Vibration Diagnostics 3a6j2z

|

u u

=

AGENDA

What is Vibration



Vibration Characteristics



Vibration Instruments



Vibration Measurement



Vibration Analysis Steps



Imbalance



Misalignment



Looseness



Antifriction Bearings



Gear Problems



Electrical Defects



Fluid Vibrations



Case Studies

Vibration is the response of a system to an internal or external stimulus causing it to oscillate or pulsate.

Vibration has three measurable characteristics Amplitude Frequency Phase u Amplitude tells us the magnitude of vibration. Amplitude can be measured in displacement (mils or r R velocity (in/sec or mm/secR or acceleration (gR. Frequency less than 600 M

=

Use Displacement

Frequency is 600 to 120 000 M

=

Use Velocity

Frequency over 120 000 M

=

Use Acceleration

Displacement (proximity probesR is measured in (= thousandth of inchR Velocity (seismic probesR is measured in Acceleration (accelerometersR is measured in Displacement is measured in (= millionth of meterR Velocity is measured in

Acceleration is measured in The average thickness of a human hair is about 5 mils which is the same as TRIP set point of all radial probes installed at K-2502 Syn Compressor!!

V Frequency tells us how many times the machine is moving/vibrating per unit of time. Frequency is the indicator of the vibration stimulating problem. Units are M or RPM. !u" Phase angle tells us in which direction a specific section of the machine is vibrating relative to some point. Phase is the angular measurement between a reference point and the vibration peak in a time waveform. Phase angle is used to distinguish between several problems as indicated by the frequency.

Vibration data collection can be collected using: Mounted instruments / probes Proximity Probes Seismic Probes Portable instruments Overall vibration meters (VibrapenR Vibration Analyzers (CSI 2130R

" " The seismic or velocity probe/transducer will measure total vibration at whatever point it is attached to usually the bearing housing of a running machine. . )" The proximity probe measures actual shaft movement usually within its bearing housing onto or through which it is usually mounted. Ov #$ %&v$ ' ( These are used for surveillance purpose only. These cannot be used for analyzing or identifying problems. #$ u )*%&"+,-( These are used for analyzing and identifying problems that cause vibration issues

&'

u !"

u # $$ #

u %$$ %

Vibration Analyzers can be used to collect vibration data in the following manner

Vibration Time Waveform



Vibration Spectrum



Vibration Phase Angle

Each of the above methods of vibration data helps in a unique way to pinpoint the vibration stimulating problem.

/ vV % '0v& ( Time wave form gives the amplitude and direction of vibrating motion of a specific part of the machine in time domain.

This is usually over a small period of time. The same for a longer duration is called trend.

#$ "' % '0v&12 )( Vibration spectrum is a calculated graph plotted through Fast Fourier Transformation of a Time Wave Form. This vibration graph shows the amplitude of vibration in frequency domain.

Vibration problems can be diagnosed in relation to the frequencies being synchronous sub synchronous or non synchronous. ") V2 are those that are intergral multiples of shaft speed (1XR. Imbalance Misalignment Looseness Vane freq etc are examples of Synchronous Frequencies ") V2 are those that are not integral multiples of the shaft speed. Bearing Defects except for Fundamental Train Freq. FTF Cavitation and Resonance are examples of Nonsynchronous frequencies. "$") V2 are those that are below shaft speed. Cage frequency of roller element bearings oil whirl are examples of Subsynchronous frequencies.

This was all about Vibration Basics.

Let¶s now proceed towards How to Analyze a Vibration Signature and How to Diagnose a Problem through Vibration Spectrum«

Before analyzing a vibration problem the following should be known to the analyst. Without this data Vibration Analysis is like groping in the darkness. Machine configuration. Identify all major components and sketch on a paper like motor pump gearbox etc. Determine as many as possible forcing frequencies before taking vibration data Shaft speeds in the machine train Bearing fault frequencies Belt frequencies Gear Mesh Frequencies Blade frequencies

Now collect the vibration data and analyze it for possible problems. -

('! &' ( . / 0- 0(* 1-2'&

o !., 3+ o-44 + 56 43-5 o78





+ & ,''( ! ,

$



%









% # 9(,'(3 1 :

$







+ & ,''( ! ,

%

o o !"#$









%





o &' () *

%



3 1 # 9 $$ !.,%

Since we don¶t live in an ideal world almost always there is one problem or the other which is brewing up in the machine. It is not possible to attend the machine on the onset of slightest of problems diagnosed by vibration analysis as it results in machine down time man hours and spares cost. As a rule of thumb allowable limit of Overall Vibration in the Spectrum of Centrifugal machines is 0.3 in/sec for machines at Plant-I and 7.6 mm/sec for machines at Plant-II. P.T.O

/ $ Imbalance is the result of a shaft¶s center of mass not rotating at the center of rotation. This is because of a heavy spot on the rotor. This heavy spot produces a centrifugal force that forces the rotor to rotate offcenter and causes a high vibration amplitude at 1 x Turning Speed (TSR Frequency in the spectrum and sinusoidal waveform in the time domain data. Imbalance can be Static Imbalance Dynamic Imbalance

" $ Freq Domain : 1 x TS Time Domain : Sinusoidal Phase Difference :

0 Deg H to H or V to V across the rotor 90 Deg H to V on same bearing

' $ Freq Domain : 1 x TS Time Domain : Sinusoidal Phase Difference :

180 Deg H to H or V to V across the rotor 90 Deg H to V on same bearing

6 High level radial vibration: Steady 1X component ± waveform and frequency Amplitude at 1X increases steadily with speed Low level at 2X 3X etc. 6 Low level axial vibration. 6 Notes: 6 Strong 2X 3X « indicate misalignment looseness bent shaft or cocked bearings which must be corrected before checking for imbalance. 6 If 1X µbeats¶ check for broken rotor bars or cocked bearings in motors. 6 If small change in speed causes drastic changes in 1X component Resonance is suspected. 6 On horizontally mounted machine check H V and A vibration. If V is higher suspect base looseness. If A is higher suspect misalignment. Correct these problems before balancing the machine.

/

Misalignment occurs because of poor alignment between mating pieces such as coupling halves bearings shafts and pulleys.

Misalignment results in high axial and radial vibrations. The highest radial vibration usually occurs in the direction of the misalignment. The axial vibrations can be as much as 0.5 to 2 times the amplitude of the radial readings. The radial readings can appear at 1X 2X 3X and even higher multiples of shaft turning speed. However in certain cases the predominant vibration occurs at 1X TS and can be confused with unbalance. In such cases phase data is used to distinguish between unbalance and misalignment.



Freq Domain :

High1x 2x 3x. Low 4x-10x harmonics (if high suspect loosenessR.

Time Domain :

Repeatable periodic time waveform with 1 2 3 or 4 clear peaks per revolution. No high µg¶ impulses.

Phase Difference :

180 Deg in radial or axial direction across the coupling

Other

High axial vibration excessive bearing temperatures

)'1

Misalignment can be

Angular Misalignment



Parallel / Offset Misalignment



Combination of the above two

u

Freq Domain :

1x TS in axial (highR and radial direction

Time Domain :

Sinusoidal with one or two clear cycles per revolution

Phase Difference :

180o in radial vertical or axial direction across the bearings of the same machine

Other :

2x TS in radial direction if Offset Misalignment is also present.



Freq Domain :

Dominant peak in 2x TS in radial direction

Time Domain :

Sinusoidal with one or two clear cycles per revolution

Phase Difference :

180o in horizontal or vertical direction across the coupling.

Other :

1x TS peak in radial & axial direction.

O)'1

Misalignment can also be :

Coupling Misalignment



Bearing Misalignment / Cocked Bearing



Bent Shaft

'

Coupling misalignment occurs when coupling is worn or the coupling is not installed properly. The characteristics of a misaligned coupling are a combination of o Angular and Parallel misalignment.

_

Bearing misalignment occurs when the bearings are not installed in the same plane they are 30 relative to the shaft or if the machine distorts due to thermal growth or soft foot. Freq. Domain :

Normally highest peak at 1x A. Harmonics at 2x 3x or number of balls x TS in the axial direction

Time Domain :

Waveform often shows truncated or flattened pattern indicating a rub. It may also appear periodic or sinusoidal with low amplitude.

Phase Difference:

180o phase difference in axial direction from top to bottom or side to side on the same bearing

_ " 1 Freq. Domain :

Dominant peak at 1x TS in radial and axial direction if the shaft is bent near the center. Dominant peak at 2x TS if the shaft is bent near the coupling.

Time Domain :

The waveform is a mixture of misalignment and imbalance

Phase Difference:

180o in axial direction across the bent shaft. 90o or 270o from horizontal to vertical direction on a bearing. 180o from top to bottom or side to side in axial direction

¢

¢

/ Mechanical looseness occurs when structural or rolling element components do not fit properly. Freq. Domain: A large number of TS harmonics characterizes looseness in the spectrum. In some cases and stages looseness exhibits subharmonics of 1/2 x TS. The highest amplitude typically appears radially in the vertical direction. Time Domain: Waveform shows a great deal of energy and impacting in a random high frequency pattern. Phase:

The phase has no set relationship and is unsteady.

Base looseness of horizontal machines will often appear as high 1X level in V; greater than H component.

V Foot looseness may emerge due to weakness in the base plate or foundation deterioration of the grouting loose hold-down bolts or a cracked or broken foot. Freq. Domain: Dominant frequency is at 1xTS radially. Time Domain: Waveform is periodic. Phase:

180o phase diff. between machine foot vertically and concrete base.

V To help determine a soft foot issue loosen each foot one at a time always keeping the others tight and then retighten while the spectrum is still measuring. Any notable reduction in the one-times (1xR energy during the loosening process can be a strong indication of the relief of machine frame distortion. (

(

Before the loosening sequence: All four feet are tightened; the 1X amplitude is 0.127 inches/second. Radial measurements were taken in the vertical direction.

After the loosening sequence: One of the four feet was loose; the 1X amplitude is now 0.048 inches/second.

Antifriction bearing faults can be identified by bearing defect frequencies in a spectrum. Following are the bearing defect frequencies: FTF

=

Fundamental Train Frequency or Cage Frequency

BSF

=

Ball Spin Frequency

BPFO =

Ball Frequency Outer race

BPFI =

Ball Frequency inner race

These frequencies are specific for each bearing type. Although complex formulae are available for calculation of these frequencies for each bearing yet it is convenient to use bearing software that are designed for the specific purpose of bearing defect frequency determination.

These bearing defect frequencies only appear in a spectrum if there is a fault in the bearing. Following is an example of these bearing frequencies Bearing : 6214 (P-2504 O/B bearingR Pump RPM

=

3570

# of Balls

=

10

BPFO

=

14655

BPFI =

21045

BSF

=

9653

FTF

1465

=

Note that FTF is sub synchronous and the rest are non synchronous.

Gear problems are usually a bit tricky to identify. Before analyzing a signature for gear problems the Gear Mesh Frequency (GMFR of the gear should be known GMF = (# Gear teethR x (RPM of the same gearR v V"0_ 0 It should be noted that the gear mesh frequency will always be present in the spectrum whether or not there is a problem with gears. Gear problems are detected with the help of side bands of the GMF GMF Sidebands = GMF TS The appearance of the sidebands of the GMF in the spectrum indicate that a problem exists with the gear. As the amplitude of these side bands increases it means that the gear problem is worsening.

_3 The problem of a broken tooth cannot be detected by the spectrum alone. The time waveform needs to be looked at for this. If one tooth is broken then a pulse will be generated once per revolution of the gear with the broken tooth.

u 0 1 Usually in order to find out if the vibration problem is because of motor or because of the driven machine the motor is discoupled from the driven equipment and the vibration is checked. However it should be kept in mind that electrical defects disappear from vibration spectrum if the motor is operated in discoupled state. In discoupled state only the mechanical defects of the motor can be picked up in the vibration spectrum like bearing defects rotor unbalance etc.

) *

# V2 ) Usually people think of Vane as the pump¶s primary vibration frequency. Vane is not always a defect frequency but is an operational frequency. It is normal to see this frequency in pump vibration data.

A 5-Vane Impeller of Centrifugal Pump

Vane frequency = (shaft speedR x (no. of blades or vanesR A problem with impeller¶s vane appears as high vane- frequency harmonics with sidebands ± radial or axial. Blade damage may cause imbalance. Impeller looseness may look like mechanical looseness.

v

' + "+, - , .+ % $$ - ,++ ", /, 0 " + " + +/ " /

Vibration due to Cavitation is typically seen as harmonics of turning speed upto the impeller vane frequency. Higher frequency broad band noise can also be found. This broad band vibration is usually present above the impeller blade frequency. µRaised floor¶ is an indication of cavitation. If there is a raised noise floor - look in the time waveform. The waveform will show you the impacts rubs bursts of energy from cavitation and so on. Cavitation is a common problem during startups. It can be rectified by increasing the fluid head (pressureR at suction. ã

Here is an interesting example. The first spectrum is 800 lines. Notice the broad bases of the peaks. Could it be resonance? The next spectrum is 3200 lines - you can see there is more to it. And the third spectrum is zoomed in to the base of the peaks.

Time waveform of cavitation showing bursts

ã ã ã &' 1"

ã ã ã &' 2"

ã ã ã / u+ """3+//

—





ã ã ã ã / (/" + / " $ %

" (15':.& 3 &&+&







% o !., !. % 3+ o-44 + 56 43-5 o % 78





$%



%

$



!,



$ % 5: !. " !.

%



!,

!&&:;< . .% %- 07 0(* 17 '8 (&





$ # 9(,'(3 1 :





-,,'(6 :



o !. % % o !"#%











%

$ "' '( !,:





ã ã ã ã u+ " ""3 /4/ / u 3" ++ 5 /6 4 " ã ""/

u+ / !/ 4"3 u+ " 4/""3 / 4 o 4 /

ã ã ã ã u" + + /

ã ã ã ã ã ! " #$%% &%'#%()%*+,-,+./01 2 $%% &%' (2+,-0+0.1 2 $%% &%' 3(+0-+01 2 #$%%*&%'#%()%*0+0/0.+,1 2 $%%*&%' (2+,-+!0!1 2 $%%*&%' 3(+!!0+.-01 2

ã ¢ ã ã 7 3"4 40"" # " / +"" #ã* / " / ã ã '8oo! o!* &&!'oã #u98!#u98!8 ..... ............................. ....... ..... ....... ....... .... ...

= 3 * "" " & 1 / " /

ã.:7. ' 9ã" 1;;:;; ã<ã" % : 26; ã#<ã" # 2 72; ãu<ã" u5 6; ã<ã" % : ; ã#<ã" # 62 7; ãu<ã" u5 72;

ã ¢ ã ã 2(/ +#ã*

# / + + + / 3+ + 3 ! >22 + ?) .: 2: + ?.77 ã.

u u&o! + + ++" +

o 0 " + 94 X + + " + " +3 +/ "+ / 3+3 " + /+ + "3 "" ã 3 3" / " " " + ," # 3" 3" " " / +/"" +