type Manifestation Technology Parameter Analysis Interval Set Up

Dry/Marked Bearing Failure High frequency Vibration Demodulation Spikes @ Brg defect On Line Fmax = BPFI x 8

bearings resonance @ bearing (Ultra Sound) frequencies (20 dBG) Radial

defect frequency intervals See sect. 2.5 G acceleration

Time

Waveform Peaks in waveform On Line = 3 revs

Spalled Bearing Failure Harmonics of bearing Vibration Velocity Spikes @ Brg defect On Line Fmax = BPFI x 8

bearings defect frequencies freqys. & harmonics (higher if not using

and/or sidebands. resonance parameter)

See sct. 2.5 Radial

Waveform 7 G (ball brg) G acceleration

Misalignment Bearing Failure/ High axial vibration Vibration Velocity Multiples of run speed On Line Fmax = 10 x rpm

Shaft Failure axial > radial Axial

Looseness Bearing Failure/ Run speed harmonics Vibration Velocity Multiples of run speed On Line Fmax = 10 x rpm

Shaft Failure/ & subharmonics up to 15x. Possible Vertical

Structural Failure sub-harmonics

Bad S.C.R.s Reduced Power Vibration @ SCR Frqys Vibration Velocity 240 Hz, 1/3 DC pulse, On Line Fmax = 120 kcpm

Motor Burnout s/band on DC(Sect.2.5) Drive End

Changes in SCR temps. Thermography Thermal Image Bad SCR is colder 6 month Compare SCR to SCR

Different current / Current Draw Amps Unbalanced supply 6 month Compare phase to phase

input phase

Winding Motor burnout Stator temp. rise/ Thermography Thermal image Hot spot on stator 6 month 1 baseline

defects Uneven temp. distribution (spot temp)

Earth current leakage Insulation test Megger Should be > 1.5 MΩ 6 month Motor de-energized

Rotor Rotor failure Line frequency + Motor load Motor current Sidebands of slip x No. 6 month Press idling

defects sidebands current analysis frequency of poles around 60 Hz

(Sect. 2.2)

Motor burnout Vibration @ rotor bar Vibration Velocity Sidebands of 1x On Line Fmax = RBPF x 3

pass frequency (Sect 2.2) Radial

c. Ron Frend - PreDiCon

95

Vibration Analysis on Rotating Equipment

Conclusion

The method outlined in this paper for setting up the type of readings in a predictive maintenance system will

result in quantifiable, repeatable and useful measurements if these factors are taken into consideration:

• The engineer or technician defining the measurements must be familiar with previous failures

encountered on machines similar to the machine under scrutiny

• The engineer or technician defining the measurements should be cognizant of all

• available predictive maintenance technologies.

• The engineer or technician defining the measurements must be aware of the limitations of the

technology which is used for the measurement.

Finally, don’t limit yourself to only taking vibration measurements - there are a host of other technologies

out there which complement vibration in a predictive maintenance application.

CHUTE ASSEMBLIES

CLAMP/SEAL FAILURE INSPECTION - CCTV

CAM SIEZURE FORIEGN BODY INGRESS INSPECTION - CCTV

CORROSION (reference marks req'd)

FRETTING/CLEARANCE INCREASE

DRIVE ASSEMBLY

BEARING CLEARANCE INCREASE MATERIALS / LOADING DESIGN

OVER TEMPERATURE LUBRICATION / LOADING DESIGN

CORROSION BEARING DEFECTS

BRINNELLING

MATERIAL COMPATIBILITY / CORROSION DESIGN

HEAD 1

96

Vibration Analysis on Rotating Equipment

BEARING COLLAPSE

PROBLEMS

OVER TEMPERATURE

ULTRA-LOW SPEED

FORIEGN BODY INGRESS BEARING DEFECTS

PARTIAL ROTATION

CORROSION

INNER BEARING MOUNTS

BRINNELLING

PROBLEM

FREQUENCY RESOLUTION NOT POSSIBLE

TIME-BASE READINGS REQ'D IN : ACCELERATION

STRESS WAVE

RADIATION

HEAD 2

97

Vibration Analysis on Rotating Equipment

External

Base Cause Failure Type Manifestation Technology Parameter Analysis Interval Set Up

Dry/Marked Bearing Failure High frequency Vibration Demodulation Spikes @ Brg defect On Line Fmax = BPFI x 8

bearings resonance @ bearing (Ultra Sound) frequencies (20 dBG) Radial

defect frequency intervals See sect. 3.5 G acceleration

Time

Waveform Peaks in waveform On Line = 3 revs

Spalled Bearing Failure Harmonics of bearing Vibration Velocity Spikes @ Brg defect On Line Fmax = BPFI x 8

bearings defect frequencies freqys. & harmonics (higher if not using

and/or sidebands. resonance parameter)

See sct. 3.5 Radial

Waveform 7 G (ball brg) G acceleration

Looseness Bearing Failure/ Run speed harmonics Vibration Velocity Multiples ofBPFO On Line Fmax = 10 x rpm

Shaft Failure/ & subharmonics or BPFI in velocity Vertical

Structural Failure Sect. 3.5 1 Baseline

98

Vibration Analysis on Rotating Equipment

Measurement Windows

Many people get confused with this topic. What is windowing? Let's go back to the time record. What

happens if a signal is not exactly periodic within the time record? We said that its amplitude is divided into

multiple adjacent frequency bins. This is true but it's actually a bit worse than that. If the time record does

not start and stop with the same data value, the signal can actually smear across the entire spectrum. This

smearing will also change wildly between records because the amount of mismatch between the starting

value and ending value changes with each record.

If a sine wave is passing through zero at the beginning and end of the time series, the resulting FFT

spectrum will consist of a single line with the correct amplitude and at the correct frequency. If, on the

other hand, the signal level is not at zero at one or both ends of the time series record, truncation of the

waveform will occur, resulting in a discontinuity in the sampled signal. This discontinuity causes problems

with the FFT process, and the result is a smearing of the spectrum from a single line into adjacent lines.

This is called "leakage"; energy in the signal "leaks" from its proper location into the adjacent lines.

Leakage could be avoided if the time series zero crossings were synchronized with the sampling times, but

this is impossible to achieve in practice. The shape of the "leaky" spectrum depends on the amount of

signal truncation, and is generally unpredictable for real signals.

In order to reduce the effect of leakage, it is necessary that the signal level is forced zero at the beginning

and end of the time series. This is done by multiplying the data samples by a "smoothing window" function,

which can have several different shapes. The difference between each smoothing window is the way in

which they transition from the low weights near the edges to the higher weights near the middle of the

sequence. If there is no windowing function used, this is called "Rectangular", "Flat", or "Uniform"

windowing.

While the smoothing window does a good job of forcing the ends to zero, it also adds distortion to the time

series which results in sidebands in the spectrum. These sidebands, or side lobes, effectively reduce the

frequency resolution of the analyzer; it is as if the spectral lines are wider. The measured amplitude of the

weighted signal is also incorrect because a portion of the signal level is removed by the weighting process.

To make up for this reduction in power, windowing algorithms give extra weight to the values near the

middle of the sequence.

Windows are functions defined across the time record which are periodic in the time record. They start and

stop at zero and are smooth functions in between. When the time record is windowed, its points are

multiplied by the window function, time bin by time bin, and the resulting time record is by definition

periodic. It may not be identical from record to record, but it will be periodic (zero at each end).

In the frequency domain, a window acts like a filter. The amplitude of each frequency bin is determined by

centering this filter on each bin and measuring how much of the signal falls within the filter. If the filter is

narrow, then only frequencies near the bin will contribute to the bin. A narrow filter is called a selective

window - it selects a small range of frequencies around each bin. However, since the filter is narrow, it falls

off from center rapidly. This means that even frequencies close to the bin may be attenuated somewhat. If

the filter is wide, then frequencies far from the bin will contribute to the bin amplitude but those close by

will not be attenuated significantly.

The net result of windowing is to reduce the amount of smearing in the spectrum from signals not exactly

periodic with the time record. The different types of windows trade off selectivity, amplitude accuracy, and

noise floor.

Several types of window functions are available including Uniform (none), Flattop, Hanning, Blackman-

Harris, and Kaiser.

Uniform

The uniform window is actually no window at all. The time record is used with no weighting. A signal will

appear as narrow as a single bin if its frequency is exactly equal to a frequency bin. (It is exactly periodic

99

Vibration Analysis on Rotating Equipment

within the time record). If its frequency is between bins, it will affect every bin of the spectrum. These two

cases also have a great deal of amplitude variation between them (up to 4 dB).

In general, this window is only useful when looking at transients which do not fill the entire time record.

Hanning

The Hanning window is the most commonly used window. It has an amplitude variation of about 1.5 dB

(for signals between bins) and provides reasonable selectivity. Its filter roll off is not particularly steep. As

a result, the Hanning window can limit the performance of the analyzer when looking at signals close

together in frequency and very different in amplitude.

Flattop

The Flattop window improves on the amplitude accuracy of the Hanning window. Its between-bin

amplitude variation is about 0.02 dB. However, the selectivity is a little worse. Unlike the Hanning, the

Flattop window has a wide pass band and very steep rolloff on either side. Thus, signals appear wide but do

not leak across the whole spectrum.

Blackman-Harris

The Blackman-Harris window is a very good window to use with the spectrum analyzer. It has better

amplitude accuracy (about 0.7 dB) than the Hanning, very good selectivity and the fastest filter rolloff. The

filter is steep and narrow and reaches a lower attenuation than the other windows. This allows signals close

together in frequency to be distinguished, even when their amplitudes are very different.

Kaiser

The Kaiser window, which is available on IRD analyzers, combines excellent selectivity and reasonable

accuracy (about 0.8 dB for signals between exact bins). The Kaiser window has the lowest side-lobes and

the least broadening for non-bin frequencies. Because of these properties, it is the best window to use for

measurements requiring a large dynamic range.

Averaging

In general, averaging many spectra together improves the accuracy and repeatability of

measurements.

RMS Averaging

RMS averaging computes the weighted mean of the sum of the squared magnitudes (FFT times its complex

conjugate). The weighting is either linear or exponential.

RMS averaging reduces fluctuations in the data but does not reduce the actual noise floor. With a sufficient

number of averages, a very good approximation of the actual random noise floor can be displayed.

Since RMS averaging involves magnitudes only, displaying the real or imaginary part or phase of an RMS

average has no meaning. The RMS average has no phase information.

Vector averaging averages the complex FFT spectrum. (The real part is averaged separately from the

imaginary part.) This can reduce the noise floor for random signals since they are not phase coherent from

time record to time record.

100

Vibration Analysis on Rotating Equipment

Vector averaging requires a trigger. The signal of interest must be both periodic and phase synchronous

with the trigger. Otherwise, the real and imaginary parts of the signal will not add in phase and instead will

cancel randomly.

With vector averaging, the real and imaginary parts as well as phase displays are correctly averaged and

may be displayed. This is because the complex information is preserved.

Peak Hold

Peak Hold is not really averaging, instead, the new spectral magnitudes are compared to the previous data,

and if the new data is larger, then the new data is stored. This is done on a frequency bin by bin basis. The

resulting display shows the peak magnitudes which occurred in the previous group of spectra.

Peak Hold detects the peaks in the spectral magnitudes and only applies to Spectrum, PSD, and Octave

Analysis measurements. However, the peak magnitude values are stored in the original complex form. If

the real or imaginary part or phase is being displayed for spectrum measurements, the display shows the real

or imaginary part or phase of the complex peak value.

Linear Averaging

Linear averaging combines N (number of averages) spectra with equal weighting in either RMS, Vector or

Peak Hold fashion. This type of averaging is useful for eliminating transients.

Exponential Averaging

Exponential averaging weights new data more than old data. Averaging takes place according to the

formula,

New Average = (New Spectrum * 1/N) +(Old Average) * (N-l)/N

where N is the number of averages.

Exponential averages "grow" for approximately the first 5N spectra until the steady state values are reached.

Once in steady state, further changes in the spectra are detected only if they last sufficiently long. Make

sure that the number of averages is not so large as to eliminate the changes in the data that might be

important.

What is real time bandwidth? Simply stated, it is the frequency span whose corresponding time record

exceeds the time it takes to compute the spectrum. At this span and below, it is possible to compute the

spectra for every time record with no loss of data. The spectra are computed in "real time". At larger

spans, some data samples will be lost while the FFT computations are in progress.

What about narrow spans where the time record is long compared to the processing time which is what we

normally see when taking vibration measurements? The analyzer computes one FFT per time record and

can wait until the next time record is complete before computing the next FFT. The update rate would be

no faster than one spectra per time record. With narrow spans, this could be quite slow.

And what is the processor doing while it waits? Nothing. With overlap processing, the analyzer does not

wait for the next complete time record before computing the next FFT. Instead it uses data from the

previous time record as well as data from the current time record to compute the next FFT. This speeds up

the processing rate. Remember, most window functions are zero at the start and end of the time record.

Thus, the points at the ends of the time record do not contribute much to the FFT. With overlap, these

points are “re-used" and appear as middle points in other time records. This is why overlap effectively

speeds up averaging and smoothes out window variations.

101

Vibration Analysis on Rotating Equipment

Overlap Percentage

The amount of overlap is specified as a percentage of the time record. 0% is no overlap and 99.8% is the

typical maximum. The maximum overlap is determined by the amount of time it takes to calculate an FFT

and the length of the time record and thus varies according to the span. For vibration analysis of rotating

machinery a good overlap is 50% as this ensures that no data is zeroed out by the smoothing windows, yet

sufficient samples are gathered for a valid analysis.

Octave Analysis

The magnitude of the normal spectrum measures the amplitudes within equally divided frequency bins.

Octave analysis computes the spectral amplitude in logarithmic frequency bands whose widths are

proportional to their center frequencies. The bands are arranged in octaves with either 1 or 3 bands per

octave (1/1 or 1/3 octave analysis). Octave analysis measures spectral power closer to the way people

perceive sound, that is, in octaves. For vibration measurements, octave analysis is now only used for sound

analysis or for quality acceptance checks at the end of an assembly line but is included here for the sake of

completeness.

The center frequency of each band should be calculated according to ANSI standard S1.11 (1986).

Typically the shape of each band is a third-order Butterworth filter whose bandwidth is either a full or 1/3

octave. The full octave bands have band centers at:

Center Freq. = 1 kHz x 2n

The 1/3 octave bands have center frequencies given by:

Center Freq. = 1 kHz x 2(n-30/3)

102

Vibration Analysis on Rotating Equipment

Analysis

The following procedure gives an overview of the procedure to take in the analysis of

vibration frequency spectra.

Predictive Maintenance Procedure

Date:July 11, 2003

Procedure Name: Signature Analysis

Purpose: To provide a routine procedure for analysis of vibration in order to promote

understanding of the relationship between vibration frequencies and their causes.

Activities:

1) Select the first plot of the machine which will be at the driver outboard and will

be an enveloped acceleration reading. Identify the run speed accurately. If you

have frequency information for the machine ensure that the reference speed is

accurate - if not you must change the speed reference before continuing.

2) In the envelope spectrum see if any of the generated frequencies coincide with (or

are close to) any significant spikes. Remember that the bearing frequencies may

not be completely accurate if the bearing which has been nominated in the

frequency setup has been replaced with an equivalent.

• As a rule of thumb:

• Ball Pass Frequency Outer Race ≅ Run Speed x No. Of Rolling

Elements x 0.4

• Ball Pass Frequency Inner Race ≅ Run Speed x No. Of Rolling

Elements x 0.6

• Cage Frequency ≅ Run Speed x 0.4

• Note that this vibration is not necessarily direction specific.

3) Once a spike at a bearing frequency has been identified you should check the

baseline for this type of machine for the trend. If the trend is deteriorating then

further checks are necessary. Be careful that you do not confuse a run speed

harmonic or an electrical frequency with a bearing defect frequency. One

common bearing frequency is just over 3 x run speed for BPOR on a 8 element

bearing. The run speed of an electric motor cannot exceed the electrical speed so

harmonics of run speed cannot have a frequency even slightly more than 3600

cpm (for a 1200 rpm motor) or 5400 cpm (for a 1800 rpm motor) or 10,800 cpm

(for a 3600 rpm motor) - if the spike is even at a slightly higher frequency then it

is likely caused by a bearing defect otherwise it is likely a run speed harmonic or

an electrical frequency. Check the time domain signal for “angel fish” patterns.

103

Vibration Analysis on Rotating Equipment

4) In the single spectrum plot double click on the convert the display to dB(G)

(referenced to 0.001 G). Identify the carpet amplitude in dB(G) - this is the

average amplitude (excluding spikes). Identify the amplitude of the spikes above

the carpet level. The following rules of thumb apply to enveloped acceleration

levels in dB(G):

• > 20 dB(G) rise in carpet level - under lubrication

• 10-15 dB(G) spike above carpet - minor marking

• > 20 dB(G) spike above carpet - marked race.

5) If the spectrum shows multiples of run speed then there is an impact every rev of

the rotor, with possible looseness if there are many multiples. If the bearing

defect frequencies have sidebands of cage frequency then there is a FALSE

BRINELLING problem. If the bearing defect frequency and harmonics have

sidebands of run speed then there is probably a defect on the inner race. As the

defect deteriorates then the carpet level will rise and the sidebands and harmonics

will increase in amplitude up to a certain amplitude then stop. The carpet level

will continue to rise as random marking occurs around the bearing and may rise to

mask the spikes completely.

6) Move to the velocity spectrum for the same point but with amplitude set to linear.

Check to see if there are any spikes in velocity at the bearing defect, harmonics of

the defect and/or sidebands of the defect - look particularly for the third and fifth

harmonics. If any spike at these frequencies exist then there is physical spalling of

the race. If the amplitude of the spike reaches 1 mm/s then the spalling is severe.

7) In the velocity spectrum the following patterns indicate the associated defects:

Dominant Secondary Harmonics Sidebands Dominant Defect Suggested

Frequency Frequency of Direction Maximum

Dominant Amplitude

Frequency @

Dominant

Frequency

1x Nil Nil Nil Radial Imbalance 6 mm/s

1x 1/2 or 1/3 Multiple Nil Radial Looseness 3 mm/s

x

1x 3x 2 or 3 Nil Axial Misalignment 4 mm/s

BPFO / BPFI 1x Multiple 1x / Cage Radial Bearings 0.5 mm/s

7,200 cpm Rotor Bar 2 or 3 2xLF of RBF Any Electrical 5 mm/s

Any Any Nil Nil Any Resonance 7 mm/s

Gear Mesh 1x 3 or 4 1x Radial Gearing 1 mm/s

2x Belt 1x 2 or 3 N/A Radial Belts 5 mm/s

104

Vibration Analysis on Rotating Equipment

Severity charts

Figure 95 shows a general severity chart for vibration which is in widespread use. There are many versions

of these charts. The best use for these charts is for new or rebuilt equipment acceptance limits.

Dependency on these charts can be confusing, especially if the bedplate is flexible or on resilient mounts.

105

Vibration Analysis on Rotating Equipment

Two-channel analyzers offer additional measurements such as transfer function, cross-spectrum, coherence

and orbit. These measurements are discussed below.

Transfer Function

The transfer function is the ratio of the spectrum of channel 2 to the spectrum of channel 1. For the transfer

function to be valid, the input spectrum must have amplitude at all frequencies over which the transfer

function is to be measured.

Cross Spectrum

The cross spectrum is defined as:

cross spectrum = FFT2 conj(FFT1)

The cross spectrum is a complex quantity which contains magnitude and phase information. The phase is

the relative phase between the two channels. The magnitude is simply the product of the magnitudes of the

two spectra. Frequencies where signals are present in both spectra will have large components in the cross

spectrum.

Orbit

The orbit is simply a two dimensional display of the time record of channel 1 vs. the time record of channel

2. The orbit display is similar to an oscilloscope displaying a "Lissajous" figure.

Coherence

Coherence measures the percentage of power in channel 2 which is caused by (phase coherent with) power

in the input channel. Coherence is a unit-less quantity which varies from 0 to l. If the coherence is 1, all the

power of the output signal is due to the input signal. If the coherence is 0, the input and output are

completely random with respect to one another. Coherence is related to signal to noise ratio (S/N) by the

formula:

S/N = γ2/(l -γ2)

where γ2 is the traditional notation for coherence.

Correlation

The two channel analyzer may also compute auto and cross correlation. Correlation is a time domain

measurement which is defined as follows:

Auto Correlation(τ) = ∫x*(t)x(t-τ)dt

Cross Correlation(τ) = ∫x*(t)y(t-τ)dt

where x and y are the channel 1 and channel 2 input signals and the integrals are over all time. It is clear

that the auto correlation at a time “t” is a measure of how much overlap a signal has with a “delayed-by-t”

version of itself, and the cross-correlation is a measure of how much overlap a signal has with a “delayed-

by-t” version of the other channel. Although correlation is a time domain measurement some analyzers use

frequency domain techniques to compute it in order to make the calculation faster.

106

Vibration Analysis on Rotating Equipment

Spectrum

Although the linear magnitude scale is used most often for displaying spectra, another way of displaying

amplitude is the Log Magnitude. The Log Mag display graphs the magnitude of the spectrum on a

logarithmic scale.

Why is the Log Mag display useful? Remember that the 16 bit analyzer has a dynamic range of about 90

dB. below full scale. Imagine what something 0.01% of full scale would look like on a linear scale. If we

wanted it to be 1 inch high on the graph, the top of the graph would be 833 feet above the bottom - It turns

out that the log display is both easy to understand and shows features which have very different amplitudes

clearly.

The real and imaginary parts are always displayed on a linear scale. This avoids the problem of taking the

log of negative voltages.

Phase

In general, phase measurements are only used when the analyzer is triggered. The phase is relative to the

pulse of the trigger.

The phase is displayed in degrees or radians on a linear scale, usually from -180 to +180 degrees.

The phase of a particular frequency bin is set to zero in most analyzers if neither the real nor imaginary part

of the FFT is greater than 0.012% of full scale (-78 dB below f.s.). This avoids the messy phase display

associated with the noise floor. (Remember, even if a signal is small, its phase extends over the full 360

degrees.)

107

Vibration Analysis on Rotating Equipment

Advanced functions

Sometimes the representation of the spectrum is carried out by the use of rotating vectors instead of sine

waves. For complex signals this is often much more convenient. Consider the vector below.

a a+b

ϖt

0 φ

b

The parallelogram rotates at the vibration frequency ϖt so the two vibrations must be at the same frequency

for this representation to work. For numerical addition, however the vectors must be resolved

geometrically which loses almost all of the advantage of ease. There exists a simpler method of handling

the vectors numerically by employing imaginary numbers. A complex number can represented graphically

by a point in a plane where the real numbers 1,2,3 etc. are plotted horizontally and the imaginary numbers

are plotted vertically. With the notation

j = √-1

these imaginary numbers are j, j2, j3, etc. Harmonic motions are represented by rotating vectors. A

substitution of the variable angle ϖt for the fixed angle of the vector (α) leads to

a(cos ϖt + j sin ϖt)

representing a rotating vector, the horizontal projection of which is the harmonic motion. But this

horizontal projection is also the real part of the vector. So if we say that a “vector represents harmonic

motion” what we really mean is that the horizontal projection of the rotating vector represent that motion.

Similarly if we say that “a complex number represents harmonic motion” we imply that the real part of such

a number, written in the form “a(cos ϖt + j sin ϖt)” represents that motion. Almost all of the algorithms in

the analyzer which involve phase make reference to the imaginary number. We do not necessarily need to

make much use of this number but we do need to know where it comes from. The first use that we will

make of the imaginary number is the Nyquist chart which is useful for identifying resonances.

108

Vibration Analysis on Rotating Equipment

There is quite a lot of confusion over the terms cascade and waterfall plots in the vibration world. A

cascade plot is a 3-D representation of the amplitude against speed with respect to time and the waterfall is

the 3-D representation of the amplitude against frequency with respect to time such as the example shown

above. The advantage of these plots is that we can record data over a certain period of time and see how all

of the relevant frequencies are affected. This could be a very high speed machine over a short time duration

or a slow speed machine over a very long time period.

Whichever is the particular application, the collection of data must be considered very carefully before data

acquisition takes place. The OR25 series analyzer is limited to displaying data at the acquired Fmax or

1/10th or 1/100th. Ensure that the data collection rate is valid for the necessary analysis.

Triggering

As mentioned previously a trigger may be set up on a machine to control the acquisition of data. This

trigger is often a 5 volt, once per rev pulse or it could be from an encoder giving many pulses per rev. A

many pulse per rev trigger requires an external clock input to the collector which is programmed to the

number of pulses per rev. A third type of trigger is a single pulse or voltage change and is known as an

event trigger.

As the name implies this is a simple pulse, once per rev of the shaft, which triggers data collection at a

specific time. This function is useful in cascade plots, orbits, or for Bodé plots.

Encoder trigger

The encoder will give many pulses per rev and must be input to the external clock input. This input is

necessary for torsional vibration and should ideally be used for time synchronous averaging where the

speed may be expected to change from moment to moment.

109

Vibration Analysis on Rotating Equipment

Event trigger

This is probably the simplest trigger and may be setup as a channel reading or an external trigger. The

trigger level is set and the analyzer just sits there waiting for the trigger amplitude to reach the pre-

defined level. The analyzer is actually collecting data all of the time but is only storing as much as the

memory will allow. When the trigger is activated the data storage will commence. If the trigger is set

up as a pre-trigger, then some data collected before the trigger event will be collected as well as some

data after the trigger. A post-trigger will collect data only after the event. This can be very useful if the

subject machine is suffering from a transient fault.

Bodé plots

When the cross channel properties are calculated by the analyzer, they can be displayed as amplitude or

phase against speed (Bodé) plots or real-imaginary (Nyquist) plots. The Bodé plot is useful when assessing

the resonant condition of a machine or even bearing deterioration as in the example below.

Orbits

Displacement sensors measure the amount of shaft displacement, or rather, the total motion of

the shaft's orbit within the bearing. The display is set so that the output of one transducer is set to

one (X) axis while the output of the other sensor is on the other (Y) axis. The displacement

sensors are mounted orthogonally which is so that each sensor is 45° from the vertical. In the first

example we see an elliptical orbit which is typical of preload caused by misalignment.

110

Vibration Analysis on Rotating Equipment

BALANCE

Diagnosis of a degrading balance condition is performed by

concentrating on the synchronous amplitude which coincides with the

rotor speed. This can be accomplished by viewing the spectra from

any single Eddy Probe sensor. A similar diagnosis can be made by

viewing the filtered signals from two orthogonally mounted Eddy

Probes sensors as orbits. As the balance condition deteriorates the

size, and sometimes the shape, of the orbit will grow larger until the

peak-to- peak amplitude exceeds acceptable limits.

CRACKED SHAFT

A crack in a rotor, or shaft, can generate several different effects on

how the machine behaves: a change in the vibration level, a change in

the operating phase angle, and/or a change in the resonance frequency

as the machine starts or stops. Spectral analysis can be used to identify

this fault, but observing filtered, synchronous orbits with the phase

angle superimposed on the orbit allows rapid identification of this

condition.

analysis. By superimposing the phase angle input signal onto the orbit a shift in this parameter can

be easily determined. By noting the operating speed at which the resonance frequencies occur, a

change in this frequency may indicate the "possibility" of a crack existing.

The "possibility" must be emphasized and carefully analyzed because many other causes can

produce these changes, such as, a damaged or loose bearing support, foundation problems, loose

rotating parts...basically anything that can influence the "system" mass, damping, and/or stiffness.

A loose rotating part can generate unusual vibration signals. They may

fluctuate in amplitude and the phase angle may shift, also. This fault is

diagnosed easiest using filtered, synchronous orbit analysis. Imagine a

111

Vibration Analysis on Rotating Equipment

mass, such as an impeller, which has come loose; it can rotate freely on the shaft independently.

As the loose part rotates it influences the balance condition of the rotor which appears as a cyclical

increase and decrease in the synchronous amplitude. This is observable using a spectrum analyzer,

but the changes may be too rapid for the sampling rate of the instrument. An oscilloscope set up to

observe a filtered orbit will sample continuously so that the changes can be seen. The phase

shifting can, also, be observed using an oscilloscope.

The inception of a loose part condition will produce a "nervous" filtered, synchronous orbit. The

orbit will appear to vibrate slightly as this condition is created; the part may be slipping and then

sticking on the shaft just prior to becoming a full fledged loose rotating part.

OIL WHIRL

Oil whirl and oil whip are sometimes listed as a single machine fault,

but closer observation of the vibration signals and the machine

conditions causing these signals will produce different, distinct signal

displays for each condition. This fault is caused by a condition which

prevents the rotor from creating a stable oil wedge on which ride. An

improperly designed bearing is the usual source for oil whirl

conditions, but a change in the fluid viscosity or machine alignment

state are other possibilities.

Generally, an oil whirl condition precedes an oil whip condition. Spectral and orbit analysis can be

used to identify either condition. This phenomenon creates an individual sub-synchronous

frequency which can occur within a frequency range from 35% to 48% of rotor speed, depending

upon the machine/bearing design or construction. As the machine accelerates the whirl frequency

will increase as machine speed increases.

Observing oil whirl as a filtered, synchronous orbit produces a distinctive display. The orbit will be

more or less round in shape with an amplitude that nearly approximates the bearing clearance, and

when the phase angle is superimposed upon the display, the orbit will appear to have two phase

marks on it. This characteristic is due to filtering at shaft speed and the fault being generated at a

sub-synchronous frequency. The two phase marks will not be displayed symmetrically on the orbit

because the whirl frequency is not at exactly ½ machine speed.

OIL WHIP

Oil whip occurs during the later stages of an oil whirl condition and it

has a distinctive orbital display. The display, with the phase input

superimposed on the display, appears to have several phase marks.

This display will be round in shape and the amplitude will greater that

the amplitude noted during oil whirl.

The size of the orbit be will larger because the shaft uses up the entire

bearing clearance as an oil wedge can no longer be established by the

rotor and the shaft is in direct metal-to-metal contact with the bearing. The orbit display will no

longer rotate because the oil whirl frequency has coincided with the first natural resonance, or

critical speed, and has "locked" onto this frequency. Oil whip is a dangerous condition because the

rotor uses up the entire bearing clearance and is in direct metal-to-metal contact that will wear

away the bearing rapidly and destroy the rotor if not corrected.

EXCESSIVE PRELOAD

All journal bearing machines have some amount of preload so that a

stable oil wedge can be established. The preload may be internally or

112

Vibration Analysis on Rotating Equipment

externally produced. Internal sources of preloads are from gear meshing or hydraulic loading

during pumping actions. External preloads may be from coupling misalignment or piping and

support system thermal changes. These sources of preload create an elliptical orbit that is flattened

in the direction of the preload vector.

As the preload increases the orbit is further flattened. As excessive preload increases further the

orbit begins to collapse to form a "banana" shape as the shaft tries to continue its normal rotation

pattern and direction.

After the orbit has been flattened into the "banana" shape a 2X frequency is present on spectra

displays. Heavy preloads further distort the orbit into a figure eight shape. As preload increases the

shaft centreline will shift in the direction of the preload vector.

RUB

A common problem in newly rebuilt or modified rotors is a slight rubbing condition as the rotor is

initially operated. Rotor rubs are not a phenomenon which continues

over an extended period; they usually increase the clearances until the

rub has been cleared or, if not corrected,

they will wear away the internal

clearances until the machine cannot be

operated. The shape of the orbit display

will differ depending upon the

relationship of the shaft speed to the first

natural frequency and the severity of the rub.

frequencies that occur at multiples of a fundamental frequency. The

fundamental frequency will depend upon the relationship of the

shaft speed to the first natural resonance frequency. At shaft speeds

up to twice the natural resonance frequency, the fundamental rub frequency will coincide with the

shaft speed with multiples at 2X, 3X, etc. Between twice and three times the first natural resonance

frequency, the fundamental rub frequency will be ½ shaft speed with multiples at 1X, 3/2X, 2X,

5/2X, etc. Between three and four time the natural resonance frequency, the fundamental rub

frequency will be shaft speed with multiples at 2/3X, 1X, 4/3X, 5/3X, 2X, 7/3X, etc.

The severity of the rub will affect the shape of the orbit. A light rub will produce a "tear drop"

shaped orbit, with the point of the tear drop coinciding with the impact spot. As the rub gets

heavier the orbit will flattened and may appear as an excessive preload.

At higher machine speeds (above twice the first natural frequency) the unfiltered orbits will begin

to have internal loops with the fundamental rub frequency inversely proportional to the number of

internal loops. These internal loops will have their own phase marks displayed and the loops will

be located symmetrically on the display.

113

Vibration Analysis on Rotating Equipment

Introduction to resonance

The majority of vibration analysis programs concentrate on detecting dynamic machinery faults such as:

• Bearing defects

• Gearing defects

• Misalignment

• Out-of-balance.

Eventually, however, a vibration is encountered which does not fall into any of the above categories and the

vibration analyst is left scratching his head. He knows the complete structure of the machine. He knows the

bearing numbers, the speeds, the number of teeth on the gears, the belt and pulley dimensions and still none

of the pre-calculated defect frequencies line up with the big spike in the frequency spectrum. In many cases

the big spike is related to a resonance of a natural frequency or perhaps running a machine near a critical

speed.

Many consultants in the field of vibration analysis will tell you that 90% of all failures are directly caused

by vibration resonance. This is not the case. Most failures are due to material fatigue from normal

operation leading to mechanical failure. The vibration analyst running the average predictive maintenance

system knows that most of his machinery failures are from the “bread and butter” causes listed above. The

vibration consultant, on the other hand, is usually only called in when the in-house analyst is having trouble

identifying a troublesome vibration and these problems usually are caused by resonance.

The purpose of this course is to help the in-house vibration analyst identify and cure vibration resonance

problems without having to call in outside assistance.

By the end of this course you will be able to:

• use several techniques to identify resonant conditions

• recommend appropriate measures to cure the resonance

114

Vibration Analysis on Rotating Equipment

What is resonance?

Imagine a child on a swing. You are the adult who pulls the child back and releases the swing so that the

child swings back and forth. The rate at which the child swings is the NATURAL FREQUENCY of the

child/swing system. Every time that the child swings back to your position, you give the swing a small push

– you are the EXTERNAL EXCITING FORCE. You time the small pushes to coincide with the rate of

swing of the child – in other words the exciting force is at the same frequency as the natural frequency of

the system. Each small push (as long as it is large enough to overcome friction or damping) will add to the

amplitude of the swing. In this case the child/swing system is in RESONANCE with the exciting force.

Imagine what would happen if you pushed at a different frequency. Sometimes the pushes would add to the

amplitude of the swing but sometimes the pushes would reduce the amplitude of the swing. In this case the

system is not in resonance with the exciting force.

In the world of maintenance the exciting forces are from a myriad of sources with the strongest usually

being related to run speed. Such forces include imbalance, misalignment, gear mesh, bearing frequencies

etc. In fact, anything which can be measured as a vibration may be considered to be an exciting force.

The problem we have is that we can easily determine the exciting force frequencies but cannot always tell

which system has it’s natural frequency being excited. When we say a system, we mean any part or

component (or combination of either) of the machine and/or the structure. Looking at the average machine

and associated structure, we can see a host of interconnectivity which can give rise to a large number of

natural frequencies. Fortunately for us, even though there may many natural frequencies, most of the

systems have inherent damping. The damping acts like the friction in the child/swing system. If the

exciting force is not large enough to overcome the damping, then the amplitude at that frequency will

diminish. If, however, there is little damping then the interaction of an exciting force with a natural

frequency can give rise to very large amplitudes of vibration resulting in heavy load cycling and eventual

failure of the material.

Natural Frequency

The natural frequency of a system is that frequency at which the system will oscillate or vibrate when

excited by a single external exciting pulse. Any oscillating object has a natural frequency, which is the

frequency an oscillating object tends to settle into if it is not disturbed.

Mechanical

For example, the natural frequency of a pendulum 1 m (39 in) long is 0.5 Hz, which means the pendulum

swings back and forth once every 2 seconds. If the pendulum is struck lightly once every 2 seconds, the

amplitude of the swing increases gradually until the amplitude of oscillation is very large. The phenomenon

in which a relatively small, repeatedly applied force causes the amplitude of an oscillating system to

become very large is called resonance.

Let us consider the pendulum mentioned above. The natural period of oscillation of a pendulum is

calculated using the following formula:

115

Vibration Analysis on Rotating Equipment

π

2

L 1

4. . dφ =

g 2

x . sin( φ )

2

1

0

Notice that the time period of the swing is affected slightly by the angle of the pendulum but the effect is

only slight. The main influences on the natural time period are the length of the pendulum and gravity. To

calculate the natural frequency from the time period we simply divide the time period into one second.

In this case the time period is:

1

f Hz

Ts Equation 2 Calculating frequency from time period

1

=

Ts

In a conventional mechanical system, such as the one shown below, we have to consider all aspects of the

system that affect the natural frequency.

116

Vibration Analysis on Rotating Equipment

• Mass

• (Spring) Stiffness

• Damping

Let us look at the spring/mass system in a little more detail.

Think of a spring with a mass of a certain weight hanging off the end. If we pull the mass down and then

release it, the spring/mass system will move vertically up and down at a rate which is determined by the

spring rate, the weight and gravity.

Newton's second law states that:

Force = mass x acceleration

Equation 3 Newton's 2nd law

So at any instant we can determine the acceleration of the mass, and by applying Newton's laws of motion

we can determine the instantaneous velocity and displacement.

1.

u .t

2

s (a t) Equation 4 Newton's law of motion

2

Where s = displacement (x)

u = velocity (x2)

a = acceleration (x3)

t = time

In the single degree of freedom system such as the mass/spring system shown, if the spring force has

magnitude kx, and the extension "x" is zero then there is no force.

The sign of the spring force is negative as it is always acting in the opposite direction to the extension. So:

Spring force = "-kx"

The damping force is also negative, being "-cx2", because the damping is always in the opposite direction

to the velocity.

The forcing mechanism is defined harmonically as Po sin ωt, and this could be as simple as someone

pushing and pulling the mass with their hand.

Note: Po sin ωt is a combination of direct force "Po" and angular position "sin ωt"

where ω = rotational velocity &

t = time

117

Vibration Analysis on Rotating Equipment

-kx –cx2 + Po sin ωt

Newton's law gives:

d2x

m ------ = mx3 = -kx –cx2 + Po sin ωt

dt2

Equation 5 Differential equation of motion of a single-degree-of-freedom system

or

mx3 + cx2 + kx = Po sin ωt

This equation is known as the "differential equation of motion of a single-degree-of-freedom system". You

can easily see the three components of mass, damping and stiffness directly affect the amplitudes of

oscillation in, acceleration velocity and displacement respectively.

In order to have a resonant condition we need to have the right combination of stiffness and mass to create

the natural frequency and a source vibration that matches the natural frequency. The damping controls the

amplitude response. The damping qualities of a structure control the resonant response. The higher the

damping effect the lower the amplitude of vibration will be as a result of a resonant condition. For example

a spring is a good example of a low damped system, a relatively small amplitude of input results in a large

amplitude response. A shock absorber is a good example of a highly damped system where a large input

results in a small response.

Pumping systems in general and hydraulic systems in particular are often associated with hydraulic

resonances. There are two different types of vibration in liquids:

1. Sonic vibration and

2. Pulse vibration

Sonic vibration

To think of a hydraulic resonance, consider a pipe filled with liquid. If we impact one end of the pipe, a

vibration will travel through the liquid in the pipe and the speed at which the vibration travels will be at the

speed of sound. Sound travels at different speeds in different liquids but in fresh water it travels at roughly

five times the speed of sound in air.

Table 1 Speed of sound in liquids

water 1497

methyl alcohol 1123

Benzene 1326

castor oil 1500

Glycerin 1923

118

Vibration Analysis on Rotating Equipment

Mercury 1440

air (reference) 320

liquid helium 240

The sound waves created in a substance travel with a certain speed through the substance. The speed with

which the sound wave or sound energy travels depends on the interaction between neighboring molecules of

the substance. The speed of sound decreases as the mass of the molecules increases because the mass of the

molecules is related to the force per acceleration of the molecule. However, the speed of sound increases as

the strength of the interaction between the molecules increases.

The exact relation between the velocity of sound, V, and the properties of the substance through which the

sound travels is given by:

V = √(elastic property/inertial property).

Equation 6 Velocity of sound in materials

For longitudinal waves in a solid, V(solid) = √(Young's modulus/ mass density).

For sound waves in water, V(water) = √(bulk modulus/ mass density)

The speed of sound in water is approximately 1500 m/s. It is possible to measure changes in ocean

temperature by observing the resultant change in speed of sound over long distances. The speed of sound in

an ocean is approximately:

c = 1449.2 + 4.6T - 0.055T^2 + 0.00029T^3 + (1.34-0.01T)(S-35) + 0.016z

Equation 7 Speed of sound in the ocean

S salinity in parts per thousand

z is depth in meters

Now, if we know how fast the impulses travel through the liquid, we can determine the resonant frequency

of oscillation for a certain length of pipe:

c = f ⋅λ

Equation 8 General formula relating speed, wavelength & frequency

where c = speed of sound in the material

f = resonant frequency

λ = length of pipe (or direct sub-multiples)

If this frequency coincides with a run speed or vane pass frequency then a serious resonant situation could

occur.

119

Vibration Analysis on Rotating Equipment

Impulsive vibration

The other type of vibration in liquids is from the pulses generated by the system itself, for example:

• Vane pass

• Piston pulse

• Hydraulic valve operation

• Impeller eccentricity

As each hydraulic pulse enters the piping system, it pushes liquid in front of it because most liquids are

virtually incompressible. The shock wave from this pulse then travels right through the system giving

system-wide excitation at that impulsive frequency.

It has been the authors experience that heavy vibrations on pumping systems are usually caused by an inter-

reaction of these impulsive and sonic vibrations. The usual outcome is a very noisy operating environment,

but mechanical failure often results. The impulsive frequency can also excite any mechanical resonances in

the system from pipework or supporting structures, particularly if they are both at the same frequencies.

Before we get too deep into the vibratory aspects of air-borne sound for vibration analysis, let us look at the

historical use of air resonance – music. The subtle timbre of a particular instrument is caused by the

particular frequency in resonance and several harmonics and/or sub-harmonics of that frequency.

In musical instruments harmonics are a series of subsidiary vibrations that accompany a primary, or

fundamental, wave-motion vibration. They result when the vibrating body, for example, a stretched string

or an enclosed air column, vibrates simultaneously as a whole and in equal parts (halves, thirds, quarters,

and so on), producing wave frequencies that are in simple ratios with the fundamental frequency (2:1, 3:1,

4:1, and so on). In musical sound the full-length vibration produces the fundamental tone (or first harmonic

or first partial), which is usually perceived as the basic pitch of the musical sound. The subsidiary

vibrations produce faint overtones (second and higher harmonics or partials). As the series progresses, the

vibrating segments become smaller, the frequencies higher, and the musical pitches closer together. The

harmonic series for the tone C is given in the accompanying notation; black notes show pitches that do not

correspond exactly with the Western tuning system. Harmonics contribute to the ear's perception of the

quality, or timbre, of a sound: On a flute, certain harmonics of the series are most prominent; on a clarinet,

others.

In “over-blowing” a wind instrument, the player isolates and makes predominant one of the higher

harmonics, thus extending the range of the instrument upward. Unvalved brass instruments, such as the

bugle, produce only the tones of the harmonic series; the valves of the trumpet and the slide of a trombone

add extra tubing, giving the player a new fundamental wavelength with a new harmonic series; the

instrument can thus produce more tones. The harmonics of string players are flutelike tones produced when

they cause the string to vibrate solely in halves (or thirds, and so on).

As in liquids (described above), to calculate the resonant frequency in a pipe, for example, we must first

know the speed of sound in air.

120

Vibration Analysis on Rotating Equipment

v = √( (γγ P) / ρ)

Equation 9 Newton-Laplace eq. for the speed of sound in a gas

where P =pressure

ρ = density,

and γ = ratio of the specific heat of the gas at a constant pressure over the specific heat at a constant

volume.

γ = cp over cv

Equation 10 Ratio of specific heats (gamma)

≈ 1.67 for monatomic gases,

1.40 for diatomic gases,

1.30 for triatomic gases,

and in the range of 1.2 to 1.1 for polyatomic gases.

Applying the Ideal Gas Law, the equation for the speed of sound in a gas becomes

v = √ (( γ k T) / m) = √ (( γ R T) / M)

Equation 11 Speed of sound in a gas

γ is the specific heat ratio,

T is the temperature in Kelvin,

k is Boltzman's constant,

m is the mass of one molecule,

R is the ideal gas constant, and

M is the molecular weight of the gas.

The speed of sound in a gas depends on the temperature, molecular weight, and molecular structure, but not

on the pressure of the gas. The speed of sound in air at a temperature of 0°C (32°F) and 50% relative

humidity is 331.6 m/s. The speed is proportional to the square root of absolute temperature and it is

therefore about 12 m/s greater at 20°C (64°F). The speed is nearly independent of atmospheric pressure but

the resultant sound velocity may be substantially altered by wind velocity.

Just as in the liquid pipe, we can calculate the natural frequency of the air inside a pipe by using the

formula:

c = f ⋅λ

Eq.8

If a tuning fork is set in vibration and held over the open end of a tube closed at one end, sound waves

which travel down the tube, will be reflected back from the closed end. If the returning wave is exactly in

phase, a condition of constructive interference exists which we call resonance and the loudness of the sound

will be greatly increased. The displacement of the air will be greatest at the open end of the tube, whereas

the pressure difference will be greatest at the closed end, and vice versa.

The lowest frequency where a condition of resonance exists in the tube occurs when the wavelength is four

times the length of the tube. This is called the fundamental mode of vibration. Higher modes of resonance

exist when the wavelength is 4/3(third harmonic), 4/5(fifth harmonic),... of the length of the tube. The first

four resonance modes for tubes open at both ends, and closed at one end are shown below:

121

Vibration Analysis on Rotating Equipment

The antinodes (A) are displacement antinodes (pressure nodes) and the nodes (N) are displacement nodes

(pressure antinodes). If the frequency is fixed, as in the case of the tuning fork held over the open end of a

tube, the resonance lengths are v/4f, 3v/4f, 5v/4f and so on for a tube closed at one end, and v/2f, 2v/2f,

3v/2f for a tube open at both ends.

When resonance exists, the displacement is a minimum (node) at the closed end, but the antinode is not

exactly at the open end. It is actually a small distance beyond it. This extra distance beyond the end of the

tube is called the end correction . The acoustic length of the tube is equal to its physical length plus the end

correction.

In a typical laboratory experiment with a water reservoir-resonance tube apparatus about 1m long, and

tuning forks of frequency about 500 Hz, three positions of resonance can usually be found. If the positions

of resonance from the open end of the tube are L1, L2, and L3, the wavelength of the sound wave is equal

to 4(L1 + e) or 4/3(L2 + e) or 4/5(L3 + e), where e is the end correction. From these relations the

wavelength may be calculated from 2(L2 - L1) or 2(L3 - L2) or (L3 - L1), and the end correction may be

calculated from 1/2(L2 - 3L1) or 1/2(3L3 - 5L2) or 1/4(L3 - 5L1). Since the frequency of the tuning fork is

known, both the velocity of sound may be determined by using the relation c = f⋅⋅λ, and the value of the end

correction may be determined from the relations above.

Karman Vortices

Everyone knows the story of the Tacoma Narrows suspension bridge. The bridge began to sway violently

when the wind blew steadily at a certain velocity. The mechanism for the resonance-destruction of the

bridge was the same mechanism that causes a flag to flutter – Karman Vortices.

122

Vibration Analysis on Rotating Equipment

As the air passes over the object a vortex will form on one side of the object and down wind. The increased

pressure will cause a deflection of the air flow so that another vortex is formed on the other side of the

object just as the first vortex dissipates. The vortices are alternately clockwise and anti-clockwise and are

shed from the object in a regular manner resulting in an alternating force on the object, perpendicular to the

air flow. There is a distinct relationship between the frequency (f) of the force, the velocity of the wind (V)

and the width or diameter (D) of the object.

f * D = 0.22

V

Equation 12 Karman vortices

In the case of the Tacoma Narrows bridge, a steady wind blew along the valley onto the side of the bridge.

As the wind hit the side structure of the bridge it developed Karman vortices. The vortices for the wind at

this speed developed an alternating force which was exactly at one of the natural modes for the complex

structure of the bridge.

Critical speed is, in general, any shaft rotational speed which is associated with high vibration amplitudes.

Often, the speed which corresponds to a rotor lateral mode resonance frequency excited by rotor unbalance,

in which case it is more correctly called the balance resonance speed.

123

Vibration Analysis on Rotating Equipment

Rotor balance resonance speed is the shaft rotational speed (or speed range) which is equal to a lateral

natural frequency of the rotor system., When the speed increases or decreases, the observed vibration

characteristics due to rotor unbalance are:

1) a peak in the 1X amplitude and

2) a rapid change in the 1X vibration phase angle.

The resonance speed is at the point which is 90° phase away from the angle of the heavy spot for that

resonance mode. This may be slightly different than the rotational speed peak amplitude point. These

changes may not happen at the same frequency due to non-linearity, damping, and/or asymmetry in the

system stiffness.

Campbell Diagram.

A diagram used in rotating machinery design. A tool for selecting and checking shaft operational rotational

speeds and other possible forcing function frequencies against the spectrum of natural frequencies to avoid

resonances. The X axis represents the various possible excitation frequencies, i.e., rotational speed (1X),

oil whirl (.40X to .48X), blade or vane passing frequencies, gear mesh frequencies, etc. The Y axis

represents the rotor lateral and torsional natural frequencies. The term is sometimes used incorrectly to

describe the cascade plot and waterfall plot.

124

Vibration Analysis on Rotating Equipment

125

Vibration Analysis on Rotating Equipment

Mode Shape

Whenever cracking of the machine or structure is reported, resonance should be suspected. Cracking almost

always occurs at one of the mode shape nodes. The material does not break due to excessive stresses but

due to the many millions of reversed stresses concentrated about the nodes, causing fatigue. The

phenomenon is similar to bending a wire back and forth until it breaks. The break has the characteristics of

a pure fatigue failure which is often mistakenly called “crystallization.” The break usually shows a

crystalline appearance and sharp edges.

This could account for so many familiar situations whereby pump shafts “mysteriously” break within a few

weeks to a few years, even though the calculated stresses indicate the shaft should last a lifetime without

breaking. Premature foundation or floor cracking, pipe weld breakage (often repetitively), gear cracking or

failure (usually in equal-spaced, radial lines), gear driven shafts broken with the characteristic 45° torsion

break, have all occurred as the result of metal or concrete fatigue at one of the nodes due to resonance.

Sometimes rotating machinery does not show excessive vibration at the usual points of measurement, such

as at the bearings. However, by feeling with fingers along the small diameter, lubrication, cooling water or

instrumentation piping, it can be determined if there are tell tale large “loops” (antinodes) and nodes,

characteristic of a resonant condition. Most often the vibration frequency is high enough and the tube or

pipe rigidity low enough to allow a higher resonance with several nodes. While it is unlikely that the

vibration on the tubing will cause vibration on the machine the problem should still be addressed to prevent

failure of the tube. It may take many months or even years for the tube or pipe to crack at one of the nodes,

but when it does, lost lubricant can wreck an expensive machine or a flammable product can be released.

The same phenomenon is very common for pressure gauges cantilevered on a pipe. If non-resonant, there is

no problem. But sometimes the pipe, with a gauge acting as a weight, is resonant to a specific frequency of

machine vibration. Eventually it will break, usually at its point of connection.

When hand feeling a component or structure reveals a possible resonant situation further confirmation

should be obtained by plotting out the mode shape on graph paper. This simple procedure can be performed

with the most rudimentary of instruments and is very effective in identifying the mode shapes. Remember to

126

Vibration Analysis on Rotating Equipment

plot both directions (horizontal and vertical). On piping sections the direction indicating the greatest “curl”

is the resonant direction.

The procedure for performing the mode shape plot is as follows:

1. Divide the component to be tested into a number of roughly equal parts and identify each location,

usually with a letter to avoid confusion.

2. Measure the vibration amplitude at the frequency of interest at each location. Displacement units are

preferred but velocity or acceleration can also be used. Overall readings can be used but filtered data is

preferred.

3. Divide each reading by 2 (for displacement).

4. On a horizontal zero line drawn on a sheet of graph paper mark the measurement locations spaced 1” or

25mm apart.

5. Plot the results above and below a horizontal zero line on a sheet of graph paper using a scale of ½ “ = 1

mil of vibration; 1 cm = 25 microns.

6. Join the dots “curve fit” if required.

Phase

If a force is slowly applied to a spring system, the force and its resulting deflection will move in the same

direction at the same time, or “in phase” with each other. If a vibrating force is applied to the same object or

“spring system,” and if the frequency of that vibrating force is below the critical speed range or resonance

range, then the force and resulting deflection will remain in phase. However, when the vibration frequency

enters the resonance frequency or the critical speed range, the deflection will time-lag the force. When the

vibration frequency reaches the actual critical speed or resonance frequency, the force will precede the

deflection by 90° . As the frequency increases further, the force increases its lead until the force finally

precedes the resultant deflection by 180° . The 180° relationship will remain the same for all frequencies

above the resonant frequency range, until the second resonance frequency range is approached. Then the

process starts all over again with another phase angle reversal, and so on. This phenomenon is shown in the

127

Vibration Analysis on Rotating Equipment

resonance diagram of a rotor’s vibration amplitude and phase relative to resonance of some part, such as the

rotor support system, part of a steel base or beam, or more rarely, resonance of the rotor itself.

This phenomenon can be used to help us determine resonance. Typically the running machine cannot be

shut down or have its speed varied for resonance tests, yet the phase change method to determine whether a

resonant condition exists can still be used. Remember that almost all troublesome resonances occur in non-

rotating parts, such as pipes, bases, pedestals, beams, decks and so on. These parts can usually have their

resonant frequencies changed while the machine is still running. Parts may be temporarily braced to add

rigidity and, therefore, moved to a higher resonance frequency that is at least 20 to 25 percent away from

the source frequency.

128

Vibration Analysis on Rotating Equipment

Another simple yet effective way to confirm whether or not a part is vibrating in resonance is the “bump”

test. With the machine shut down and a vibration pickup held or attached to the machine, simply bump the

machine or structure with a force sufficient to cause it to vibrate. Since an object will undergo free

vibration at its natural frequency when bumped or struck, the natural frequency generated in this way will be

indicated on the analyzer. If the vibration diminishes very quickly it may be necessary to bump the machine

several times in succession in order to sustain free vibration long enough to register on the analyzer

frequency meter. Although care should be taken that the frequency at which the "bumps” are struck is not

read as a resonant frequency. The Real Time Spectrum analyzer providing instantaneous display of the

vibration amplitude versus frequency data is an ideal instrument for determining a natural frequencies this

way.

One method of determining the natural frequency of a structure is to calculate it. The calculations are fairly

straightforward for simple structures. For the more complex structures found in the plant the calculations

are far more difficult and therefore not usually as accurate as those obtained through simple bump or impact

tests. Bump tests, however, are almost always limited to determining the first critical frequency of the part

bumped. Also, if the part tested is connected to another part that is relatively flexible, then it is sometimes

difficult to be sure that the resulting vibrations are originating from the part being bumped rather than from

the part to which it is connected.

129

Vibration Analysis on Rotating Equipment

Bump tests are also ideal for rotors, rolls, fan blades, beams, columns, floors, covers, pedestals and welded

steel bases. They are also practical for relatively rigid parts, including cast iron or cast steel bearing

supports and brackets.

This test is based on the principle that when a “spring system” or part is bumped or deflected in some way,

it will vibrate for several cycles at its natural frequency. Its amplitude decreases with each cycle, but its

frequency remains the same. The low frequency of the repeated bumps is not measured, only the frequency

that is generated between bumps.

Normally, bump tests are performed with the machine shut down. However, in certain cases a bump test can

be performed while the machine is running. A running machine spectrum is obtained. Another is taken

while bumping the part. The two spectra can then be compared to determine if there is a resonance. If peak-

hold averaging with running display capability is available, then its use is preferred. With the instrument set

up accordingly, take data with the machine running. Then with the machine still running, bump the

structure repeatedly. If the bumps generate natural frequency peaks that correspond to a peak produced by

the running machine (before bumping), then the amplitude on the analyzer screen should increase. The

peak’s width will also increase if the peak is affected by resonance.

Very large columns and beams, such as main steel support structures of buildings, do not as easily respond

to the regular bump test. A structure such as a platform supported by columns will have individual resonant

frequencies for each column and the platform. However, the total structure as a combination of columns

and platform, etc., also has several flexibilities depending on the direction being considered. For example,

a horizontal force can be applied in one direction at the top of the structure. Assume that the force is

applied to the horizontal beam supporting the platform. The force will deflect the beam but will also cause

the whole structure (including columns) to deflect as well. A regular impact bump test can be used for the

horizontal beam, but will not necessarily work well for the total structure that includes all the vertical

columns, beams and platform. To determine the resonant frequency of the total structure, a “reverse” bump

or pluck test works very well. Usually a clamp is attached to the top of the structure. A strong rope is tied

to the clamp. Tension is applied to the rope using any practical means such as pulling with a turnbuckle,

pulley, a “come-along” or even a lift truck or tractor. If the structure is small, such as one supporting a fan

or the total frame of a vertical pump, tension on the rope deflects the structure by several mils. For a very

large structure, such as a tower of several stories, the structure can be deflected as much as a mm or two

without harming the structure. Position the pickup and set the vibration instrument for the expected

frequency range. Now cut the rope. The structure will spring back, giving its resonant frequency in a

manner similar to regular bump tests. The rope will need to be retied, tension applied, cut, etc. Repeat the

procedure several times to make sure repeat readings are obtained. It takes more time and patience than

regular impact bump tests, but often reveals that the magnification due to resonance is not that of an

individual part but instead that of the total structure. Care should be exercised during this type of test as the

rope will whip when cut and could cause injury to anyone in its path. The area where the rope could whip

must be cordoned off and all bystanders moved out of harms way.

FFT-type analyzers have various settings which determine the type of spectra obtained. The procedure for

setting up the instrument varies slightly with each instrument but the general procedure is as follows:

1) Mount the transducer on the structure to be tested in the direction that the excessive vibration was

recorded. Ensure that the transducer weight is less than 10 percent of the system weight.

2) Set the Fmax (maximum frequency) to a suitable value higher than the suspected frequency.

130

Vibration Analysis on Rotating Equipment

3) Set the instrument up for manual amplitude scaling (disable the autorange feature) and select an

appropriate amplitude scale. Start with a sensitive scale and adjust the scale as required to suit the

strength of the impact and the size of the structure. If autorange remains active the instrument will tend

to range to the initial bump rather than the “ring down” afterwards. The amplitude scale and strength

of impact should be set to respond to the “ring down” rather than the impact.

4) Set the resolution to an appropriate number such as 400. The more the lines of resolution, the longer it

takes for the instrument to acquire the data. Most modern data collectors have the capability to generate

a “live” time display. If this is used, then a setting of 100 lines may be more appropriate. Despite the

loss of accuracy due to the lower resolution, the results are usually acceptable for most situations. Take

care that the data collection rate will be high enough to differentiate between the oscillations in the

ring-down and not cause them to be smeared.

5) If the analyzer has the capability to display continuously updated (running) spectra using “peak-hold-

type” averaging, then its use is recommended for the best results. If “peak-hold” is not available, then

the number of averages should be minimized, such as 4 to prevent excessive data collection time.

Special windows and trigger settings are available on most data collectors today. They can be used but

are not necessary for this simplified test.

6) Take a spectrum and observe the background vibration data (if any). These peaks, representing

vibration coming from other machines will have to be mentally subtracted from the bump test results.

Some condition monitoring software programs will subtract spectra as well. If you have an analyzer

which can perform negative averaging, then use this capability to remove the background vibration

after the bump test.

7) The structure should now be impacted in the direction of transducer orientation using a soft hammer or

piece of wood. A single blow should cause a response on the instrument and the strength of the blow

should be adjusted to give an appropriate response on the instrument. If multiple blows are required

then random impacts are preferred. For expected resonant frequencies in the range of normal speed

machines, such as under 5000 rpm, bumping with a rubber mallet is preferred over harder materials

such as steel or plastic. Soft wood (such as a “2x4”) can also be used. For higher expected

frequencies, harder hammer heads such as plastic are usually recommended.

A typical spectrum obtained from a bump test will look similar to regular spectra obtained from vibrating

running machines, except that the various peaks will not necessarily be related to running speed. Instead,

the peaks will relate to the resonant frequencies of various machine parts that were displaced by the

bumping action. Typically, there is one peak that has a much higher amplitude than the rest. The largest

peak usually represents the resonant frequency of the part that receives the bump test hammer’s impact,

although not necessarily in all instances. When the hammer impacts one part, it causes a deflection and

release. However, the impact also causes some deflection and release on parts that are connected to the part

receiving the hammer blow. If the other part is more flexible or has much less damping, then that part may

cause the largest amplitude peak. A single bump test, using an FFT, may not always determine with

certainty which peak represents a specific part. If multiple peaks appear then multiple tests should be

performed on adjacent structures and in different directions in order to determine which peak relates to

which structure.

Impact hammer

A better way of determining natural frequencies is to use the impact hammer. This is simply a hammer with

a force transducer mounted which will send a trigger signal to start collecting data the instant the hammer

hits the structure. An accelerometer is mounted on the structure and the resultant output is generated by the

natural frequencies of the structure.

131

Vibration Analysis on Rotating Equipment

The response can be analyzed with spectrum, Bode and Nyquist plots to identify resonances.

For specific information on the use of the impact hammer refer to the hammer operating manual. For this

course we will use medium sledge from IMI, model 086C41.

132

Vibration Analysis on Rotating Equipment

Care should be taken when defining the spectrum that sufficient resolution is used to stop the oscillations

from smearing into one another. If the sampling rate is not high enough to separate out the individual

oscillations then the time waveform will take on the form of a “pulse”. Consider table 2, above: Note that

if the individual oscillations are smeared into one pulse in the time domain, then the FFT cannot have

relevant frequency domain data. In some cases it is possible to get a quasi-triangular pulse which, when

133

Vibration Analysis on Rotating Equipment

passed through the Fourier transform, will give a frequency response which looks good but is, in fact, totally

meaningless.

134

Vibration Analysis on Rotating Equipment

At certain times it is very useful to be able to calculate the natural frequencies of certain components of the

system – especially the supporting structure. The structure can usually be broken down into a number of

components which can be calculated quite easily. This section gives guidelines for the calculation of

natural frequencies for simple structural elements.

Uniform Beams

One of the most common sources of resonant vibration is the structural supporting beam.

Table 3 Natural frequency calculation of uniform beams

135

Vibration Analysis on Rotating Equipment

The notation a1 through a5 is for each vibration mode. For example, calculating a1 for a clamped-clamped

beam will return the natural frequency of the first vibration mode, calculating a2 will return the frequency

for the second mode etc.

I

ωn a 1 . E. Equation 13 General formula - natural frequency of beams

µ .l

4

The difficulty in making this type of calculation is getting the values for E, I and µ.. Luckily, however,

standard text books such as the “Machinery’s Handbook” have all of the required information for different

types of beams.

• For low carbon steels E (Young’s Modulus) can be estimated at 30,000,000 psi.

• The weight per unit length is a part of the beam identifier. For example a wide-flange section beam

designated “W 18 x 64” has a nominal depth of 18” and nominal weight of 64 pounds per foot length.

This is µ in our calculation above.

• The moment of inertia I may be calculated but it is far easier to lift the value from tables as in the

example below:

Table 4 Standard values for uniform beams

136

Vibration Analysis on Rotating Equipment

Notice that there are two values if I. There is a value for each of the two non-axial directions (vertical and

horizontal) so we can calculate the natural frequencies for each of these directions. The way that the

example above is drawn, the section X-X will give us the moment of inertia horizontally and the section Y-

Y will give the moment of inertia vertically.

Let us calculate the natural frequency for a low carbon steel, “clamped-clamped” beam, 12 feet long,

designated W18x106.

I

ω n a 1 . E.

µ .l

4

E = 30,000,000

I (section Y-Y) = 220 in4

µ = 106 lb./ft = 8.833 lb. / in

a1 = 22.0 (from table 2)

l = 12 feet = 144 inches

= 29.00117 rad/s

= 4.616 Hz

Notice that we have been consistent with our units – we have only used pounds and inches in the equation.

Whenever you are given properties in other units, you must convert to the units for which your formula is

designed.

Plates

The procedure for calculating the natural frequency of any component is similar to that of uniform beams

except that the formula varies from case to case. For plates we should consider

• circular or rectangular plates

• point load or uniform load

• simply supported edges or clamped edges

In 7Hartog’s “Mechanical Vibrations” we have a number of formulae for different applications in the

appendix.

For a square plate, all edges clamped, length of side t, the fundamental mode is

D

36.

µ1

ωn

2

l

Where µ1 = mass per unit length

And D = the plate constant

7

Available in Borders bookstore. $12.95. (DOVER 0-486-64785-4)

137

Vibration Analysis on Rotating Equipment

3

t

D E.

2

12 1 µ

Where t = thickness (inches)

µ = Poisson’s ratio (≈ 0.3)

For a low carbon, steel plate, ¼” thick, the plate constant D will be 4.293 x 104

So that for a square plate, clamped at the edges, of side 6 feet we would have

ωn = 0.108 Hz

(Don’t forget to divide radians/second by 2π to convert to Hz)

For a plate of equal side = 3 feet

ωn = 0.611 Hz

If the same 3 foot plate were ½” thick:

ωn =1.221 Hz

If your plate is clamped at each end but free on the other two sides, then treat the plate as a

clamped/clamped beam.

138

Vibration Analysis on Rotating Equipment

To take any structure out of resonance we must do one of two things:

• Isolate the structure or component from the exciting force, or

• Change the natural frequency of the structure or component

In the case where it is not feasible to change the natural frequency we must isolate.

Vibration isolators

This section explores some of the isolation materials that we can use.

The following 8table suggests some appropriate damping ranges for popular vibration isolation materials:

Expected

Steel Springs 1.2 to 3.8 Hz 3.5 mils p-p

72 to 228 cpm

Rubber 4.0 to 10.5 Hz 0.66 mils p-p

240 to 630 cpm

Cork 8.0 to 12.6 Hz 0.058 mils p-p

480 to 756 cpm

Felt 19.6 to 30 Hz 0.045 mils p-p

1176 to 1800 cpm

It should be noted that these ranges are for virgin materials and many manufacturers offer specialist

materials which cover much wider ranges than shown on this list. Vibration isolation is based on installing

machinery on springs or resilient material of uniform stiffness. The types of spring material which are used

most often are rubber and steel. Another alternative is air springs.

Cork was formerly used as a "spring" material and its elastic properties were based on compression and

expansion of the air locked inside the cork. However, the locked-in air is eventually pressed out, thus

impairing the isolation result and causing it to have no effect at all after a period of use. Cork does not

permit a large spring travel, which is why it is no longer used in more demanding applications. The same

applies to felt, which has more or less the same kind of properties.

Steel springs are normally used in the form of coil springs or leaf springs. The benefit of these is that they

permit relatively high deflections, but their disadvantage is that they provide very little damping. Because of

this, excessive movement occurs when running the machine through the resonance range, and more often

than not special devices have to be installed in order to limit the deflections.

8

Taken from information prepared by Larry Riley (3/6/96)

139

Vibration Analysis on Rotating Equipment

Compared with steel, rubber has high internal damping, it also absorbs noise, has good elasticity and

chemical resistance; and these are properties which make rubber particularly suitable as a spring material.

The stress-strain diagram illustrates clearly the difference in the strain properties of rubber and steel. Figure

1 illustrates in diagram form a tensile test on steel, and figure 2 illustrates the corresponding test on rubber.

The stress strain curve for steel is linear up to the proportionality limit, and in this range Hooke's law

applies. This states that stress is proportional to strain, and the proportionality constant is equal to the

modulus of elasticity E. Steel is elastic up to the elastic limit and restores to its original shape (this value is

less than 0.1 %). At the tensile yield limit, or yield point, which is just above the proportionality limit, the

material is extended without the tensile force increasing. The maximum tensile force is reached at the

ultimate strength.

The tensile sequence for rubber is quite different. Rubber achieves high strain levels even at very low

tensile stress. The increase in load is neither linear; in other words, rubber does not "stretch" proportionally

to the load. Consequently, the modulus of elasticity is not constant, and Hooke's law does not apply in this

case apart from approximately in a small strain range of the magnitude of 5-10 %. There is no yield point,

but the stress increases continuously until there is abrupt failure.

Springs

Springs have long been a favorite with fan manufacturers for separating fan vibration from the floor.

The principle relating to vibration-isolating with springs is that they are placed between the machine and the

base or plinth. To ensure effective isolation, the springs must be calculated very accurately, otherwise the

result could be impaired performance. In favorable cases, the force can be reduced to only 2 or 3% of the

force of a rigidly mounted machine. In such cases, you could say that the vibrations are practically

140

Vibration Analysis on Rotating Equipment

eliminated. The relative magnitude of the transmission of force depends entirely on the tuning ratio Z. If Z

is high, the force transmission percentage will be small.

The magnification factor B, that is, the change in force acting on the base of a machine supported on springs

compared with a machine which is rigidly mounted, conforms to the formula:

B=

Equation 14 Amplitude magnification due to springs

Where D = damping

Z = ratio of actual frequency to resonance frequency

B = magnification factor

The factor D depends on the internal damping of the spring material. In rubber D has the value 0.04 - 0.1,

depending on the hardness of the rubber. The term “4D2 Z2” can generally be neglected completely except

in the resonance range, that is, when Z=1. If Z=1, that is, the machine speed (rpm) = the natural vibrations

of the system, it is said that there is resonance, and the vibrations will be infinitely large if there is no

damping. Here, then, a rubber spring has a distinct advantage over a steel spring, which has minor internal

damping and in which the amplitude, in theory, grows to a very high value in the resonance point.

As can be seen in figure 10, B (at Z=D√2) has dropped to 100% and when Z is further increased, B drops

rapidly. Vibration isolation is therefore of significance first when the operating frequency considerably

exceeds the natural frequency. For practical applications, Z should be between 3 and 5, which means that 88

and 96 % of interference forces are eliminated.

As a rule you usually know the operating speed of a machine (interference frequency). If the system's

natural vibration coefficient can be altered in any way, and thereby influence Z, then it is possible to change

the force transmitted. This is exactly what happens when vibration isolation is achieved i.e. by means of

springs. The low elasticity and shear moduli of rubber are used to achieve a low natural frequency.

To sum up, transmission of vibration forces can be affected in three ways:

1) Rigidly mounted machines transmit vibration forces in unchanged form to the base, which is therefore

forced to be a part of the movement of the machine. The magnification factor can be regarded as being

100%.

2) In the case of an unsuitable spring system, the magnification factor can increase considerably and may

amount to several hundred per cent.

3) The force transmission percentage is reduced substantially by correct calculation and suitable

mountings being installed between the machine and base. Typical reductions can be from 100 down to

10 %, but in favorable circumstances can be as low as 2 %.

All machines have more than one resonance point as, through many interacting movements, they can vibrate

in many different ways. The resonance points can be determined, but the methods of calculation are often

141

Vibration Analysis on Rotating Equipment

very difficult. Experience has shown that all resonance frequencies that can arise do not need to be

clarified. It is usually quite sufficient to calculate the most significant ones which can be determined easily.

The desired level of isolation and the interference frequency determine where the resonance frequency shall

be.

To calculate the natural frequency of a spring:

Equation 15 Natural frequency of a spring

k (linear spring) = Gd4

8nD2

where G = torsional modulus (typically 11,200,000 psi for ASTM A229 oil tempered spring steel)

d = spring wire diameter

n = number of turns

D = coil diameter

Let us calculate the response of a motor/fan system sitting on a frame which, in turn, is supported by by four

springs. To make the calculation easy we will assume that the weight of the unit is equally distributed over

the four springs.

Calculate the natural frequencies of the unit using the following assumptions:

Weight of motor and fan = 500 lbs.

Spring coil diameter = 1”

Wire diameter = 0.192”

Number of turns (spring) = 6

so:

k = (11,200,000 * 0.1924)/(8*6*12)

= 317.089 lbs./inch

ωn = (317.089*4/500)0.5/2π

= 0.2535 Hz

Rubber

Compared with other engineering materials, rubber is very ductile. In some cases, the strain may be higher

than 1000 %, and by far the highest proportion of this strain is elastic. Metals, on the other hand, have very

small strains below the elastic limit. Compared with metals, the tensile strength of rubber is rather low,

however. The maximum level that can be achieved with rubber is 25-30 MPa. However, because of the

high strain, rubber has a very large work absorption capacity compared with the best grade of steel. If a

material is subjected to a load below the elastic limit, the deformation will, according to Hooke's law, be

proportional to the load. This does not apply to rubber under tension or compression. This means that

rubber does not have any constant tensile or compression module of elasticity. Metals will normally be

softer towards the end of a tensile test, while the opposite is often the case with rubber. Rubber does not

have any yield point, and the modulus is increased until there is abrupt failure.

142

Vibration Analysis on Rotating Equipment

High elastic ductility is therefore the most pronounced feature of rubber. Just how easy it is to deform

rubber is shown by the fact that the modulus of elasticity of compression for rubber within the normal

hardness range is between 2 and 12 MPa; while the modulus of elasticity of steel is 210,000 MPa. This

means that soft rubber is about 100 000 times softer than steel.

Damping capacity is another important feature of compounded rubber. This is of particular importance

when operating a machine that is supported on springs through the resonance range. The resonance

deflection with rubber springs is only 1/5 to 1/50 compared with the deflection when using steel springs

with the same stiffness, see figure 17.

material with different internal damping

With a spring made of natural rubber working under compression or shear load, the direct loss of energy is

between 6 and 30% depending on the hardness of the rubber. This energy loss is such that it is possible in

many cases to use rubber springs as dampers. But care must be taken when it comes to damping in a rubber

element. If the element works with high amplitudes, a substantial amount of energy is converted into heat,

and the heat which is generated may lead to the rubber element being destroyed.

damping properties of rubber. The elliptical area

indicates the loss off energy

In the case of single impact, the vibrations sequence will be as shown in figure 19. The left-hand curve

represents a steel spring, while the right-hand curve represents a rubber spring. These two curves clearly

show just how quickly the vibrations degenerate in the rubber, while in steel springs they diminish slowly.

As sound-insulating material, rubber is one of the very best. The effect of sound isolation increases with the

for steel and rubber springs

thickness of the rubber. Rubber is an excellent absorber of impact sound, which occurs in foundations,

floors, buildings, etc.

High chemical resistance is yet another valuable feature when using rubber. Steel, on the other hand, can

rust easily when exposed to air and acids leading to fatigue problems. Rubber is fully resistant to moisture

and common acids, and at normal temperatures, does not give rise to ageing problems. The highest

temperature to which natural rubber should be continuously exposed is about +75°C (167°F). At extremely

143

Vibration Analysis on Rotating Equipment

low temperatures, -30°C (-22°F) to -40°C (-40°F), rubber stiffens and becomes rigid. At even lower

temperatures it becomes hard, brittle and non elastic (though this is reversible).

Rubber mountings may sometimes have to be installed at places where they can be exposed to the risk of

coming into contact with oil. The harmful effects of oil can often be avoided through the design of the

mounting or by fitting a mechanical shield. An alternative would be to use oil-resistant rubber material, but

compared to natural rubber it has inferior damping properties.

When calculating compression characteristics of rubber, it should be noted that the deflection is not directly

proportional to the load, as the modulus of elasticity in compression increases with the degree of stress.

The modulus of shear, however, remains constant for normal stresses. The factor with the most effect on

stiffness calculation is the ratio between loaded and free surface area of rubber - this is the so-called shape

factor (often designated S). With thin rubber sections, a very high modulus of elasticity can be achieved.

Equation 16 Modulus of elasticity for rubber

E8 = E0 (1+2.const.S2)

In other respects, the stiffness of a rubber spring is determined by the dimensions and the hardness of the

rubber.

hardness and shear modulus

Figure 20 illustrates the relationship between rubber hardness and shear modulus. Figure 21 shows the

dependence of the bulk modulus on the shape factor. The latter curve applies at 10% deformation.

modulus upon the shape factor

Rubber at a shape factor of 0.25 is about 6-8 times softer when in shear than when in compression (for the

same rubber hardness). Since only 3-4 times the stress value in compression can be considered, it may be

said that rubber is best used in shear to achieve large deflections and good isolation properties, particularly

at low interference frequencies. The given modulus values apply at a static compression test where the end

surfaces can not slide, but when rubber is subjected to rapid vibrations it appears to be stiffer. Thus a

rubber mounting displays static and dynamic spring constants. The dynamic spring constant is dependent

144

Vibration Analysis on Rotating Equipment

on the frequency and amplitude. The effect of frequency, however, is small within the ranges in which

modern vibration mountings are most effective: from about 13 to 400 Hz.

The effect of amplitude is such that rubber will be stiffer the lower the amplitude. The ratio between

dynamic and static stiffness is dependent on the material and the geometrical shape of the rubber body in

the mounting.

One important question is how large a continuous stress can rubber be subjected to? When it comes to

shear, normally 0.25 MPa is considered to be a normal stress factor, but in some instances it can go as high

as 0.5 MPa without problems. Higher shear stresses are not suitable as they change to tensile stress.

Experience has shown that tensile loads are detrimental to rubber bonded designs. If subjected to high

shear stresses, rubber should be pre-compressed. When it comes to compression, 1 MPa is a normal stress

for the rubber bonded to metal with the restriction that the deformation will not exceed 20-25% of the

rubber thickness.

With hard rubber, we can except the stress to increase slightly. The fact that the element is made of rubber

and bonded to metal does not represent any restriction in the use of the rubber's strength properties; quite

the reverse, in fact. The bond between the rubber and the metal is often stronger than the rubber material

itself. If the area under load is large in relation to the free area (high shape factor), the modulus of elasticity

will be large and the deformation small. Higher compression stresses can then be utilized. Examples of

applications are bearings for bridges and elements for vibration-isolation of buildings. The surface pressure

here may be as high as 15-20 MPa.

One of the reasons why the aforementioned relatively low stresses for normal rubber mountings must be

taken into consideration is that the rubber element will be permanently deformed if it is subjected to too

high a load. This is called "compression set". The most important factor for reducing compression set is

that the rubber is of a high quality and correctly vulcanized. Moreover, it is essential that the rubber spring

is not exposed to high temperatures.

As we mentioned at the beginning of this section, if it is not feasible to isolate the source from the structure,

we must de-tune the structure so that the natural frequency is removed from the forcing frequency or

somehow reducing the amplitude. This can be achieved by changing the mass, the stiffness or the damping.

Consider the generalized formulae:

ωn = k √(Stiffness/Mass) and

Ra = k (Forcing_Function/Damping)

Where ωn = natural frequency

Ra = amplitude of vibration

k = a constant

Let us consider what effect each of these three factors will have on the resonant condition of the

machine/structure system.

Damping

Looking at the formulae above we see that changing the damping will have zero effect on the natural

frequency of the system. In the real world, damping is usually associated with friction, either at the

molecular level or because of friction between machine components. The damping qualities of a structure

control the resonant response. The higher the damping effect the lower the amplitude of vibration will be as

a result of a resonant condition. For example a spring is a good example of a low damped system, a

relatively small amplitude of input results in a large amplitude response. A shock absorber is a good

example of a highly damped system where a large input results in a small response.

In order to have a resonant condition we need to have the right combination of stiffness and mass to create

the natural frequency and a source vibration that matches the natural frequency. The damping controls the

145

Vibration Analysis on Rotating Equipment

amplitude response. In small systems it is sometimes feasible to add a damper such as a shock absorber, but

usually we cannot change the damping of the system significantly.

Remember the formula for natural frequency?

I

ωn a 1 . E.

µ .l

4

Remember that µ is the mass per unit length and l is the length. So increasing the mass of the structure will

lower the natural frequency of the structure. Adding mass to a structure is often a very easy way of

checking your analysis of the problem. Just by adding sand bags to a structure we can check how much

weight we need to add to the structure to remove the natural frequency away from the exciting force

(without adding stiffness).

Take care when adding the finished mass (often in the form of steel) that you do not increase the stiffness of

the structure.

I

ωn a 1 . E.

µ .l

4

The combination of E and I may be considered to be the stiffness of the beam. Examining the formula it is

obvious that increasing the stiffness will increase the natural frequency of the beam. But the ratio of

increase is not linear. In other words, increasing the stiffness by 10% will increase the natural frequency by

only 3.16% (√10).

By increasing the value for I we can increase the natural frequency of the system. If stiffness is added to the

machine, great care should be exercised. In the first place, only stiffen in the direction which has a problem

with the natural frequency being co-incident with the exciting force.

• If you stiffen in places where it is not required, you could end up with a totally new set of

resonance problems.

• As you stiffen the structure, you will add mass. But additional mass reduces the natural frequency.

It is very possible that the additional stiffeners will increase the natural frequency because of their

stiffening properties but reduce the natural frequency due to the extra mass. In other words you

have done a lot of work for no change in the vibration characteristics of the machine.

If you think that adding stiffness is the correct solution for the resonance problem, be very careful about the

mass that is added. Additional stiffness can often be added by using turn-buckles at a very small mass

penalty.

146

Vibration Analysis on Rotating Equipment

Sources of vibration

Exposure to localized vibration occurs when a part of the body comes in contact with a vibrating object,

such as power hand tools. Localized vibration from hand-held power tools can increase grip force and

symptoms of hand-arm vibration syndrome (HAV) and carpal tunnel syndrome. Exposure can occur while

holding:

• vibrating tools (e.g., pneumatic and electric rotary tools such as grinders [380 m/sec]),

• impact tools such as chippers (425 m/sec) and riveters, and

• gasoline powered tools such as chain saws (70 m/sec).

Guidelines established by the American Conference of Government Industrial Hygienist (ACGIH, 1993)

suggest that hand vibration in excess of 12 mm/sec should have exposures of less than 1 hour. (The tool

manufacturer should be contacted to obtain tool operation characteristics.) The ACGIH guidelines are based

on symptoms of white finger or Raynaud's phenomenon. ACGIH and other agencies that recommend

vibration guidelines update their guidelines when new information becomes available. •Holding vibrating

work pieces such as when grinding or polishing. •Holding vibrating controls (e.g., lawn mowers).

Exposure to whole-body vibration can also occur while standing or sitting in vibrating environments or

objects, such as trucks or heavy machinery. Prolonged exposure to whole-body vibration has been

associated with back and neck musculoskeletal disorders.

Sources of whole body vibration include virtually all modes of transportation (the jarring and jolting

associated with off-road vehicles, for example). Even a building's vibration may reach levels of concern

(such as in process control rooms or at machine operators' control platforms).

Some examples of whole body vibration sources include the following:

• on-road and off-road vehicles (buses, trucks, tractors, dozers and construction, forestry or mining

equipment);

• rail vehicles;

• aircraft (helicopters, fixed-wing aircraft);

• sea vessels (hovercraft, ships); and

• buildings (generators, pumps, heavy machinery, ventilation systems).

The initial effects of whole body vibration are discomfort and irritation. This can contribute to fatigue and a

reduced ability to perform certain tasks. Prolonged or extreme exposure to whole body vibration has been

associated with gastrointestinal disorders, neural disorders, low back pain and spinal degeneration. In many

cases, vibration-related syndromes are aggravated by other occupational factors (work history, skill level,

work environment, posture) and non-occupational factors (age, health, fitness).

Frequency ranges

The specific effects also depend on the vibration frequencies to which the worker is exposed. Very low

frequency vibration (less than one Hz or cycle per second) can result in motion sickness; frequencies

between four and eight Hz are likely to affect the gastrointestinal and spinal systems; higher frequency

vibration (15 to 40 Hz) can interfere with vision. Whole body vibration usually originates from a moving

vehicle or from the operation of heavy machinery. Much can be gained by routine maintenance of the

147

Vibration Analysis on Rotating Equipment

vibration source and existing isolation systems. For more persistent problems, vibration measurements will

likely be necessary to make an educated selection of appropriate vibration control strategies.

15 – 40 Hz Vision difficulties

Whole body vibration transmitted to vehicle operators may be reduced by proper maintenance of roadways,

vehicle maintenance and control of vehicle speed. Tire pressure, track and suspension systems, as well as

seating and engine mounts should be routinely checked and serviced when required. Vehicle cab isolation is

available in some vehicles and can effectively reduce vibration exposure. If vibration control available

through vehicle and roadway maintenance is insufficient, improved seating selection may provide further

protection for the worker.

Generally, two types of seating systems are considered when addressing a vibration problem: conventional

foam seats or suspension seats. Conventional foam seats tend to amplify vibration frequencies in the range

of greatest susceptibility for the human body. Anti-vibration suspension seats include a damper and a spring

in the seat system. The spring may be constructed of steel or it may be a column of air. There are large

differences between seats and the performance of an individual seat may change during its lifetime. It is

important to know the vibration levels and frequencies in the vehicle in order to select an appropriate

suspension seat.

Control room vibration may result from machinery operation or heavy vehicles traveling in the vicinity.

This vibration may be addressed in a number of ways:

1. reduce vibration at the source (in the case of heavy machinery this may be accomplished by balancing

moving parts)

2. modify a vibrating system to reduce the likelihood of excessive vibration due to resonance (in the case

of a cover panel, this may be accomplished by •

2.1. stiffening the system by welding a steel section to it);

2.1.1. increase the damping in the system (damping materials may be coated on the system);

or •

2.1.2. isolate the system from the vibration source (mount the system on anti-vibration

mounts).

148

Vibration Analysis on Rotating Equipment

Glossary

A

Accelerometer

Transducer for measuring vibration in the form of acceleration. It is one of the most rugged transducers

for vibration measurements and also has the widest frequency range.

AC

Literally Alternating Current but often used to imply a time-fluctuating signal.

Acceleration

The rate of change of velocity in inches/second/second or G’s (acceleration due to gravity is 1 G). This

is a good indicator of the forces inside a machine since F = m.a (Newton’s 2nd law).

ADC

Acronym for Analog to Digital conversion

Alarm

Alarms are used to identity specific operating conditions or to define the boundaries between safe and

unsafe conditions. When an amplitude reaches or exceeds the alarm amplitude then the software

automatically generates an exception report.

Anti Aliasing

Nyquist's theorem says that as long as the sampling rate is greater than twice the highest frequency

component of the signal, then the sampled data will accurately represent the input signal. Certain

analyzers pass the input signal passes through an analog filter which attenuates all frequency

components above Fmax by 90 dB to make sure that Nyquist's theorem is satisfied. This is the anti-

aliasing filter.

ASCII

American Standards Code for Information Interchange.

Attenuation

The weakening of a signal by distance from the source or a mechanical interface.

Averaging

In general, averaging many spectra together improves the accuracy and repeatability of measurements.

B

Band Pass Filter

A measurement filter that removes data below the low cutoff frequency and above the high cutoff

frequency. The band pass filter only passes the data between the cutoff frequencies.

Bin

See “resolution”

Blackman-Harris Window

149

Vibration Analysis on Rotating Equipment

The Blackman-Harris window is a very good window to use with the spectrum analyzer. It has

better amplitude accuracy (about 0.7 dB) than the Hanning, very good selectivity and the fastest

filter rolloff. The filter is steep and narrow and reaches a lower attenuation than the other

windows. This allows signals close together in frequency to be distinguished, even when their

amplitudes are very different.

C

Calibration

The process of multiplying or dividing the voltage signal from a transducer by a factor that

represents a specific engineering quantity. e.g. 100 mV/G for an accelerometer.

Coherence

Coherence measures the percentage of power in channel 2 which is caused by (phase coherent

with) power in the input channel. Coherence is a unitless quantity which varies from 0 to l. If the

coherence is 1, all the power of the output signal is due to the input signal. If the coherence is 0,

the input and output are completely random with respect to one another. Coherence is related to

signal to noise ratio (S/N) by the formula:

S/N = γ2/(l -γ2)

where γ2 is the traditional notation for coherence.

Correlation

The two channel analyzer may also compute auto and cross correlation. Correlation is a time

domain measurement which is defined as follows:

Auto Correlation(τ) = ∫x*(t)x(t-τ)dt

Cross Correlation(τ) = ∫x*(t)y(t-τ)dt

where x and y are the channel 1 and channel 2 input signals and the integrals are over all time. It is

clear that the auto correlation at a time t is a measure of how much overlap a signal has with a

delayed-by-t version of itself, and the cross-correlation is a measure of how much overlap a signal

has with a delayed-by-t version of the other channel. Although correlation is a time domain

measurement the some analyzers use frequency domain techniques to compute it in order to make

the calculation faster.

Cross Spectrum

The cross spectrum is defined as:

cross spectrum = FFT2 conj(FFT1)

The cross spectrum is a complex quantity which contains magnitude and phase information. The

phase is the relative phase between the two channels. The magnitude is simply the product of the

magnitudes of the two spectra. Frequencies where signals are present in both spectra will have

large components in the cross spectrum.

CPM

Abbreviation for Cycles per Minute - the most common format for displaying frequency in

vibration analysis.

150

Vibration Analysis on Rotating Equipment

D

Database

A collection of information files that are ties together by a common topic. A database allows rapid

access of the database files.

Data Collector

Hardware device for collecting vibration data off line.

Decibel (dB)

A convenient unit for displaying logarithmic data.

dB = 10 log (X/Xο)

where X or Xο is

a power or squared quantity

or dB = 20 log (Y/Yο)

where Y or Yο is a linear quantity.

Decibels are usually referenced to a standard. e.g. 0.1 GdB re 0.001G.

DC

Literally Direct Current. In PdM terms this may be used to check the accelerometer or for a

process measurement.

Differentiate

Differentiation changes displacement to velocity and velocity to acceleration relative to time using

the formulae:

Acceleration = Velocity /(2.π.f)

Velocity = Displacement /(2.π.f)

or Acceleration = Displacement /(2.π.f)2

where f = frequency

n.b. remember 1G = 386 in/s2 or 9.806 m/S2

E

Envelope Measurements

A.k.a. demodulation (see appendix)

Exception

To have an alarm condition. Data is outside the defined “safe” area.

Exponential Averaging

Exponential averaging weights new data more than old data. Averaging takes place according to

the formula,

New Average = (New Spectrum - I/N) +(Old Average) - (N-l)/N

where N is the number of averages.

151

Vibration Analysis on Rotating Equipment

Exponential averages "grow" for approximately the first 5N spectra until the steady state values are

reached. Once in steady state, further changes in the spectra are detected only if they last

sufficiently long. Make sure that the number of averages is not so large as to eliminate the changes

in the data that might be important.

F

Flattop Window

The Flattop window improves on the amplitude accuracy of the Hanning window. Its between-bin

amplitude variation is about 0.02 dB. However, the selectivity is a little worse. Unlike the

Hanning, the Flattop window has a wide pass band and very steep rolloff on either side. Thus,

signals appear wide but do not leak across the whole spectrum.

FFT Spectrum Analyzer

FFT Spectrum Analyzers take a time varying input signal, like you would see on an oscilloscope

trace, and compute its frequency spectrum.

Fourier or FFT

Fourier's theorem states that any waveform in the time domain can be represented by the weighted

sum of sines and cosines. The FFT spectrum analyzer samples the input signal, computes the

magnitude of its sine and cosine components, and displays the spectrum of these measured

frequency components.

Frequency

The rate at which periodic events happen. Typical units are Hertz (Hz), Cycles per Minute (cpm)

or orders (multiples of run speed).

Fundamental Frequency

Primary frequency, such as operating speed, to which other frequencies may be referred back.

H

Hanning Window

The Hanning window is the most commonly used window. It has an amplitude variation of about

1.5 dB (for signals between bins) and provides reasonable selectivity. Its filter rolloff is not

particularly steep. As a result, the Hanning window can limit the performance of the analyzer

when looking at signals close together in frequency and very different in amplitude.

Harmonic

Frequencies at direct multiples of a fundamental frequency. The fundamental is not necessarily the

run speed.

Hertz (Hz)

Common frequency units in cycles per second. Named after, Heinrich Rudolf Hertz (1857-94).

German physicist born in Hamburg and educated at the University of Berlin. From 1885 to 1889

he was a professor of physics at the technical school in Karlsruhe and after 1889 a professor of

physics at the university in Bonn. Hertz clarified and expanded the electromagnetic theory of light

that had been put forth by the British physicist James Clerk Maxwell in 1884. Hertz proved that

electricity can be transmitted in electromagnetic waves, which travel at the speed of light and

which possess many other properties of light. His experiments with these electromagnetic waves

led to the development of the wireless telegraph and the radio. The unit of frequency that is

measured in cycles per second was renamed the hertz; it is commonly abbreviated Hz.

152

Vibration Analysis on Rotating Equipment

A measurement filter that removes data below its low cutoff frequency.

I

Input Couple

Measurements may be AC or DC coupled. Use DC coupling for process measurements and use

AC for vibration measurements. DC coupling includes both DC and AC signals whereas AC

coupling does not include any DC offsets.

Integrate

Integration is the opposite to differentiation and changes acceleration to velocity and velocity to

displacement where A = V x (2.π.f)

K

Kaiser

The Kaiser window, which is available on IRD analyzers, combines excellent selectivity and

reasonable accuracy (about 0.8 dB for signals between exact bins). The Kaiser window has the

lowest side-lobes and the least broadening for non-bin frequencies. Because of these properties, it

is the best window to use for measurements requiring a large dynamic range.

L

Leakage

Errors resulting from Fourier transforming nonperiodic time domain data. This effect is reduced

by using windows such as Hanning or Flat Top.

Linear

Linear scaling displays all data on an equal basis. Small signals may be hard to detect with linear

scaling but become visible with logarithmic scaling.

Linear Averaging

Linear averaging combines N (number of averages) spectra with equal weighting in either RMS,

Vector or Peak Hold fashion. This type of averaging is useful for eliminating transients.

Line of Resolution

see “resolution”

Logarithmic

Although the linear magnitude scale is used most often for displaying spectra, another way of

displaying amplitude is the Log Magnitude. The Log Mag display graphs the magnitude of the

spectrum on a logarithmic scale using dBEU (Engineering Units) as units.

The 16 bit analyzer has a dynamic range of about 90 dB. below full scale. Imagine what

something 0.01% of full scale would look like on a linear scale. If we wanted it to be 1 inch high

on the graph, the top of the graph would be 833 feet above the bottom - It turns out that the log

display is both easy to understand and shows features which have very different amplitudes clearly.

The real and imaginary parts are always displayed on a linear scale. This avoids the problem of

taking the log of negative voltages.

A measurement filter that removes data above the cutoff frequency.

153

Vibration Analysis on Rotating Equipment

O

Octave Analysis

The magnitude of the normal spectrum measures the amplitudes within equally divided frequency bins.

Octave analysis computes the spectral amplitude in logarithmic frequency bands whose widths are

proportional to their center frequencies. The bands are arranged in octaves with either 1 or 3 bands per

octave (1/1 or 1/3 octave analysis). Octave analysis measures spectral power closer to the way people

perceive sound, that is, in octaves.

The center frequency of each band should be calculated according to ANSI standard S1.11 (1986).

Typically the shape of each band is a third-order Butterworth filter whose bandwidth is either a full, 1/3,

or 1/12 octave. The full octave bands have band centers at:

Center Freq: = 1 kHz x 2n

The 1/3 octave bands have center frequencies given by:

Center Freq: = 1 kHz x 2(n-30/3)

Operating System

A form of software that controls and supervises how the computer operates. It loads programs, handles

input and output operations and accepts and executes commands issued by the user.

Orbit

The orbit is simply a two dimensional display of the time record of channel 1 vs. the time record of

channel 2. The orbit display is similar to an oscilloscope displaying a "Lissajous" figure.

Orders

A frequency axis scale which is useful for viewing data as a function of the operating speed. The first

order corresponds to the operating speed, the second order is two times the operating speed and so on.

Overall

A single value representing the vibration or some other measurement parameter. For vibration

measurements this value includes vibration at all frequencies.

Overlap Processing

What about narrow spans where the time record is long compared to the processing time which is what

we normally see when taking vibration measurements? The analyzer computes one FFT per time record

and can wait until the next time record is complete before computing the next FFT. The update rate

would be no faster than one spectra per time record. With narrow spans, this could be quite slow.

And what is the processor doing while it waits? Nothing. With overlap processing, the analyzer does

not wait for the next complete time record before computing the next FFT. Instead it uses data from the

previous time record as well as data from the current time record to compute the next FFT. This speeds

up the processing rate. Remember, most window functions are zero at the start and end of the time

record. Thus, the points at the ends of the time record do not contribute much to the FFT. With

overlap, these points are “re-used" and appear as middle points in other time records. This is why

overlap effectively speeds up averaging and smoothes out window variations.

Typically, time records with 50% overlap provide almost as much noise reduction as non-overlapping

time records when RMS averaging is used. When RMS averaging narrow spans, this can reduce the

measurement time by a factor of two.

The amount of overlap is specified as a percentage of the time record. 0% is no overlap and 99.8% is the

maximum (511 out of 512 samples re-used). The maximum overlap is determined by the amount of

time it takes to calculate an FFT and the length of the time record and thus varies according to the span.

154

Vibration Analysis on Rotating Equipment

Peak

The maximum value as seen in the time domain data. For a sine wave

peak = RMS x 1.414

Peak Hold Averaging

Peak Hold is not really averaging, instead, the new spectral magnitudes are compared to the

previous data, and if the new data is larger, then the new data is stored. This is done on a

frequency bin by bin basis. The resulting display shows the peak magnitudes which occurred in

the previous group of spectra.

Peak Hold detects the peaks in the spectral magnitudes and only applies to Spectrum, PSD, and

Octave Analysis measurements. However, the peak magnitude values are stored in the original

complex form. If the real or imaginary part or phase is being displayed for spectrum

measurements, the display shows the real or imaginary part or phase of the complex peak value.

Phase

In general, phase measurements are only used when the analyzer is triggered. The phase is relative

to the pulse of the trigger. The phase is displayed in degrees or radians on a linear scale, usually

from -180 to +180 degrees. The phase of a particular frequency bin is set to zero in most analyzers

if neither the real nor imaginary part of the FFT is greater than about 0.012% of full scale (-78 dB

below f.s.). This avoids the messy phase display associated with the noise floor. (Remember, even

if a signal is small, its phase extends over the full 360 degrees.)

Power Spectral Density (PSD)

The PSD is simply the magnitude of the spectrum normalized to a 1 Hz bandwidth. This

measurement approximates what the spectrum would look like if each frequency component were

really a 1 Hz wide piece of the spectrum at each frequency bin.

What good is this? When measuring broad band signals such as noise, the amplitude of the

spectrum changes with the frequency span. This is because the line width changes so the

frequency bins have a different noise bandwidth. The PSD, on the other hand, normalizes all

measurements to a 1 Hz bandwidth and the noise spectrum becomes independent of the span. This

allows measurements with different spans to be compared. If the noise is Gaussian in nature, then

the amount of noise amplitude in other bandwidths may be approximated by scaling the PSD

measurement by the square root of the bandwidth. Thus the PSD is displayed in units of V/√Hz or

dBV/√Hz.

Since the PSD uses the magnitude of the spectrum, the PSD is a real quantity. There is no real or

imaginary part or phase.

R

Rayleigh’s Principle

This principle states that ∆f is the lowest measurable frequency for a time record length T:

∆f = 1/T

What is real time bandwidth? Simply stated, it is the frequency span whose corresponding time record

exceeds the time it takes to compute the spectrum. At this span and below, it is possible to compute the

spectra for every time record with no loss of data. The spectra are computed in "real time". At larger

spans, some data samples will be lost while the FFT computations are in progress.

Rectangular Window (Uniform or No Window)

155

Vibration Analysis on Rotating Equipment

The rectangular windows applies an even weighting (no window) over the time period. It does not

reduce leakage errors and should only be used for impulsive or transient data that dies out within the

time sample period.

Resolution

The accuracy of a reading based on the number of discrete values used to define it. For a frequency

domain spectrum this refers to the number of lines (or bins) of resolution that are combined to display

the spectral data. For example a spectrum of 500 Hz Fmax with 400 bins or lines would have a line

resolution of 1.25 Hz.

RMS (Root Mean Square)

The square root of the average of a set of squared values. For a sine wave RMS = Peak x 0.7071

RMS Averaging

RMS averaging computes the weighted mean of the sum of the squared magnitudes (FFT times its

complex conjugate). The weighting is either linear or exponential.

RMS averaging reduces fluctuations in the data but does not reduce the actual noise floor. With a

sufficient number of averages, a very good approximation of the actual random noise floor can be

displayed.

Since RMS averaging involves magnitudes only, displaying the real or imaginary part or phase of an

RMS average has no meaning. The RMS average has no phase information.

Route

For off-line data collectors this is a an ordered list of points containing the sequence for collecting data.

S

Sampling Rate

The original digital time record comes from discrete samples taken at the sampling rate. The

corresponding FFT yields a spectrum with discrete frequency samples. In fact, the spectrum has less

than half as many frequency points as there are time points. Suppose that you take 1024 samples at

2560 Hz. It takes 0.4 Seconds to take this time record. The FFT of this record yields 400 frequency

points or lines, but over what frequency range? The highest frequency will be determined by the in-built

ratio of F-max to data sampling rate - 2.56. The lowest frequency is just the F-max divided by the

number of lines:

F-max = data sampling rate / 2.56

No. Of Lines = No samples / 2.56

Bin resolution = F-max / No. of lines

= (2560 / 2.56) / (1024 / 2.56)

= 2.5 Hz (the same as the lowest measurable frequency)

Everything below 2.5 Hz is considered to be DC. The output spectrum thus represents the frequency

range from DC to 1000 Hz with points every 2.5 Hz.

Sideband

A frequency which occurs either side of a fundamental frequency. Sidebands occur because of a

modulation of the fundamental by another frequency.

Spectrum

156

Vibration Analysis on Rotating Equipment

The spectrum is the basic measurement of an FFT analyzer. It is simply the complex FFT. Normally,

the magnitude of the spectrum is displayed. The magnitude is the square root of the FFT times its

complex conjugate. (Square root of the sum of the real (sine) part squared and the imaginary (cosine)

part squared). The magnitude is a real quantity and represents the total signal amplitude in each

frequency bin, independent of phase.

If there is phase information in the spectrum, i.e. the time record is triggered in phase with some

component of the signal, then the real (cosine) or imaginary (sine) part or the phase may be displayed.

The phase is simply the arc tangent of the ratio of the imaginary and real parts of each frequency

component. The phase is always relative to the start of the triggered time record.

T

Tachometer

A device for measuring the speed of rotation.

Time Record

The time record measurement displays the filtered data points before the FFT is taken For baseband

spans (spans that start at DC), the time record is a real quantity. For non-baseband spans (zooms) the

heterodyning discussed earlier transforms the time record into a complex quantity which can be

somewhat difficult to interpret.

Time Synchronous Averaging

see “vector averaging”

Transfer Function

The transfer function is the ratio of the spectrum of channel 2 to the spectrum of channel 1. For the

transfer function to be valid, the input spectrum must have amplitude at all frequencies over which the

transfer function is to be measured.

Two-Channel Measurements

Two-channel analyzers offer additional measurements such as transfer function, cross-spectrum,

coherence and orbit.

U

Uniform Window

The uniform window is actually no window at all. The time record is used with no weighting. A

signal will appear as narrow as a single bin if its frequency is exactly equal to a frequency bin. (It is

exactly- periodic within the time record). If its frequency is between bins, it will affect every bin of

the spectrum. These two cases also have a great deal of amplitude variation between them (up to 4

dB).

In general, this window is only useful when looking at transients which do not fill the entire time

record.

157

Vibration Analysis on Rotating Equipment

V

Vector (Synchronous Time) Averaging

Vector averaging averages the complex FFT spectrum. (The real part is averaged separately from the

imaginary part.) This can reduce the noise floor for random signals since they are not phase coherent

from time record to time record.

Vector averaging requires a trigger. The signal of interest must be both periodic and phase

synchronous with the trigger. Otherwise, the real and imaginary parts of the signal will not add in

phase and instead will cancel randomly.

With vector averaging, the real and imaginary parts as well as phase displays are correctly averaged

and displayed. This is because the complex information is preserved.

W

Windowing

What is windowing? Let's go back to the time record. What happens if a signal is not exactly

periodic within the time record? We said that its amplitude is divided into multiple adjacent

frequency bins. This is true but it's actually a bit worse than that. If the time record does not start

and stop with the same data value, the signal can actually smear across the entire spectrum. This

smearing will also change wildly between records because the amount of mismatch between the

starting value and ending value changes with each record.

Windows are functions defined across the time record which are periodic in the time record. They

start and stop at zero and are smooth functions in between. When the time record is windowed, its

points are multiplied by the window function, time bin by time bin, and the resulting time record is

by definition periodic. It may not be identical from record to record, but it will be periodic (zero at

each end).

In the frequency domain, a window acts like a filter. The amplitude of each frequency bin is

determined by centering this filter on each bin and measuring how much of the signal falls within

the filter. If the filter is narrow, then only frequencies near the bin will contribute to the bin. A

narrow filter is called a selective window - it selects a small range of frequencies around each bin.

However, since the filter is narrow, it falls off from center rapidly. This means that even

frequencies close to the bin may be attenuated somewhat. If the filter is wide, then frequencies far

from the bin will contribute to the bin amplitude but those close by will not be attenuated

significantly.

The net result of windowing is to reduce the amount of smearing in the spectrum from signals not

exactly periodic with the time record. The different types of windows trade off selectivity,

amplitude accuracy, and noise floor.

Several types of window functions are available including Uniform (none), Flattop, Hanning,

BlackmanHarris, and Kaiser.

158

Vibration Analysis on Rotating Equipment

Index

Decibel, 152

Demodulation, 2, 6, 25, 26, 27

—A— Developed Fatigue, 7, 67

AC, 154 Differentiate, 152

acceleration, 2, 6, 7, 13, 14, 15, 17, 22, 23, 30, 69, 70, displacement, 2, 6, 7, 15, 16, 17, 18, 22, 77, 79, 80,

71, 72, 74, 93, 94, 103, 104, 117, 118, 119, 127, 110, 117, 118, 121, 122, 127, 152, 154

150, 152, 154 Displacement, 127

accelerometer, 6, 7, 22, 23, 24, 28, 71, 87, 88, 90,

131, 151, 152 —E—

Accelerometer, 7, 41, 89, 90, 150

Accelerometers, 2, 24 eccentric rotor, 32

Aerodynamic cross coupling, 3, 60 Elasto Hydrodynamic Lubrication, 3, 62

aerodynamic forces, 3, 59 envelope, 6, 13, 27, 31, 69, 93, 103

Air, 120 Envelope, 6, 13, 93, 152

Analysis, 2, 3, 4, 9, 70, 87, 91, 92, 93, 101, 102, 103, Enveloping, 6, 27

106, 155, 156 External manifestation, 3, 92

Analyzer, 153

Armature, 2, 32

Average, 152

—F—

Averaging, 4, 100, 101, 150, 152, 154, 156, 157, 158, Failure type, 3, 91

159 fan, 6, 7, 9, 10, 11, 12, 13, 29, 45, 59, 109, 130, 140,

142

—B— fans, 9, 59, 60, 69

FFT, 2, 5, 6, 18, 19, 20, 21, 28, 36, 38, 39, 40, 41, 42,

Balancing, 3, 75, 78, 80, 82, 85 43, 99, 100, 101, 102, 107, 130, 131, 133, 153,

Base cause, 3, 91 155, 156, 157, 158, 159

bearing, 3, 6, 7, 10, 13, 14, 22, 26, 28, 29, 30, 38, 41, Filter, 154, 155

43, 45, 60, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, Firing Cards, 3, 37

72, 73, 74, 78, 91, 92, 93, 94, 95, 103, 104, 110, Flattop, 99, 100, 153, 159

111, 112, 113, 114, 115, 130 Fourier, 6, 20, 28, 134, 153, 154

Bearing, 2, 3, 6, 7, 13, 26, 45, 62, 63, 64, 65, 66, 67, Frequency, 2, 3, 4, 5, 19, 21, 34, 67, 103, 104, 115,

68, 70, 114 135, 148, 149, 153

Benzene, 119 Fundamental, 153

Bin, 157

Blackman-Harris, 99, 100, 151

Bode, 132

—G—

Boltzman, 121 Gas, 121

bowed rotor, 32 Glycerin, 119

BPIR, 13

Bump test, 129, 130

—H—

—C— Hanning, 93, 94, 99, 100, 151, 153, 154, 159

Harmonic, 108, 153

Campbell, 124 Harmonics, 120

Choking, 3, 61 Hertz, 153

Coherence, 106, 151 Hydraulic valve, 120

Comparitor Card, 3, 40, 41 hysteresis whirl, 60

Compression mode, 6, 24 Hz, 21, 26, 153, 156, 157, 158

Correlation, 106, 151

CPM, 151

Cracking, 126 —I—

Critical speed, 123 Imbalance, 2, 3, 6, 9, 45, 46, 47, 48, 49, 50, 75, 78,

104

—D— Imbalanced phases, 33

Impact hammer, 131, 132

Damping, 117, 139, 143, 145, 146 Impeller, 120

DC, 2, 3, 6, 21, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, Induction Motors, 2, 32

152, 154, 158 Integrate, 154

159

Vibration Analysis on Rotating Equipment

Rubber, 139, 140, 142, 143, 144, 145

—K—

Kaiser, 93, 94, 99, 100, 154, 159

—S—

S.C.R, 3, 38

Setup, 4, 91, 92

—L— Severity, 4, 105

Leakage, 154 Shear mode, 6, 24

Line, 154 Sideband, 158

Liquids, 118 Simple Harmonic Motion, 2, 15

Logarithmic, 154 Single Plane, 3, 79, 80, 82

Looseness, 2, 3, 6, 7, 11, 54, 55, 56, 57, 58, 104 spall, 6, 13, 14, 64, 65, 94

lubricant wedge, 62, 63, 64 spectrum, 6, 7, 9, 10, 11, 12, 13, 14, 15, 18, 20, 21,

22, 29, 30, 31, 33, 34, 36, 37, 38, 39, 40, 41, 43,

50, 69, 70, 71, 72, 73, 74, 92, 94, 99, 100, 101,

—M— 102, 103, 104, 106, 107, 108, 112, 114, 124, 130,

magnetic center, 33 131, 132, 133, 151, 153, 154, 155, 156, 157, 158,

Mass, 117, 145, 146 159

Mercury, 119 Spectrum, 2, 20, 21, 101, 106, 107, 129, 151, 152,

Misalignment, 2, 3, 6, 10, 51, 52, 53, 104, 114 153, 156, 158

Mode shape, 127 Speed of sound, 118, 119, 121

mounting techniques, 7, 89, 90 Spring, 117, 142

Stone Walling, 3, 61

Surging, 3, 60

—N—

Natural frequency, 135, 142 —T—

Newton, 8, 22, 23, 45, 117, 118, 121, 150

Nyquist, 132 Tachometer, 158

Technology, 3, 91, 92

time, 6, 15, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 29,

—O— 30, 32, 33, 34, 41, 50, 59, 64, 66, 73, 79, 84, 85,

Octave, 4, 101, 102, 155, 156 87, 92, 94, 99, 100, 101, 102, 103, 106, 109, 110,

Orbit, 7, 106, 111, 155 113, 115, 116, 117, 118, 127, 130, 131, 133, 150,

Orbits, 4, 110 151, 152, 153, 154, 155, 156, 157, 158, 159

Overall, 155 Time, 2, 4, 7, 19, 72, 73, 101, 129, 157, 158, 159

Overlap, 4, 101, 102, 155 tolerance stack up, 7, 76

Transfer Function, 106, 158

Triggering, 4, 109

—P—

Parameter, 2, 4, 22, 91, 92 —U—

PEAK, 2, 18

Peak Hold, 101, 154, 156 Uniform, 5, 99, 100, 135, 157, 158, 159

PFA, 3, 4, 7, 91, 93, 95

Phase, 4, 7, 50, 83, 84, 107, 127, 128, 129, 156 —V—

Piston pulse, 120

PRELOAD, 112 Vane pass, 120

vector, 7, 80, 81, 82, 83, 84, 101, 108, 113, 158, 159

velocity, 2, 6, 10, 11, 13, 15, 16, 17, 18, 22, 23, 31,

—R— 33, 50, 61, 64, 65, 69, 74, 92, 94, 104, 117, 118,

119, 121, 122, 123, 127, 150, 152, 154

Resolution, 154, 157

Vibration isolators, 139

resonance, 4, 7, 9, 25, 26, 28, 31, 33, 59, 69, 94, 111,

112, 113, 114, 115, 116, 118, 120, 121, 122, 123,

124, 126, 127, 128, 129, 130, 139, 140, 141, 142, —W—

143, 146, 147, 149

Resonance, 2, 4, 8, 28, 104, 123, 141 waterfall, 4, 6, 30, 31, 69, 109, 124

RMS, 2, 6, 16, 18, 19, 100, 101, 154, 155, 156, 157 waveform, 6, 7, 15, 18, 19, 20, 26, 27, 29, 32, 33, 50,

rotor bars, 32, 33, 34 69, 72, 73, 94, 99, 133, 153

Rotor Bars, 2, 34 whip, 112, 130

Route, 157 whirl, 60, 112, 124

160

Vibration Analysis on Rotating Equipment

windings, 32, 33, 71

161

- Vibration analysis lecturesUploaded byRam Warun
- Vibrations in Rotating Machinery (7th Int'l Conf.) (IME, 2000) WWUploaded byyounever123
- Vibration DiagnosticsUploaded byahadryk
- Rotating Machinery Vibration - From Analysis to Troublesh OotingUploaded byguhpjm
- 36199644-Vibration-Intermediate.pdfUploaded byJameel Khan
- Basics of Vibration AnalysisUploaded bybagusindarto
- Section II - Basic Vibration TheoryUploaded byagiba
- Vibration Training Quick ReferenceUploaded byJose G Lopez
- Vibration Adv 0402Uploaded bySajid Rafique
- Vibration AnalysisUploaded bymadan023
- VA II TrainingUploaded byjawadhussain1
- Machine Vibration ( a Beginners Guide)Uploaded bymavric444
- Vibration Analysis Handbook - James TaylorUploaded bymmbataineh
- Level II MetricUploaded byAshish Chaturvedi
- Vibration Basic 0400Uploaded bySajid Rafique
- Basic Vibration AnalysisUploaded bygpradipta
- Vibration Analysis Manual Rev0Uploaded byiradanke
- Vibration Analysis Level- 1 Updated [Compatibility Mode]Uploaded byjawadhussain1
- M. Rades - Dynamics of Machinery 3Uploaded byrades3702
- VibrationUploaded bysatswk
- Vibration Dignostic ChartUploaded byIbrahim Sammour
- Vibration Training Mobius1Uploaded byazhar.simebest
- Machinery Failure Analysis & Trouble ShootingUploaded bymsavidge50
- vibration analysisUploaded byvalentinbecerra
- Vibration Analysis of Centrifugal FansUploaded byMohd Asiren Mohd Sharif
- Machinery Malfunction Diagnosis and Correction Robert Eisenmann (1)Uploaded byvinothenergy
- Dynamic Field BalancingUploaded byMuralitcl
- CAT-IIUploaded bymustafa
- Rotating Machinery Practical Solutions to Unbalance and Misalignment - Robert B McMillan,Uploaded bySelvaraj Kuppannan
- Vibration Diagonistic Chart.pptUploaded byjawadhussain1

- Trib OlogyUploaded byronfrend
- Vibration Analysis Measurement ManualUploaded byronfrend
- Rutina de Termografia - CarroceriasUploaded byronfrend
- Rutina de Termografia - CarroceriasUploaded byronfrend
- Estimating Convective Heat Transfer CoefficientsUploaded byronfrend
- Level 3 NVQ Refinery Ops Control Room NOSUploaded byronfrend
- Level 3 NVQ Refinery Ops Control Room NOSUploaded byronfrend
- Vibration Signature Analysis PROC_V18Uploaded byronfrend
- Vibration Signature Analysis PROC_V18Uploaded byronfrend
- Thermography LimitsUploaded byronfrend
- 3906-qep-dec11Uploaded byronfrend
- Hospital heat loss analysis Report 5 Appendix 10Uploaded byronfrend
- 450OILUploaded byronfrend
- PROC_V19 PL302 downloadUploaded byronfrend
- Hospital heat loss analysis Report 1 Appendix 6Uploaded byronfrend
- RBI and FFSUploaded byronfrend
- How to Bump Test With MicrologUploaded byYuda Satria
- MV VFD SpecUploaded byronfrend
- Spec - Thermal Imaging SystemUploaded byronfrend
- avsde4Uploaded byAsim Idrees Osman
- Life Cycle CostingUploaded byronfrend
- 10.1.1.123.8421Uploaded byjavier_valero
- Guidlines Tanks of Tank InspectionUploaded byVeky Pamintu
- Integration of safety in machine designUploaded byronfrend
- DC Motor Brush LifeUploaded byronfrend
- Guardian Angel FlowchartUploaded byronfrend
- ISO 10816-1 normativa vibracionesUploaded byumendibil
- Risk Based Inspection Jacketed PlatformsUploaded byronfrend
- Pre Shutdown ChecklistUploaded byronfrend
- Body Shop Thermography ExamplesUploaded byronfrend

- 2014 CMOST PresentationUploaded byc3781794
- Magsanga TrailUploaded byGlenda Tapang
- CartomancyUploaded bylucentum56
- Board-of-Aeronautical-Engineering-SB-1.pdfUploaded byCruz Plus
- 07 Horvath Drasovean Remarks on the Connections Between Banat and the Great Hungarian PlainUploaded byАлександар Јаблановић
- 76561_truvision_family_bro_v3.pdfUploaded byWERMERM
- cdj-100sUploaded bymarcelo_eletronica
- Al Bielek & His Claims About the Philadelphia ExperimentUploaded byJason Lamb
- obesity and malnutritionUploaded byapi-346311171
- Sudden Cardiac DeathUploaded byforensicmed
- pone.0022577Uploaded bySiva Karthik
- Physics Project on Optical Fiber and Its ApplicationsUploaded byShlok Sah
- Thermal Engg II Sem(3)Uploaded byGaddam Ravikumar
- cue sheetUploaded byapi-426578805
- Sump Pump P-534 Service ManualUploaded bykaash028
- 95 Scs 4 Sr. No. 9Uploaded byBhupender Ramchandani
- 195434301-Philippines-Ancient-Biblical-History.pdfUploaded byHope Earl Ropia Borongan
- Specs VaibhavUploaded byNishit Chadha
- HP2-896Uploaded byJulces Gutierrez Quintero
- Natural Fibre CompositeUploaded byAdika
- avensis_2006_tns600Uploaded byExpobiro Srbija
- Bmw PresentationUploaded bypdamodar2007
- X-Series 01.Intro v.3 en Low ResUploaded byChandra Nuriko
- 2018-2019 science pacing guideUploaded byapi-417439119
- Gardner Denver PZ-11revF3[1].PptUploaded bypipeo25
- NoahStJohn_AFFORMATIONS_First2Chapters.pdfUploaded bybigjlk
- Raincheck BrochureUploaded byNandini Lochan Sharma
- Marketing Poultry BirdsUploaded byabacusdotcom2964
- Chemical Evidence for Wine Production Around 4000 BCEUploaded byEduardo Vilugrón
- Pembuktian Pythagoras.docxUploaded byDini Putri Utami