Vibration Analysis Level- 1 Updated [Compatibility Mode]

Warmest Welcome to Vibration Analysis Level I Course

Vibration Analysis Level‐I y IMRAN AHMAD Director Technical SUMICO Technologies (Pvt) Ltd +92 321 427 6092 [email protected] p www.sumico.com.pk

Timings • • • • • • •

1st Session Session Tea Break 2nd Session S i Lunch/Prayer Break 3rd Session Tea Break Tea Break 4th Session

0900 1045 0900‐1045 1045‐1100 1100 1300 1100‐1300 1300‐1400 1400‐1530 1530‐1545 1530 1545 1545‐1700

• • • • • •

Typical Machinery Problems that Can Be Found Using Vibration Analysis Unbalance Mi li Misalignment t Mechanical looseness Structural problems Structural problems Bent shaft Bearing faults g

Typical Machinery Problems that Can Be Found Using Vibration Analysis • • • • • •

Gear faults Belt problems Lubrication problems Electrical motor faults Cavitations and turbulence others

What is CBM & Why ? What is CBM & Why ? • To try and maximise the plants production and increase the mean time between outages many industries are moved g y towards a ‘Condition Based Maintenance’ approach. • Condition Based Maintenance or CBM is an activity that attempts to predict and trend component failure non‐ intrusively given the end user valuable advanced warning of the problem at hand. • Maximising asset reliability is of the utmost importance in today’s global economy. – As competition and the pressure to produce products cheaper increases the higher consequence of machine/production failure becomes.

CBM Overview • Most machine faults generate some kind of signature that is unique to the particular fault developing. By using the correct technology to detect these signatures we can not correct technology to detect these signatures we can not only tell that a fault is developing, but distinguish what the fault type is. • There are several technologies available to help determine There are several technologies available to help determine the condition of the machine being monitored and the type of fault developing and these are: – – – –

Vibration Analysis y Tribology Sonics Thermography

Diagnosing a machine is just like a person… Di i hi i j lik

Vibration: The ‘pulse’ of the machine

Oil: The ‘life blood’ of the machine

Thermography: ‘Taking its temperature’

Total Picture

Motor Current: The ‘brain waves’ of the machine

Overview of Technologies Overview of Technologies •

Vibration Analysis –

Used to Detect, Analyse and Confirm plant machinery problems. This y p yp can be done in three ways: • • •

On‐line (4500T & CSI6500) for automated and continuous monitoring and protection of critical plant items Portable (2130 Analyser) Route based data collection and analysis Wireless used for remote monitoring of moving or inaccessible equipment

g Overview of Technologies •

Tribology –

Is the analysis of ‘interactive surfaces in relative motion’. • •

Lubricants are analysed on‐site using the 5200 mini‐lab series. The results are plotted in a simple to understand tri‐vector plot showing the ‘Chemistry’ ‘Contamination’ and ‘Wear’ of each lubricant, this allowing the lubricant to be changed on condition rather than on a time based interval lubricant to be changed on condition rather than on a time‐based interval. Wear

Contamination

Chemistry


Sonics –

Through a process known as ‘heterodyning’ Ultrasonic sounds that are non‐audible to human ears are converted back down to a d bl h db kd frequency that is audible to human ears, allowing the operator to hear and recognise faults developing within plant operating systems, such as: such as: • • • • •

Mechanical – Bearings, Rubs, Gear Defects etc Electrical Defects Valve Operation Steam Trap Operation Leak Detection – Pressurised Systems and Vacuum Systems


Thermography –

Thermal Imaging is used to locate potential problems by detecting g g p p y g abnormal temperature fluctuations at a glance. •

This can be used in a wide array of circumstances but is most commonly used in electrical control panels


Corrective technologies allow the engineer to set‐up the machine to try and prevent premature machine failure from such causes as Imbalance and Misalignment d Mi li t – When these forces are induced upon a machine components such as bearings, seals and even supports fail due to stress – Technologies such as Laser Alignment and Balancing prevent these from being Technologies such as Laser Alignment and Balancing prevent these from being so much of a problem

y g Machinery Health Manager A1 - Recirculation Pump #5 -M2H MOTOR O O INBOARD O BRG. G - HORIZONTA O O

PK In/Sec

RCP#5 C # 0.025 0.020 0.015 0.010 0.005 0

•

PK In/Sec



ALERT

0

0.05 0.04 0.03 0.02 0.01 0 0

Acc in G-s s

Trend Display 36-65xTS

FAULT

1.0 0.5 0 -0.5 -1.0 -1.5 0

The machines due to be monitored are defined within the Each technology is stored and analysed from a single software platform, software. software allowing the analyst to: allowing the analyst to: 100

–– – – 1

200 300 Days: 11-Aug-95 To 11-Dec-96

400

500

Route Spectrum 11-Dec-96 17:33:57 OVERALL= .0604 V-DG PK = .0605 LOAD = 100.0 RPM = 3593. (59.89 Hz)

40

As much information as possible about the machines being monitored Store all data and information in one database isEasily cross reference data for conformation of analysis preferred when building the database. 80

120 160 Frequency in kCPM

200

240

Route Waveform 11 D 96 17 11-Dec-96 17:33:57 33 57 RMS = .4233 PK(+/-) = 1.13/1.22 CRESTF= 2.89

Collaborate all data into one single report Collaborate all data into one single report. 2

3 4 Revolution Number

5

6

7

Overview O i of Condition Monitoring Maintenance Philosophies Maintenance Philosophies

Definition of Maintenance Definition of Maintenance • The The act of causing to continue act of causing to continue (Webster) • Keeping equipment in repair (Oxford)

Maintenance Reactive Maintenance – Often called ‘Breakdown Maintenance’ and has the concept ‘fix it when it breaks’. breaks • This is probably the most common type of maintenance in industry today but can be the most costly, especially on critical machines. • Maintenance costs are usually higher due to the catastrophic failure that occurs.

Predictive Maintenance – Also known as ‘Condition Based Maintenance’. • This approach uses non-intru technologies to determine the actua condition of a machine and its rate of failure. • This can be very effective in extending machine life with big financial savings if implemented properly.

Planned Maintenance Also known as ‘Shutdown Maintenance’. This is based upon p ‘Timed Intervals’ between maintenance. Can be very effective if maintenance and resources are aimed at the machines that need it the most. However it can be very difficult to distinguish which machines actually need maintenance.

Proactive Maintenance Often referred to as ‘Root Cause Analysis’. This philosophy works hand in hand with Predictive Maintenance, eliminating the source of the fault to try to prevent it from re-occurring.

y Today’s Industrial Demand • It should be unacceptable to deliver – less performance for more money l f f – same performance for more money

• It could be acceptable to deliver – same performance for less money same performance for less money – more performance for the same money – more performance for more money

• The desire is More Performance for Less Money!!!!

j Predictive Maintenance Objectives

• To To confirm good confirm good‐condition condition machines machines • To detect developing problems • To determine the nature and severity of the d i h d i f h problem • To schedule repairs that can best fit with production and maintenance needs

q Predictive Maintenance Techniques

• • • • • • •

Vibration measurement Vibration measurement Electrical testing Motor current analysis l i Reciprocating machine testing Thickness testing Visual inspection Visual inspection And many more…

Predictive Maintenance Basic Facts

• Every Every mechanical or electrical faults on a mechanical or electrical faults on a machine has a distinct vibration behavior. • Any change in the vibration signature Any change in the vibration signature indicates changes in the dynamic operating condition of the machine condition of the machine

( ) Predictive Maintenance Mechanism (VA)

• Establish Establish a database of all the machines that a database of all the machines that need to be monitored • Establish a data collection route that best Establish a data collection route that best optimize the data collection time • Download route into the data collector D l d i h d ll • Collect data • Upload collected data into the database

Predictive Maintenance Mechanism

• Run Run exception reports to detect the exception reports to detect the problematic machines • Analyze only the machines in the exception Analyze only the machines in the exception reports • Generate repair work to be performed G i k b f d • Again collect data on the machine on which work is being done.

Predictive Maintenance Predictive Maintenance Rules + Experi

Start NO YES Create Ref.

Regular Meas.

Compare limits

Fault Diagnostics

Input m/c specs Create New Ref. & Limits

Fault correction

Vibration Fundamentals Vibration Fundamentals How Much Vibration is Too Much ? 1. Use Absolute Vibration Levels - Given Gi by b machine hi makers k - Published Vibration Severity Standards eg. ISO 2372, VDI 2056, BS 4675

2. Use Relative Vibration Levels

ISO 10816 3 ISO 10816‐3 11

0 44 0.44

7.1

0.28

4,5

0.18

3,5 ,

0.11

2,8

0.07

2,3

0.04

1.4

0.03

0,71

0.02

mm/s rms

rigid

flexible

rigid

flexible

pumps > 15 kW

rigid

flexible

medium sized machines

radial, axial, mixed flow

integrated driver

external driver

Group 4

Group 3

15 kW < P  300 kW

motors 160 mm  H < 315 mm Group 2

rigid

flexible

inch/s rms

Foundation

large machines 300 kW < P < 50 MW

Machine Type

motors 315 mm H Group 1

Group A B

C D

newly commissioned unrestricted long-term operation restricted long-term operation vibration causes damage

ISO 10816 3 ISO 10816‐3 140

5.51

113

4.45

90

3 54 3.54

71

2.80

56

2.20

45

1.77

36

1.42

28

1.10

22

0.87

18

0.71

11

0.43

µm rms

rigid

flexible

rigid

flexible

pumps > 15 kW

rigid

flexible

medium sized machines

radial, axial, mixed flow

integrated driver

external driver

Group 4

Group 3

15 kW < P  300 kW

motors 60 mm  H < 3 315 5 mm 160 Group 2

rigid

mil rms

flexible

Foundation

large machines 300 kW < P < 50 MW

Machine Type

motors 315 3 5 mm  H Group 1

Group A

newly commissioned

B

unrestricted long-term operation

C

restricted long-term operation

Vibration standards are guidelines Vibration standards are guidelines

Just Tolerable Just Tolerable

Allowable

Just Tolerable All Allowable bl Allowable

Good

Good

Large Machines with rigid and heavy foundations whose G d Good natural Frequency Small <300 kW on special exceeds Machines< 15 kW foundations machine speed 15 kW< Medium Machines <75kW

Group K

Group M

Group G

45 28 18 11.2 71 7.1 4.5 2.8 18 1.8 1.12 1.71 0 45 0.45 0.28 0.18

Velociity mm//s RMS

Nott N Permissible

10 times = 2 20dB

Not Permissible

Not Permissible

2.5 time es = 8dB

ISO2372 ( BS 4675 , VDI 2056 )

Predictive Maintenance Database Setup Predictive Maintenance Database Setup • Identify Identify which machines to monitor which machines to monitor • Identify each machine characteristics • Define analysis requirements for each fi l i i f h machine • Define acceptable levels and alarm limits • Define data collection point locations and p monitoring methods

Which Machine to Monitor? Which Machine to Monitor? • • • •

Machine that are vital to the operation Machine that are vital to the operation Machines that are expensive to repair Machines that are trouble makers hi h bl k Machines that are in remote or inaccessible locations

Why Machine Characteristics? Why Machine Characteristics? • Initially Initially, the knowledge of the machine design the knowledge of the machine design and its operating characteristics is mandatory to successfully establish a good database to successfully establish a good database • Later, this knowledge will provide the basis for analyzing the data accurately for analyzing the data accurately

What Machine Information Is needed? What Machine Information Is needed? • • • • • • •

Machine speed p Machine load Bearing type g yp Coupling type yp Gear type and teeth count Blades and vanes g yp p g Machine drawings and typical operating conditions

Machine Analysis Requirements Machine Analysis Requirements • List List all possible problems of the machine all possible problems of the machine • Determine the particular effects that each problem impose on the machine problem impose on the machine • Determine the best method to monitor the severity of the problem i f h bl

Manpower Required Depends on Manpower Required Depends on • Number of data collection points: Number of data collection points: – Complexity of the machine – Number of machines to be monitored Number of machines to be monitored

Manpower Required Depends on Manpower Required Depends on • Analysis time required Analysis time required – Complexity of the machine – Complexity of the problem Complexity of the problem

• Frequency of analysis – Machine classification – Machine history

Machine Complexity Machine Complexity • Simplex machines Simplex machines – Constant speed and load – Direct drive (coupling) Direct drive (coupling) – 5‐10 measurement points

Machine Complexity Machine Complexity • Compound Machines: Compound Machines: – Constant speed and load – Intermediate drive (gearbox and belts) Intermediate drive (gearbox and belts) – 10‐20 measurement points

Machine Complexity Machine Complexity • Complex Machines Complex Machines – Variable speed and load – Multiple components Multiple components – More than 20 measurement point

Machine Classifications: Machine Classifications: • Vital Machines: Vital Machines: – – – –

Irreplaceable Halt production Halt production Hard to find parts E Expensive to repair i t i

Machine Classification Machine Classification • Critical Machines: Critical Machines: – – – – –

Halt part of production Expensive to repair Expensive to repair Costly replacement H d t fi d Hard to find parts t Frequent repairs

Machine Classifications: Machine Classifications: • Support Machines: Support Machines: – – – –

Not too expensive to repair Parts are readily available Parts are readily available Affect but don’t halt operation M d t l Moderately costly repair or replacement tl i l t

Machine Classifications: Machine Classifications: • Other Machines: Other Machines: – – – –

Parts are readily available Replacement is easy and inexpensive Replacement is easy and inexpensive Do not affect operation directly N hi t No history of repair f i

Monitoring Frequency Monitoring Frequency • Vital Machines – On‐line Monitoring or every 1‐2 weeks

• Critical Machines – Every 2‐4 weeks

• Support Machines – Every 4‐8 weeks

• Other Machines – Every 8‐12 weeks E 8 12 k

Methods of Data Collection Methods of Data Collection • On‐line On line Continuous Monitoring Continuous Monitoring • Manual Data collection through portable data collection

Continuous Monitoring Continuous Monitoring • Real‐time Real time data acquisition through dedicated data acquisition through dedicated sensors and instrumentation that monitor the machine during every second of its operation machine during every second of its operation. • Sometimes the instrumentation supplied with relays for automatic shutdown when alarm relays for automatic shutdown when alarm levels are exceeded.

Manual Data Acquisition Manual Data Acquisition • Using Using a portable instrumentation with a portable instrumentation with sensors, data can be captured on a scheduled intervals • Data then is dumped back to a PC for trending analysis and reporting trending, analysis, and reporting.

Manpower Required for Data Collection Manpower Required for Data Collection • • • •

Level of expertise: Tech Level of expertise: Tech Amount of training: Minimum Frequency of training: once a year f i i High level of commitment

Manpower Required for Data Analysis Manpower Required for Data Analysis • Level Level of expertise: Engineer or highly of expertise: Engineer or highly knowledge mechanic • Duties: analyze data and run and manage the Duties: analyze data and run and manage the program • Amount of training: Varies A f i i V i • High level of commitment

Predictive Maintenance •

Results: – – – – – – –

Increase machine availability Save on maintenance cost Reduce spare parts Reduce spare‐parts inventory Increase machine life Avoid unnecessary repairs Avoid unnecessary repairs Organize maintenance activities I Improve plant safety l t f t

Introduction to Vibration Analysis l Introduction to Vibration Analysis Introduction to Vibration Analysis

General Description Vibration General Description‐Vibration •

There are many different parameters we can measure to help us determine machinery health:

Voltage Power Pressure Viscosity

Current Flow Flow Temp Torque Speed Density Emission

Particles Load • None contains as much information as the vibration signature!!! None contains as much information as the vibration signature!!! • Not only does it provide the severity of the problem but can also point to the source of the problem •

‘Vibration’ can be simply stated as ‘A response to some form of excitation’ – The ‘excitation’ is generally referred to as the ‘Forcing Function’

•

Vibration is the motion of a body about a reference position caused by a force

General Description – Forcing Function •

When a forcing function is applied to a shaft within a plain bearing the free shaft within a plain bearing the free movement will cause the shaft to vibrate within the bearing – Here we are measuring actual shaft movement movement



When a forcing function is applied to a shaft within a bearing housing where there is very little free movement, then the vibration will transmit i to the h casing i – Measuring the casing movement of a specific component as result of the forcing function

Vibration from Mechanical Faults b at o o ec a ca au ts

Vibration from Mechanical Faults





Vibration Characteristics

• Amplitude

How Much

• Frequency

How Often

• Phase.

When

General Description – Measuring Response •

You can also look at vibration as the amount of ‘Time’ it takes to complete a particular cycle – If If we examine the motion of a forcing function on a fan blade we examine the motion of a forcing function on a fan blade ‘Heavy Heavy Spot Spot’ over over a period of time a distinct signature will occur.



This motion is called a sine wave. – The horizontal axis is measuring Time – The vertical axis is measuring Amplitude



This is known as a ‘Ti ‘Time Waveform’ W f ’ – Amplitude versus Time

Time Waveforms Time Waveforms •

Unfortunately there are multiple sources of forcing functions that can emit from a machine or component. – Thus resulting in the time waveform becoming complex in nature

Ex15

0.4

•

The plot shown on the right is a complex time waveform. l ti f

0.2

EX 8

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

•

This is just one format (domain) for analysing vibration data vibration data. Data can also be analysed in a ‘Spectrum’ – (Amplitude Vs Frequency) through a process known as the FFT known as the FFT

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

0.3

– Amplitude versus Time

•

A8 - Example 15 -F2V Fan Outboard Vertical

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

Los - Example 8 -P2V Pump Outboard Vertical

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

0.1

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

0.0 0.8

-0.1

0.6 -0.2

-0.3 0.4 -0.4 0

50

100

0.2

150 200 Time in mSecs

250

300

350

Time: 120.44 Ampl: -.07595

0 0

6000

Label: Looseness

12000 18000 Frequency in CPM

24000

30000

Freq: q 736.86 Ordr: 1.000 Spec: .245

Fast Fourier Transform Fast Fourier Transform – FFT Process FFT Process •

When a problem starts to develop within a rotating component it will generate a vibration signature. This signature should be captured in the time waveform – Distinguishing that signature can be very difficult when looking at a time plot Di ti i hi th t i t b diffi lt h l ki t ti l t

•

To understand the problem we need to understand the frequency – ‘How often is it occurring?’

• •

The ‘FFT’ is a process that determines the frequency of a signal from a time waveform. f The FFT is named after an 18th century mathematician named ‘Jean Baptise Joseph Fourier’. He established: – ‘Any periodic signal can be represented as a series of sines and cosines’. – Meaning if you take a time waveform and mathematically calculate the f k f d h ll l l h vibration frequency, it can be converted to a more familiar format

Tim e

Amplitude e

Ampliitude

Amplitude e

How the Vibration Spectrum is Created How the Vibration Spectrum is Created

q y Frequency Domain The frequency domain (Spectrum) plots the data as ‘Amplitude’ in the (Y) axis and ‘Frequency’ in the (X) axis. This data is derived from the time domain – mathematical manipulation of the time waveform. p • Recall the waveform and spectrum from the previous slide. If you tried to determine all the frequencies from the waveform plot, you would need all day just to analyse one point of data. • As the FFT plots the frequencies from the waveform for you the analysis of this data becomes easier and reduces the amount of time needed for analysis of each point.

•

Ex15

0.4

A8 - Example 15 -F2V Fan Outboard Vertical

0.3

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

0.1

0.0

-0.1

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

0.8 RMS Velocity in mm m/Sec

Acceleration in G G-s

0.2

EX 8

1.0


0.6

0.4

-0.2

0.2 -0.3

-0.4 0

50

100


250

300

350

Time: 120.44 Ampl: -.07595

0 0

6000

Label: Looseness


24000

30000

Freq: 736.86 Ordr: 1.000 Spec: .245

Introduction to Vibration Analysis l Units of Measurements Units of Measurements

Measuring Amplitude and Frequency Measuring Amplitude and Frequency •

You can measure amplitude from a time waveform as shown:

+

0 to Peak RMS

Amplitud de

Average

-

Time ‘t’

Peak to Peak

•

The period ‘t’ is the time required for one revolution of the shaft in this illustration which equals one cycle of the waveform illustration, which equals one cycle of the waveform – During this period, the amplitude of the waveform reaches a positive (+) peak, returns to rest, and reaches a negative (‐) peak before returning to rest

Measuring Amplitude and Frequency Measuring Amplitude and Frequency

•

•

You can calculate the different amplitudes when one of the values are known: – RMS = 0.707 times the peak value RMS 0 707 times the peak value – Avg = 0.637 times the peak value – Pk‐Pk = 2 times the peak value

0 to Peak

+

•

RMS Average

Amplitude

•

Peak (Pk) – Amplitude measured from the ‘at rest’ position (0) to the highest value (0 to Peak) Peak to Peak (Pk Peak to Peak (Pk‐Pk) Pk) – Amplitude measured from the peak positive (+) Amplitude measured from the peak positive (+) value to the peak negative (‐) value RMS (Root Mean Square) – obtained by averaging the square of the signal level over a period of time and then taking the square root result A Average (Avg) – (A ) Amplitude value that averages the peak values of the A lit d l th t th k l f th waveform

Time ‘t’t

-

•

Peak to P k Peak

Measuring Amplitude and Frequency Measuring Amplitude and Frequency • •

Severity of a vibration problem can be determined by the amplitude of the vibration. We can measure amplitude in one of three ways 1. Displacement – measures the distance the shaft moves in relation to a reference point. 2. Velocity – measures the displacement of the shaft in relation to time 3. Acceleration – measures the change in velocity in relation to time

•

The most common industrial applications are: The most common industrial applications are: 1. Displacement 2. Velocity 3. Acceleration –

‐ Microns ‐ Peak to Peak value ‐ mm/sec ‐ RMS ‐ G‐s ‐ Peak value

G‐s = 1 x force of gravity (G‐force) g y( )

Amplitude Relationships Amplitude Relationships The three types of amplitude measurements used to display data are directly related to each other – Changing from one amplitude unit to the next alters the way in which the data is displayed

• Ex15

0.35 5 140

A8 - Example 15 -F1H Fan Inboard Horizontal

Low frequencies require very Forlittle normal speed forceoperating to move an object ranges, velocity data provides the best indication of Increasing frequency that machine hi the condition diti

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

•

0.25 100

3 0.20 0 20 80

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

0.15 60 2

0.10 40 1 0.05 20

00 0

20000

Label: Large Fan Unit - Easy

40000 Frequency in CPM

– High and low frequency h dl f events can be seen

R t S Route Spectrum t 22-Aug-02 11:30:50 OVERALL= 3.45 V-DG RMS==.3909 PK 3.44 P-P 104.98 LOAD = 100.0 RPM = 831. (13.85 Hz)

60000

Velocity is the default unit for standard data collection techniques



Displacement measures low frequency events ignoring high frequencies – Relative shaft motion



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

Frequency Units Frequency Units •

Frequency refers to how often something occurs: – –

•

How often a shaft rotates? How often a rolling element hits a defected race?

There are three ways to express frequency: 1. CPM – Cycles Per Minute –

1CPM = 1RPM

y 2. Hz – Cycles Per Second –

CPM / 60

3. Orders – Multiples of Turning Speed –

•

Frequency/Turning Speed

Consider a motor has a rotational speed of 1485RPM, in terms of frequency this equates to: 1485 CPM (1rpm = 1cpm) (1rpm = 1cpm) – 1485 CPM – 24.75 Hz (1485/60) (minutes to seconds) – 1 Orders (1 x revolution of the shaft)

q y Frequency Units •

Shown below is a table showing the relationship between all three frequency relationship between all three frequency units with reference to the turning speed

Motor Turning Speed = 1500RPM CPM

1500

2250

3000

6000

12000

Hz

25

37.5

50

100

200

Orders

1

1.5

2

4

8

Frequency Domain Frequency Domain •

The vibration analyst can divide the frequency domain data into three major areas of interest h f 1. 2. 3 3.

• •

Synchronous Equal to Ts or Harmonics of Ts Sub synchronous < 1 x Ts N Non synchronous h > 1 x Ts but not an integer 1 T b t t i t

Note ‘Ts’ is the turning speed or rotational frequency (RPM) of the shaft at the position where you make the measurement measurement Each defect that can materialise in the frequency domain can be categorised into one of three types of energy listed above –

Knowing the type of energy within the data can help the analyst quickly eliminate 2/3rd of the fault types

Harmonic Orders Harmonic Orders •

Harmonics are cursors that are exact multiples of the primary frequency – They are used to locate other frequencies related to the primary cursor 1.0

RMS V Velocity in mm/Sec

0.8

•

0.6

EX3

Los - Example 3 -P2V Pump Outboard Vertical Analyze Spectrum 15-Nov-95 10:00:16

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

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

04 0.4

Therefore:

0.2

– When the primary cursors is located on 1Order all the harmonics will be synchronous h 0 Freq: 736.86 0 6000 12000 18000 24000 30000 Ordr: 1.000 – Harmonic cursors can be used to show non‐synchronous and sub‐ Frequency in CPM Spec: .245 synchronous harmonics depending upon the energy of the primary frequency

Energy in the Spectrum Energy in the Spectrum E4

05 0.5

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

RMS Velocity in mm/Sec

0.4

0.3

0.2

01 0.1

0 0

20

40

60


120

140

160

Freq: 2.937 Ordr: 1.000 Spec: .01038

y gy Synchronous Energy EX 8

1.0

Los - Example 8 -P2V Pump Outboard Vertical Analyze Spectrum 15-Nov-95 10:00:16

Synchronous energy ‐ related to 0.8 turning speed. i d

•

0.6 We can see from the spectrum that the first peak spectrum that the first peak is at 1 Orders (which means it 0.4 is 1 x turning speed)

•

All the other peaks are All th th k harmonics off, which means they are related to the first peak

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

RMS Veloc city in mm/Sec

•

0.2

0 0

6000


24000

Label: Looseness

Examples of synchronous energy: 1) Imbalance 2) Misalignment

3) Gearmesh

30000

Freq: 736.86 Ordr: 1.000 Spec: .245 245

y gy Non‐Synchronous Energy E5

2.0

•

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

Non‐synchronous energy ‐ not related to turning speed not related to turning speed

1.8

OVRALL= 2.63 V-DG RMS = 2.69 LOAD = 100 100.0 0 MPM = 3225. RPM = 380.

•

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


1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0 0

6000


24000

30000

Freq: 3888.9 Ordr: 10.24 Spec: .748

Label: Outer Race Defect Priority: 1

• Examples of non-synchronous energy: •

Bearings

Multiples of belt frequency

Other Machine Speeds

y gy Sub‐Synchronous Energy •

Sub‐synchronous energy ‐ Less than turning speed than turning speed

•

The spectrum shows the first impacting peak below 1 Order. This is sub‐synchronous energy

•

Examples of sub‐synchronous energy are: Belt Frequencies Other Machine Speeds Other Machine Speeds Cage Frequencies

• • •

Energy in a Spectrum Energy in a Spectrum 

Synchronous – N x RPM where N is an integer g



Sub-synchronous – <1 x RPM



Non-synchronous – F x RPM where F is >1x RPM but not integer g

y gy Causes of Sub Synchronous Energy • Frequencies that show below the rotational frequency (Less than 1 Order) are sub synchronous. – – – –

Another component Another component Cage frequencies Primary belt frequency Oil whirl (plain bearings)

y gy Causes of Synchronous Energy • Frequencies that are equal too or a direct multiple of running speed are Synchronous • Possible causes of Possible causes of Synchronous energy are: Imbalance Misalignment Looseness Vane pass frequency Vane pass frequency Gears etc

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

0.8 RMS Velocity in mm/Sec

– – – – –

EX 8

1.0

0.6

0.4

0.2

0 0

6000

Label: Looseness


24000

30000

Freq: 736.86 Ordr: 1.000 Spec: .245

y gy Causes of Non Synchronous Energy • Frequencies above (but not integer multiples of) turning speed are non synchronous. • Possible causes of non synchronous energy are: synchronous energy are: Another component Antifriction bearings Electrical System resonances Multiples of belt frequency Multiples of belt frequency

E5

2.0

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

1.8

OVRALL= 2.63 V-DG RMS = 2.69 LOAD = 100.0 MPM = 3225. RPM = 380.

1.6 RMS Velocitty in mm/Sec

– – – – –

1.4 1.2 10 1.0 0.8 0.6 0.4 0.2 0 0

6000



24000

30000

Freq: 3888.9 Ordr: 10.24 Spec: .748

Lines of Resolution Lines of Resolution LOR

Lines of Resolution Lines of Resolution •

Lines of Resolution (LOR) determine how clear the peaks(data) are defined within our spectrum.

•

The more lines we have over the same F‐max (Maximum frequency scale). The more accurate our data will be

•

Example. –

The diagram below shows data that has been collected using 400 LOR. Notice how the top of the peaks are capped. When the LOR are increased the data becomes more accurate.

Lines of Resolution TA16

0.20 0.5

L2 - TA 16 -M1H Motor Outboard Horizontal Analyze Spectrum 13-Mar-01 09:14:16 09:13:53

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

•

The spectrum shown displays The spectrum shown displays data at 800 L.O.R with an Fmax of 1600 Hz



The second spectrum displays the same data but with 3200 L.O.R over the same Fmax

PK = .3852 3852 .7078 7078 LOAD = 100.0 RPM = 1497. 1496. RPS = 24.95 24.94

0.16 0.4

0.12 0.3

0.08 0.2

0.04 0.1

0 0

400

800 Frequency in Hz

1200

1600

Lines of Resolution Lines of Resolution •

The range of LOR settings that we can choose from on the analyzer starts at 100 Lines and go up to 12800 Lines. 100 Li d 12800 Li

•

The average number of LOR is around 1600 Lines for a typical The average number of LOR is around 1600 Lines for a typical motor/pump set up

To change the LOR settings we need to alter our parameter set set. This is done in the Database Setup program Remember. If you double your lines of resolution you double your data collection time.

p y Spectral Summary E5

RMS S Velocity in mm/S Sec Accceleration, Velocity Dissplacement V

Am mplitudes s

2.0

BF - Example 5 -R4A ROLL BRG. #4 - AXIAL

Energy Types

18 1.8 1.6

Synchronous

1.4

Non Synchronous

1.2

Sub Synchronous

H Harmonics i Multiples of Primary Frequency

Route Spectrum 12-Jul-96 17:16:42 OVRALL= 2.63 V-DG RMS = 2.69 LOAD = 100.0 MPM = 3225. RPM = 380.

Resolution

1.0

Clarity of the spectral d t data

0.8 0.6 0.4 0.2 0 0

6000



Frequency

Hz (CPS), CPM, Orders

24000

30000

Freq: 3888.9 Ordr: 10.24 Spec: p .748

Introduction to Vibration Analysis Data Collection Data Collection

g q Transducers and Mounting Techniques • Although there are many different types of transducers available, the most common type used for day to day data collection are Accelerometers. • These transducers provide an electrical charge proportional to acceleration by stressing piezoelectric crystals typically acceleration by stressing piezoelectric crystals typically 100mV/g sensors are used.

Data Qualityy •

Whether it is your job to collect the data and/or analyse the data it is important to understand that the technologies will not give you the answer to a machines problem unless you have collected meaningful, quality data

•

There are certain considerations that must be taken prior to any data being collected, these are: –

– –

A good understanding of the internal make up of the machine, in order A good understanding of the internal make up of the machine in order to understand the best transmission path for data collection ‐ bearing locations, load zones etc. Ensure data is collected in a repeatable manner so we can compare p p two or more readings to each other ‐ trending purposes Variable speed machines ‐ it is very important to collect data with the correct running speed enter into the analyser

Transmission Path •

Damaged caused to a machine component will cause a certain amount of vibration/sound or heat to propagate away from the initial impact initial impact. –

•

In many cases the further you are away from the initial event the weaker the signal will become, resulting in the data appearing to be lower in value. –

•

It is the effect of the impact/force that we are trying to detect

In more extreme cases the impact can be lost amongst other machine noise by the time it has reached your transducer, resulting in no detection of a machine problem.

Usually the best place to acquire data from a machine, is at the bearings. –

This is because the bearings are the only part of the machine that connect the internal rotating components to the stationary components (Casing)

p Repeatable Data •

Collect data in the same manner each time. –

•

In order to aid with repeatable data the analyser requests for d data to be collected in certain locations on the machine. b ll d i i l i h hi –





This consistency will allow you to trend the machinery condition and y y y properly judge the progression of faults

These are called ‘Measurement Points’

A measurement point is determined by three characters and a description. Each character refers to a particular place on the machine being monitored – E.g. E g M1H is a typical measurement point

Measurement Points Measurement Points • A measurement point is defined as three alpha numeric digits along with their respective definition – Orientation and location on each component Orientation and location on each component

• The image on the right is taken from the screen of the 2130 analyser during the 2130 analyser during a collection ‘route’ • The measurement ‘point identifier’ can be seen in identifier can be seen in the top right while the ‘point description’ is shown just below j

Measurement Points Measurement Points • The first letter of the ‘Point Identifier’ refers to the type of machine being monitored – M M = Motor Motor

P = Pump P Pump

F = Fan F Fan

• The second character represented by a number indicates the location on the machine – Inboard (Drive End) or Outboard (Non Drive End)

• The third letter refers to the orientation of the sensor or the type of processing being done by the analyser – H = Horizontal

V = Vertical

P = Peakvue Change in DSP of Analyser

Measurement Points Measurement Points • The following example shows how the numbering system changes as you cross from one component to the next  

M1H – Motor Outboard Horizontal M1P – Motor Outboard Horizontal Peakvue

2

1

1  

P1H – Pump Inboard Horizontal P1P – Pump Inboard Horizontal Peakvue

2

• Notice how the ‘1’ is not always the ‘Outboard’ – This changes when the next component is required for data collection • The numbering system starts from 1 again The numbering system starts from 1 again

Introduction to Vibration Analysis Fault Diagnostics Fault Diagnostics Imbalance, Misalignment, Looseness

Fault Diagnostics Fault Diagnostics •

•

•

Each type of machine fault yp f or defect f reveals a specific p vibration characteristic in the spectrum and time waveform domain that distinguish that fault from another. Si l b Simply by gaining a basic knowledge of these patterns and i i b i k l d f th tt d applying a few rules of thumb we can start to analyse machine vibration and prevent machine failure. This section concentrates the characteristics / patterns and rules that apply to diagnose machine faults such as: – – –

IImbalance b l Bearings Resonance

Misalignment Mi li t (Peakvue)

LLooseness Belts

Gears G Electrical

Imbalance

Imbalance • Imbalance (Unbalance) occurs when the centre of mass differs from the centre of rotation. • If the centre of mass changes on the rotor due to a heavy spot or some other influence then a centrifugal force is produced. This results in the centre of rotation being offset from the This results in the centre of rotation being offset from the centre of mass causing the vibration to increase at the rotational frequency.

Imbalance (Types) ( yp )

Imbalance • Causes of Imbalance – – – –

Improper Assembly Material build up / dirt Material build up / dirt Wear to components Broken or missing parts

All of the above conditions will result in an unbalanced state • Diagnostic Rules for Imbalance – – – – – – –

Periodic non Periodic non‐impacting impacting sinusoidal waveform sinusoidal waveform Spectral peak at 1xTs (1 Order) Very little axial vibration Similar amplitudes between horizontal and vertical plains p p Synchronous fault type Amplitudes will increase with speed Very low harmonics of 1xTs

Imbalance Spectral Data Imbalance Spectral Data • The spectrum shown represents a simple unbalance state – Single peak at 1xTs (1 Order) – Little indication of harmonics Little indication of harmonics

6

RMS Velocity in mm//Sec

5

4

Ex2

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

• What should the waveform show?

OVRALL= 4.58 V-DG RMS = 4.56 LOAD = 100.0 RPM = 3000. RPS = 50.00

3

2

1

0 0

20000


60000

80000

Freq: 3000.0 Ordr: 1.000 Spec: 4.539

Imbalance Waveform Data Imbalance Waveform Data • Despite the waveform being displayed in Acceleration – Default unit for route based waveform data

• There is still a predominant sinusoidal waveform pattern There is still a predominant sinusoidal waveform pattern – 1 x Revolution sine wave 10 1.0 0.8 0.6

Ex2

IF - Example 2 -F1H Fan Inboard Horizontal Waveform Display 02-Feb-00 15:13:51 PK = .5289 LOAD = 100.0 RPM = 2985. RPS = 49.76

Acceleration in G-s A

0.4 0.2

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

-0.0 -0.2 -0.4 -0.6 -0.8

• Ch Changing the units to velocity would reduce the amount of high i th-1.0 it t l it ld d th t f hi h 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 frequency noise residing on the waveform Revolution Number

Imbalance Trend Data Imbalance Trend Data • The trend data is a good way of determining if there has been a change in condition, as this plots amplitude against time ( e e t e s days) (where time is in days) • Here the 1xTs parameter is being trended – Vibration has been steady at 3mm/sec for a period of time – A sudden change instate should alert the analyst to a fault developing E02N - JB1420C CONDY RECOVERY PUMP JB1420C -M1H Motor Outboard Horizontal

14

Trend Display of 1xTS

RMS Velo ocity in mm/Sec

12

-- Baseline -Value: 3.063 Date: 07-Apr-00

10

8 FAULT 6

4

2

0 0

100

200 300 400 Days: 07-Apr-00 To 21-May-01

500

Date: 21-May-01 Time:14:24:29 Ampl: 11.21

Imbalance Problem ‐ Practical Imbalance Problem • The following fan unit has an imbalance present on the rotor. – 1xTs Peak in the Spectrum – 1xTs Peak in the Waveform 1xTs Peak in the Waveform Imbalance Ex2

6

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


5

4

3

2

1

0 0

20000 Ex2

1.0

40000 Frequency eque cy in C CPM IF - Example 2 -F1H Fan Inboard Horizontal

60000

80000

Freq: Ordr: Spec:

3000.0 1.000 4 539 4.539

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

0.8

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

0.6

Acceleration in G-s

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

0.2 -0.0 -0.2 -0.4 -0.6 -0.8 -1.0 0

0.5

1.0

1.5

2.0

2.5 Revolution Number

3.0

3.5

4.0

4.5

5.0

• What would happen to the data if the following occurred to th f ? the fan?

Imbalance Case Study 1 Imbalance Case Study 1 • •

Background The following data is taken from a Recirculation Fan designed to circulate the hot air through an Oven to aid with drying the process. The oven is vertically mounted and the product comes into the oven from the top and exits at the bottom. There is one Recirculation Fan and one Extract Fan. L Loss of function from either fan results in the oven being taken offline. f f ti f ith f lt i th b i t k ffli Bm/c - TOP RECIRC FAN TRF B m/c -F2H Fan Outboard Horizontal

6

The spectral plots shows The spectral plots shows a dominant 1xTs peak (1 Order) with very little other vibration present

OVERALL= 4.70 4 70 V V-DG DG RMS = 4.70 LOAD = 100.0 RPM = 1246. RPS = 20.77

5

RMS Velo ocity in mm/Sec

•

ROUTE SPECTRUM 08-Nov-04 14:16:45

4

3

2

1

0 0

20000


80000

100000

Freq: 1246.3 Ordr: 1.000 Spec: 4.673

Imbalance Case Study 1 Imbalance Case Study 1 • The waveform from this data shown on the following page represents a sinusoidal waveform clearly shown once per revolution of the shaft – e o ut o o t e s a t here the waveform is displayed in e et e a eo s d sp ayed velocity. Bm/c - TOP RECIRC FAN TRF B m/c -F2H Fan Outboard Horizontal

8

Route Waveform 08-Nov-04 14:16:45

• All All indications point indications point towards an imbalance problem. The amplitudes should be checked in both radial directions to confirm this problem

RMS = 4.66 LOAD = 100.0 RPM = 1246. RPS = 20.77

Velocity in mm m/Sec

4

PK(+) = 7.03 PK(-) = 7.40 CRESTF= 1.59

0

-4

-8 8

-12 0

1

2

3 Revolution Number

4

5

Imbalance Case Study 1 Imbalance Case Study 1 • The plot shown indicates a multi spectral plot showing all the radial directions. Bm/c - TOP RECIRC FAN TRF B m/c - Multiple Points (08-Nov-04) Max Amp 4.27 Plot Scale

TRF B m/c -F2V

RMS Veloc city in mm/Sec

5

• It It is clear that the is clear that the amplitudes are common to both b i bearing – I b d Inboard and Outboard in all radial plains

TRF B m/c -F2H

TRF B m/c -F1V

0

0

TRF B m/c -F1H 8000


24000

Imbalance Case Study 1 Imbalance Case Study 1 • The trend data for the 1xTs parameter has been steady for a considerable amount of time. The last two readings has shown a s g ca t c ease a p tude a significant increase in amplitude Bm/c - TOP RECIRC FAN TRF B m/c -F2H Fan Outboard Horizontal

6

Trend Display of 1xTS

RMS Velocity y in mm/Sec

5 -- Baseline -Value: .428 Date: 24-Nov-00

4

• The The fan was fan was recommended to be cleaned at the next available opportunity il bl t it and for it to be re‐ tested afterwards 3

2

1

FAULT

ALERT

0

0

300

600 900 Days: 24-Nov-00 To 08-Nov-04

1200

1500

Date: 08-Nov-04 Time: 14:16:38 Ampl: 4.688

Imbalance Summary Imbalance Summary • Diagnostic Rules for Imbalance Diagnostic Rules for Imbalance – Periodic non‐impacting sinusoidal waveform – Spectral peak at 1xTs (1 Order) Spectral peak at 1xTs (1 Order) – Very little axial vibration – Similar amplitudes between horizontal and Si il lit d b t h i t l d vertical plains – Synchronous fault type Synchronous fault type – Amplitudes will increase with speed – Very low harmonics of 1xTs V l h i f1 T

Misalignment

Misalignment • When two mating shafts do not share the same collinear axis then misalignment is induced.

• Misalignment is one of the primary reasons for premature machine failure. The forces that are exerted on the machine and its components when in a misaligned state are greatly p g increased from normal operating conditions

Misalignment • Operational Deflection Shape (ODS) is a technique that machine movement based upon the phase and magnitude of data co ected o t e a a yse S o be o s a data collected from the analyser. Shown below is an image age from the ODS illustrating the forces that are exerted onto the machine and components when running in a misaligned condition

Misalignment • Misalignment can be broken into three basic categories, these are: •

Angular – Where the shaft centrelines cross producing a 1xTs peak axially



Offset – Where the shaft centrelines are parallel but they do not meet producing p g a radial 2xTs p peak



More commonly seen – A combination of the above

Misalignment g

Misalignment g •

•

Another common problem associated with alignment is ‘b i ‘bearing misalignment’. i li t’ Bearing misalignment occurs when the bearings are not mounted in the same plain t d i th l i possibly due to: – one or more of the bearings being cocked in the housing being cocked in the housing – The machine itself distorts due to thermal growth or soft foot conditions – Misalignment at the drive causes shaft bending.

Misalignment • Diagnostic Rules for Misalignment – High axial levels of vibration at 1xTs g – High radial levels of vibration at 2xTs – Repeatable period sine waveform showing 1 or 2 clear peaks per revolution (Most likely “M” revolution (Most likely M or or “W” W shape) shape) – Data can usually be seen across the coupling

• Diagnostic Rules for Bearing Misalignment Diagnostic Rules for Bearing Misalignment – High levels of vibration at 1xTs and 2xTs – Repeatable periodic sine waveform showing 1 or 2 clear peaks per revolution l ti – Data usually shown either the driver or driven component

g p Offset Misalignment Spectral Data • The spectral data shown represents a simple misalignment plot. – The primary cursor denotes the 1xTs peak while the harmonic cursors indicate a larger 2xTs peak. This type of data is common to that of Offset Misalignment ST.1 - Raw Water Pump P029 -M2H

7

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

RM MS Velocity in mm/Sec c

6


5

4

3

2

1

0 0

10000


40000

50000

Freq: 2925.0 Ordr: .983 Spec: 2.046

g g p Angular Misalignment Spectral Data • The spectral data below represents a simple misalignment plot. – The primary cursor denotes the 1xTs peak while the data was taken in the axial direction. This type of data is common to that of Angular Misalignment B29 - PUMP NO 3 3601PUM003-M2A Motor Inboard Axial

8

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

7

OVERALL= 6.33 V-DG RMS = 6.31 LOAD = 100.0 RPM = 1071 1071. (17 (17.84 84 H Hz))

RMS S Velocity in mm/Sec

6

5

4

3

2

1

0 0

30

60 Frequency in kCPM

90

120

Freq: 1.071 Ordr: 1.000 Spec: 5.966

g Offset Misalignment Waveform Data • The waveform above is showing two clear peaks per revolution of the shaft. This type of waveform resembling an ‘M’ or ‘W’ shape is common to offset misalignment. o s ape s co o to o set sa g e t – Data shown in velocity

ST.1 - Raw Water Pump P029 -M2H

40

Waveform Display 26-MAR-93 13:32:52

Velocity in mm/Sec

30

RMS = 17.00 LOAD = 100.0 RPM = 2996. RPS = 49 49.93 93

20

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

10

0

-10

-20

-30 0

0.5

1.0

1.5

2.0 2.5 3.0 3.5 Revolution Number

4.0

4.5

5.0

Misalignment Waveform Misalignment Waveform • The waveform data shown above is predominantly showing one sinusoidal waveform per revolution of the shaft. – Here the data is shown Acceleration Here the data is shown Acceleration

B29 - PUMP NO 3 3601PUM003-M2A 3601PUM003 M2A Motor Inboard Axia

0.8


0.6

PK = .2596 LOAD = 100.0 RPM = 1071. (17.84 Hz)

Acceleration in G-s

0.4

PK(+) = .6277 PK(-) = .5683 CRESTF= 3.42

0.2

0.0

-0.2

-0.4

-0.6

-0 8 -0.8 0

0.4

0.8

1.2

1.6 2.0 2.4 Revolution Number

2.8

3.2

3.6

Rev : .680 Ampl: -.306

y Case Study 3 – Kiln Main Motor Gearbox • Introduction • The Kiln drive gearbox motor had been replaced during a The Kiln drive gearbox motor had been replaced during a planned plant shutdown. • During the start up of the plant after the shutdown it was noted that the motor and gearbox were excessively noisy. Vibration data was taken during the run up of the plant to determine the cause of the problem. p

Main Motor

Kiln Gearbox

y Case Study 3 – Kiln Main Motor Gearbox • The spectral plot shown above is the data taken from the drive end of the motor. Here there is a dominant 2xTs peak.

0804

2.4

Route Spectrum 29-Mar-01 11:33:43

In addition to the misalignment the excessive forces being applied to the machine were causing excessive loading on the gears. 2.1


Max Amp 5.98

RM S V Velocity in mm /Sec

1.8

Amplitude - Mixed Units



04 - Kiln Drive -M2H Motor Inboard Horizontal

1.5

0804

04 - Kiln Drive -G2A Shaft 01 Outboard Axial

1.2

0.9

After Shutdown

0.6

0.3

26-Mar-01 12:11: 12 23-Jan-01 15:02: 00

0 0

20000

Before Shutdown 40000

25-Oct-00 09:04: 17

60000 08-Aug-00 14:06: 56

Frequency in CPM 0

5 4 3 2 1 0 29-Mar-01 09:40: 20 29-Mar-01 09:40: 09

60


240

300

Freq: 1418.3 Ordr: 1.000 Spec: 1.346

y Case Study 3 – Kiln Main Motor Gearbox • During data collection it was also observed that the grouting around the front feet of the motor had begun to crack as a result of the excessive force being applied to the motor base and feet due to the misalignment. • Conclusion – It was confirmed the engineers that replaced the motor during the shutdown and assumed as the motor was a like for like swap, as long as th k t th hi i th they kept the shims in the correct place then alignment was not t l th li t t necessary. – Corrective action was required and production was stopped so the motor could be re‐aligned and the mountings re‐secured. t ld b li d d th ti d

Misalignment Summary Misalignment Summary • Diagnostic Rules for Misalignment Diagnostic Rules for Misalignment – Periodic non‐impacting sinusoidal waveform with 1 or 2 clear peaks per revolution (Most Likely “M” 1 or 2 clear peaks per revolution (Most Likely M or “W” shape) – Spectral peak at 1xTs and 2xTs Spectral peak at 1xTs and 2xTs – Axial vibration at 1xTs – Synchronous fault type Synchronous fault type – Data can be seen across the coupling or across the component itself component itself

Looseness

How would looseness ?

Looseness • Looseness can be broken down into two main categories, Structural and Component  Structural looseness occurs when there is free movement within the machines support structure causing excessive vibration. This can be a result of: – Loose support bolts to the components feet and supports – Cracked welds – Deterioration of the base itself. 

Component looseness generally occurs when there is excessive clearance to the components p within the machine, such as: – Excessive clearance between the shaft and bearings – Excessive clearance between the shaft and an impeller etc etc.

Looseness • Diagnostic Rules for Looseness – – – – – – – –

Multiple harmonics of the 1xTs peak ‐ p p Structural Multiple Harmonics of the component that is loose ‐ Component Number of harmonics will increase as the looseness progresses Random non periodic waveform Structural Random, non‐periodic waveform ‐ Waveform shows predominant impacts ‐ Component Raised noise level around the 1xTs + harmonics Half harmonics may also be present Can be present in all Directions

p ((Structural)) Looseness Spectral Data • The spectral plot shown is demonstrating Looseness. – The 1xTs peak has been highlighted by the primary cursor and the p g g y p y relevant harmonics have been displayed. – Multiple harmonics of 1xTs are shown up to around 10 orders of 1xTs. M4441

4.0

40 - Kiln Main Drive -G2H Shaft 01 Outboard Horizontal Route Spectrum 06-Nov-02 11:02:11

3.5 OVERALL= 5.22 V-DG RMS = 5.22 LOAD = 100.0 RPM = 635. (10.58 Hz)


30 3.0

2.5

2.0

1.5

1.0

0.5

0 0

200

400 600 Frequency in Hz

800

1000

Freq: 10.58 Ordr: 1.000 Spec: 3.088

p ((Component) p ) Looseness Spectral Data • The spectral plot shown is demonstrating rotational Looseness. – The primary cursor is on 5xTs peak • The 5 Order peak is vane pass frequency (5 vanes on the impeller)

– Multiple harmonics of 5xTs are shown indicating the impeller has p g p come loose. Ex 9

1.5

L1 - Example 9 -P2A Pump Outboard Axial

Route Spectrum* 17-Aug-01 08:52:02



RMS V Velocity in mm/Sec

1.2

0.9

The raised noise level around the vane pass frequency is common to a pumping problem known as Cavitation OVERALL= 6.62 V-DG RMS = 6.13 LOAD = 100 100.0 0 RPM = 2974. (49.57 Hz)

– This would be the likely cause of the impeller problem

06 0.6

0.3

0 0

40

80

120 160 Frequency in kCPM Label: Centrifugal Pump - Medium

200

240

Freq: 14.88 Ordr: 5.002 Spec: .742

Looseness Waveform Data • Here the waveform is demonstrating a lot of energy and appears to be more random and non‐periodic. – Displaying the waveform in velocity may help to show the random non‐periodic pattern. M4441

1.2

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

0.8

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

Acceleration in G-s

0.4

0.0

-0.4

-0.8

-1.2 0

50

100

150


300

350

400

Looseness Trend Data Looseness Trend Data • Here the trend plot is showing the parameter labelled as the 3‐ 15xTs. This is measuring the amount of energy from 3 orders to 15 orders, which is where the harmonics of looseness will 5 o de s, c s e e t e a o cs o oose ess appear.

M4441

8

40 - Kiln Main Drive -G2H Shaft 01 Outboard Horizontal Trend Display of 3-15xTS

7

-- Baseline -Value: .837 837 Date: 28-Feb-02


6

5 FAULT 4 ALERT 3

2

1

0 0

10

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

40

50

Case Study 4 Case Study 4 – Reciprocator Fan Reciprocator Fan • Introduction • Data had been collected on the following fan for several Data had been collected on the following fan for several months as part of a routine periodic vibration routine. During a routine visit to the machine it was observed that there was a lot of low frequency activity showing around the bearing on lot of low frequency activity showing around the bearing on the inboard of the fan (F1H)

Case Study 4 Case Study 4 – Reciprocator Fan Reciprocator Fan • The multiple plots shown above indicate the change over time from the data taken on F1H. – It is quite apparent that the data shown here is indicating multiple harmonics of the 1xTs frequency (the rise energy as you move further away from the 1xTs). – This type of data is common to that of a looseness problem. M2237

40 - Precip Fan -F1H Fan Inboard Horizontal 2.4

Max Amp 2.74

2.0 1.6 1.2

RMS Velocitty in mm/Sec

0.8 0.4 0 29-Oct-02 11:00:02 18-Sep-02 09:13:26

29-Aug-02 15:30:18

22-Aug-02 11:14:48 0

300

600 Frequency in Hz

900

1200

29-Oct-02 11:00:02 RPM= 998.9 Freq: Ordr: Sp 4:

16.65 1.000 2.811

y p Case Study 4 – Reciprocator Fan • The waveform data taken for this particular point is not showing a random type of waveform pattern which you would expect from Structural looseness, ou d e pect o St uctu a oose ess, but but there is a more t e e sa o e a repeatable (timed interval) pattern. M2237

3

40 - Precip Fan -F1H Fan Inboard Horizonta Analyze Waveform 18-Sep-02 09:24:16

2

RMS = .3747 LOAD = 100.0 RPM = 998. (16.63 Hz)

Acceleration in G-s A

1

PK(+) = 2.36 PK(-) = 2.83 CRESTF= 7.55

0

-1

-2

-3

-4 0

100

200

300


600

700

800

Case Study 4 Case Study 4 – Reciprocator Fan Reciprocator Fan • This type of waveform would more be indicating Component looseness and may indicate a problem with a loose bearing. • Conclusion • It was recommended that the bearing should be inspected at the next available opportunity. – Upon Upon inspection it was found that the bearing was a inspection it was found that the bearing was a ‘Taper‐Lock’ Taper Lock bearing and the taper lock was loose, thus resulting in excessive clearance between the bearing and the rotor.

Looseness Summaryy • Diagnostic Rules for Looseness Diagnostic Rules for Looseness – Multiple harmonics of the 1xTs peak – Number of harmonics will increase as the looseness progresses – Random, non‐periodic waveform – Structural – Waveform shows predominant impacts ‐ f h d i i Component – Raised noise level around the 1xTs + harmonics – Half harmonics may also be present Half harmonics may also be present – Can be present in all Directions

Introduction to Vibration Analysis Fault Diagnostics Fault Diagnostics Gears, Bearings, Peakvue, Electrical, Belts, Resonance

Gear Defects Gear Defects • There are many different types of gears and gear combinations available for various speed and power requirements. • Regardless of gear type they all produce the same basic Regardless of gear type they all produce the same basic vibration patterns and characteristics when a defect is present

• The following topic will discuss the basic characteristics for the following types of gears: – Spur Gears – Helical Gears – Bevel Gears

Spur Gears Spur Gears • Spur Gears are most commonly thought of when diagnosing gears. The teeth are cut parallel to the shaft. These gears are good at po e t a s ss o a d speed c a ges but a e o s e good at power transmission and speed changes but are noisier than other gear types.

• Spur Gear Advantages – High efficiency – Low heat generation L h i

• Spur Gear Disadvantages – Can be very noisy y y

Helical Gears Helical Gears • Helical Gears have teeth cut at an angle to the shaft. These gears are much quieter than spur gears but due to the angular nature of the gear meshing, axial thrust and therefore axial vibration is higher than those of spur gears – Sometimes Sometimes to counter act the axial thrust these gears can be double up to counter act the axial thrust these gears can be double up and are known as ‘Double Helical’ or ‘Wishbone Gears’

• Helical Gear Advantages – Quiet Operation Quiet Operation

• Helical Gear Disadvantages – Less power transmission efficiency and greater heat generation than and greater heat generation than spur gears – Axial loading on bearings

Bevel Gears Bevel Gears • Bevel Gears are used to transmit power and speed to an output shaft perpendicular to the drive shaft. These gears use a bevel design to transmit the power better. – These gears are most commonly seen on right angle gearboxes (where the input shaft is at 90 degrees to the output shaft) the input shaft is at 90 degrees to the output shaft)

• Bevel Gear Advantages Bevel Gear Advantages – Converts the direction of power transmission

• Bevel Gear Disadvantages – Less efficient – Higher heat generation

Gear Analysis Gear Analysis •

Vibration analysis of gears can provide a wealth of information about the mechanical health of the gears. This section discusses the basic frequencies that may be present within a gearbox.

• Gear Mesh Frequency Spectral Data G M hF S t lD t •

•

The gear mesh frequency (GMF) refers to the frequency at which to mating gears interact with each other and is the most commonly discussed gear frequency. However, GMF by itself is not a defect frequency. The GMF should always be present in the spectral data regardless of gear condition. What is important is the amplitude as this may vary depending upon gear condition or loading of the gear.

Gears • Two mating gears will generate a frequency known as the GMF and will show in the spectral data regardless of gear condition. co d t o M4441

1.2

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


0.9

0.6

0.3

0 0

200


800

1000

Freq: 386.98 Ordr: 15.00 Spec: .864

Calculating GMF Calculating GMF – Single Reduction Single Reduction • Single Reduction Gear Train – The GMF is simply defined as the number of teeth on a gear multiplied py g p by its turning speed

GMF = (#teeth) x (Turning speed) • Example: E l – Consider the following gear train, INPUT

OUTPUT

Input

= 1490RPM

Gear 1

= 44 Teeth

Gear 2

= 71 Teeth

GMF = #teeth x turning speed GMF = 44teeth x 1490 RPM GMF = 65560 CPM

or 65560/60 = 1092.6 Hz

Calculating GMF Calculating GMF – Multi Reduction Multi Reduction • Calculating the GMF for gearboxes that have multiple trains use the following. GMF = (#teeth) x (Turning speed) Gear Ratio = (#teeth in) / (#teeth out) Speed out = (Speed in) x (Gear Ratio) Speed out = (Speed in) x (Gear Ratio)

• Example: – Consider the following gear train: INPUT

OUTPUT

Input

= 1490RPM

Gear 1 Gear 2

= 15 teeth = 21 teeth

Gear 3 Gear 4


Calculating GMF Calculating GMF – Multi Reduction Multi Reduction INPUT

Input

= 1490RPM

Gear 1 Gear 2


Gear 3 Gear 4


OUTPUT

Gear Ratio 1 p Out Speed

= 15 teeth / 21 teeth = 1490 RPM x 0.714

= 0.714 = 1064.28 RPM

Gear Ratio 2 Speed Out

= 19 teeth / 54 teeth = 1064.28 RPM x 0.351

= 0.351 = 374.47 RPM

GMF 1 = 1490 RPM x 15 teeth = 22350 CPM GMF 2 = 1064.28 RPM x 19 teeth = 20221.32 CPM

GMF Calculation Exercise GMF Calculation Exercise • Using the formulas on P153 from the manual calculate: – Speeds of all shafts – All GMF from the following gearbox arrangement All GMF from the following gearbox arrangement INPUT

OUTPUT

• • • • • •

Input

= 1000 RPM

Gear 1 Gear 2


Gear 3 Gear 4


Gear Ratio 1 = 10/40 Shaft 2 speed = 1000 x 0.25 Gear Ratio 2 Gear Ratio 2 = 10/20 10/20 Shaft 3 Speed = 250 x 0.5 GMF 1 = 1000 x 10 GMF 2 = 250 x 10

= 0.25 = 250 RPM = 0.5 05 = 125 RPM = 10000 CPM = 2500 CPM

Gears Gears – Sideband Frequencies Sideband Frequencies • Sidebands are the most common indication that a gear is defected. – Sidebands are equally spaced frequencies in the spectral data that materialise either side of the main GMF peak. – The sideband frequency spacing is equal to either the turning speed of The sideband frequency spacing is equal to either the turning speed of the input gear or the turning speed of the output gear.

• Sidebands show in the data when either the gear is worn, loose or eccentric. – The speed of the shaft with the bad gear on it will p produce the most dominant sidebands in the spectral data.

Gears • The spectral data shows GMF with sideband data. – The sidebands are equally spaced at intervals of 310 CPM. This is indicating the gear that rotates at 310 RPM is the one that is worn or g g damaged. X401A

1.0

FPP - SAND MILLS (OLD)A -G3A Shaft 02 Inboard Axial

Route Spectrum 07-Nov-02 09:11:53 (SST-Corrected)

RMS Velocity in m mm/Sec

0.8

GMF

OVERALL= 2.18 V-DG RMS = 2.17 LOAD = 100.0 RPM = 310. (5.17 Hz)

0.6

0.4

Sidebands 0.2

0 0

8000

16000 Frequencyin CPM

24000

Freq: 18363. Ordr: 59.23 Spec: .564 Dfrq: 310.82

Gears Gears – Waveform Data Waveform Data • Gears can produce different types of waveforms, the one shown below is indicating gear wear. – As As the defective teeth come into mesh the noise generated increases the defective teeth come into mesh the noise generated increases showing an increase in amplitude in the vibration data X401A

1.5

FPP - SAND MILLS (OLD)A -G3A Shaft 02 Inboard Axial Route Waveform 07 Nov 02 09:11:53 07-Nov-02

1.2

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

0.9

Acceleration in G-s

0.6

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

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

1

2


5

6

Case Study 5 Gearbox Case Study 5 –Gearbox • The following case study is from a motor gearbox unit that drives a roller. – Product Product (Fibre) is fed along the top of the roll while being washed (Fibre) is fed along the top of the roll while being washed through a series of baths. – There are several of these Wash Nip Rollers in a continuous stream, failure of any one of them results in lost production

•

Data is collected on a Data is collected on a fortnightly basis as part of a routine data collection route

Case Study 5 Gearbox Case Study 5 –Gearbox • The spectral data shown below is taken from the motor in the axial direction – (As (As the motor is mounted directly into the gearbox the first helical gear the motor is mounted directly into the gearbox the first helical gear is mounted on the end of the motor shaft). L1NG - WASH LINE NIP UNIT 3 3-32J03 -MIA MOTOR INBOARD AXIAL

0.6

•

The GMF is highlighted by Th GMF i hi hli ht d b the primary cursor at 49 Orders Th f lt f The fault frequency data d t (dotted lines) indicate the sideband data showing gear wear on the first gear gear wear on the first gear in the gear train RMS Velocity in mm/Sec

•

Route Spectrum 01-Aug-04 01 Aug 04 10:21:41

EEEEE EEEEE

0.5

OVERALL= 1.08 V-DG RMS = 1.07 LOAD = 100.0 RPM = 1175. (19.58 Hz)

0.4

>REN Wash Nip E=Gm(1>2)-S1

0.3

0.2

0.1

0 0

10

20

30

40 50 Frequency in Orders

60

70

80

Ordr: 49.00 Freq: 57551. 57551 Spec: .275 Dord: .00649

Case Study 5 Gearbox Case Study 5 –Gearbox • The waveform data is showing a distinct pattern commonly associated with gears. • The amplitude increases In noise as the damaged teeth come The amplitude increases In noise as the damaged teeth come into mesh – Producing over 2G‐s of force in both the positive and negative direction

Case Study 5 Gearbox Case Study 5 –Gearbox • The gears were inspected due to the critical nature of the asset. It was found the gear to be severely damaged. • A new gearbox was fitted and new data was taken showing the A new gearbox was fitted and new data was taken showing the difference between the good and bad gear

Bearing Defects g Rolling Element g Plain Bearings Peakvue

Rolling Element Bearings Rolling Element Bearings • Rolling element bearings have specific bearing failure modes that can be observed in the spectral and waveform data. • Bearing frequencies differ from most other frequencies present within the spectral data because unless the bearing present within the spectral data because unless the bearing has a defect there will be no frequency peaks in the data relating to the bearing. Only if the bearing has a defect will frequencies show in the spectral data.  There are four main fundamental bearing defect frequencies q these are:

g g Rolling Element Bearings

Outer Race

Inner Race

How Bearing Faults Generate Vibration g

How Bearing Faults Generate Vibration g

Rolling Element Bearings Rolling Element Bearings •

Bearing defect frequencies are calculated based upon the geometry of the bearing these calculations may include: – – – –

Number of rolling elements Pitch Circle Diameter Rolling element diameter Contact angle • Defined within Machinery Health Manager there are over 100000 predefined bearing stored in the CSI bearing warehouse BEARINGS in CSI Warehouse: c:\RBMsuite\SysData\CSI_CMP.WH

**************************************************** BRG ID Bearing Type #B/R FTF BSF BPFO BPFI 12143 RHP 6218 11 0.418 2.967 4.598 6.402 24421 SKF 6313E 8 0.376 1.894 3.009 4.991 25372 SKF I 26313 25372 SKF I‐26313 19 0.433 3.568 8.219 10.781 19 0 433 3 568 8 219 10 781

Rolling Element Bearings Rolling Element Bearings • Characteristics of Bearing Defects – High High frequency raised noise level (Hump of energy) frequency raised noise level (Hump of energy) – Non‐Synchronous harmonic peaks (Both low and high frequency) – Time waveform will show a lot of noise/impacting – Early stages of bearing wear may show better if viewed in acceleration in the frequency domain l ti i th f d i – Fundamental bearing defect frequency (First calculable q y) y p p frequency) may not be present in the spectral data

Failure Mode 1 Failure Mode 1 •

The early stages of bearing defects produce low amplitudes of vibration at higher frequencies – (Appears on the right hand side of the spectrum). ( pp g p )

•

These are normally humps of energy or peaks that are harmonics to the fundamental frequency. – (The fundamental frequency should not be visible at this stage). i ibl hi )

Failure Mode 2 •

Distinct harmonics of Non‐Synchronous peaks appear. – (These should appear lower down the scale of the spectrum – towards the left / middle of the plot)

•

Sidebands may appear around these frequencies usually equating to turning speed. – (The fault frequencies may not match exactly with the peaks in the spectrum due to the fact that the bearing geometry will have changed) bearing geometry will have changed).

Failure Mode 3 •

The fundamental frequency normally appears at this stage – (First calculable frequency of the bearing – towards the left‐ hand side of the spectral plot). This is classed as advanced stages of bearing wear.

•

Sidebands may be visible that equate to other bearing frequencies – BSF, FTF etc). )

Failure Mode 4 Failure Mode 4 •

The bearing degrades so much that the spectrum The bearing degrades so much that the spectrum becomes a mass of noise. At this point the bearing will fail at any point (If it last this long – most fail around Mode 3).

g g BPFI Rolling Element Bearings ‐ •

Typical data showing a defected inner race – Fundamental frequency showing – Harmonics low and high frequency + sidebands

g g BPFO Rolling Element Bearings ‐ •

Data showing a defect related to the BPFO – The fundamental frequency is showing – Harmonics from low to high frequency Harmonics from low to high frequency

g g BSF Rolling Element Bearings ‐ •

Bearing defect showing the BSF – Rolling elements – Sidebands around the BSF = FTF

Rolling Element Bearings Rolling Element Bearings ‐ FTF •

The FTF is the only bearing frequency that is sub‐synchronous – May not detect then with conventional vibration data – FTF defect at 0.4 orders shown in Peakvue

• Bearing

FTF & BSF FTF & BSF

BPFI & BPFO BPFI & BPFO

Rolling Element Bearings Rolling Element Bearings ‐ Waveform • As a bearing becomes defected then the amount of noise/force generated as the rolling elements impact the de ect e a ea c eases defective area increases. – This can show significant G‐levels in the time waveform. This value is trended in the software as the Peak‐Peak value

•

This data is taken from a pump with a damaged bearing – The force levels are reaching 40G‐s

y g Case Study 6 – Bearing Defect •

The spectral plot below is showing the data from the inboard vertical direction of the motor. – The primary cursor is indicating the fundamental defect BPFO f BPFO frequency + harmonics. +h i – The frequency range of the harmonics covers both low and high frequency ranges suggesting the bearing is more advanced stages of failure.

Case Study 6 Case Study 6 – Bearing Defect Bearing Defect • The time waveform is showing significant impacting levels reaching in excess of +/‐ 8G‐s of force. – This This level of impacting is higher than would be suspected for a motor of level of impacting is higher than would be suspected for a motor of this type. •

The repetitive impacting The repetitive impacting pattern shown above is common to antifriction bearing defects. g – In this instance the impacting is representing the rolling elements striking a defect on the race.

Case Study 6 Case Study 6 – Bearing Defect Bearing Defect • The trend plot above is showing the increase in amplitude of the Peak‐Peak parameter. – The The peak peak‐peak peak parameter is measuring the amount of energy in the parameter is measuring the amount of energy in the time waveform from the Peak+ to the Peak‐ • •

Conclusion C l i The motor was reported as having a bearing defect to the engineering group. As the f d fundamental defect frequency was present t ld f tf t and the trend had shown sudden increases it was recommended to change the bearing at the next available opportunity the next available opportunity.

Bearing Defects g Rolling Element

Plain Bearings Peakvue

Plain Bearings Plain Bearings • Rotating elements are not used in sleeve (plain) bearings; rather the shaft rides on a layer of lubricating oil inside the bea g jou a bearing journal. – Therefore the fundamental frequencies seen from antifriction bearings do not apply to sleeve bearings.

• Since there is no contact between the bearing and the shaft monitoring of sleeve bearings for vibration analysis usually requires the use of displacement probes mounted 45 p p degrees either side of top dead centre.

Plain Bearings Plain Bearings • As there are no rotating components in the bearing that produce high frequency noise (force) there is no need to monitor a high frequency range. Usually 10 to 15 orders of turning speed will be sufficient. • Sleeve bearings have specific defects that contribute towards bearing failure, these are: – Excessive clearance Excessive clearance – Hydraulic instability (oil whirl)

Plain Bearings – Spectral Diagnostics • Excessive Clearance – When there is excessive clearance between the rotor and the bearing then this will have an effect on the system vibration. When the bearings have excessive clearance then a ‘looseness’ bearings have excessive clearance then a looseness occurs. occurs The spectral data shown below is indicating a sleeve bearing with excessive clearance excessive clearance.  As the clearance increases then the harmonics of 1xTs will increase and can go up to 10–15xTs.

•

TBT

16

Fu - Turbine Brg Thrust End -R1Y Radial 'Y' Direction

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

OVERALL= 2.93 V-DG P-P = 22.71 LOAD = 100 100.0 0 RPM= 941. (15.69 Hz)

– Like looseness the more harmonics there are the more severe the problem will be. – A good sleeve bearing will still show a few harmonics as there is a small clearance l between b t the th shaft h ft and d bearing

P-P Dis pla c cement in Microns

12

8

4

0 0

3

6 Frequency in Orders

9

12

Ordr: Freq: Spec:

1.000 15.68 7.494

Plain Bearings Plain Bearings – Spectral Diagnostics Spectral Diagnostics • Oil Whirl – One of the major problems encountered with these types of bearings is j p yp g the possibility of hydraulic instability of the shaft within the bearing; known as oil whirl or oil whip. – Oil Whirl is a result of turbulent flow within the oil resulting in the oil pushing the shaft around of centre. TBT

16

•

Fu - Turbine Brg Thrust End -R1Y Radial 'Y' Direction

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

Oil Whi Whirll att 0 0.4 4 orders d

OVERALL= 2.93 V-DG P-P = 22.71 LOAD = 100.0 RPM= 941. (15.69 Hz)

P-P Dis pla c e m e nt in M ic rons

12

•

8

– This defect is sub‐synchronous data. – When the amplitude of the oil whirl is equal to or greater than the 1xTs peak a problem exists

IIn this instance oil whirl can be thi i t il hi l b corrected by: – Properly loading the bearing – Change the oil viscosity – Change the oil pressure Ch th il

4

0 0

The dominant peak within the spectral data will be typically at 0.4 orders. (.40‐ .48)

3

6 Frequency in Orders

9

12

Ordr: Freq: Spec:

1.000 15.68 7.494

Oil Whirl Oil Whirl

Bearing Defects Rolling Element Plain Bearings

Peakvue

Peakvue Processing Peakvue Processing •

The detection of bearing and gear defects is one of the primary expectations of a predictive maintenance program. – As analysts we can spend a lot of time tying to determine these faults. – Peakvue is a process that concentrates on these defects to help the analysts determine potential faults developing

•

Peakvue stands for the Peak Value and is a technique that detects high frequency stress waves generated from metal to metal contact, such as: frequency stress waves generated from metal to metal contact such as: – Bearing defects – Rotating elements striking a defect on the race – Gear defects – Damaged teeth in mesh – It is the detection of these high frequency stress waves that will aid with analysis analysis

Peakvue Processing Peakvue Processing ‐ Filters • In order to capture the stress wave signal the process requires the use of a filter to remove all unwanted noise that can do dominate the data ate t e data

1. Conventional Vibration Signals that are filtered from the Peakvue Signal Imbalance Misalignment Misalignment Gears Bearings Resonance

2. Peakvue filter removing low frequency noise from the stress wave data This is to prevent low frequency noise consuming the stress wave activity

3. High frequency stress wave activity occurring in the 1000Hz 20000Hz frequency range at a rate governed by a low frequency event Bearings Gears

Peakvue Processing Peakvue Processing ‐ Filters • There are two types of filters available • Band Pass Filters Band Pass Filters f

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

• High Pass Filter

f

– The high pass filter removes all data lower in frequency to that of the g p q y filter selection allowing only the high frequency stress waves to pass through

• After After the filtering process what should remain is the high the filtering process what should remain is the high frequency stress wave activity that is occurring at the rate of the excitation – such as from a bearing.

Peakvue Processing Peakvue Processing – Spectral Data Spectral Data • Shown below is a typical Peakvue spectrum with a defect present 

Stress waves are showing clearly in the data at 4.6 Orders 

•

The filter used is shown in the top The filter used is shown in the top right hand corner

Good G d Spectrum S t will ill show only a noise level



Noise removed by y filter

Peakvue Processing Peakvue Processing – Waveform Data Waveform Data • As stress waves are small in amplitude severity of the problem can be judged using the time waveform – Peak Value of force from the impact Peak Value of force from the impact

RMS Acc celeration in G-s

• The waveform can resemble a spectrum as there is no negative half to the data B42 - ZONE 5 DF FAN 1 16/16EXT01-M2P Motor Inboard Horz Peakvue

0.8 0.7

N

N

N

N

N

N

N

N

N

0.6 0.5 0.4 03 0.3 0.2

Route Spectrum 09-Jul-03 09:50:49 (PkVue-HP 1000 Hz) OVERALL= 1.37 A-DG RMS = 1.37 LOAD = 100.0 RPM = 1342 1342. (22 (22.37 37 Hz) >NTN 6217 N=BPFO -OB



For Peakvue analysis



Use the Spectrum

0.1

– Diagnose the defect

0

Acceleration in G-s

0

200


800

1000

Route R t W Waveform f 09-Jul-03 09:50:49 (PkVue-HP 1000 Hz) RMS = 2.97 PK(+) = 8.35 CRESTF= 2.81

8 7 6 5 4 3 2 1 0 0

4

8

12

16 20 24 Revolution Number Label: Bearing Fault - BPFO NTN6217

28

32

36

Freq: 1.250 Ordr: .05587 Spec: .01367



Use the Waveform – Determine the severity

y Case Study 7 – Peakvue on Fan Bearingg • The following machine is a pre‐heater pre heater fan designed to fan designed to heat the product prior to it entering a Kiln – There is no standby for this machine – Failure results in stopped production

• The following data was taken from the above fan unit. – The problem bearing resided on the fan inboard bearing. – Data was collected on a monthly basis. Both conventional vibration data and Peakvue data were taken during the route collection.

y Case Study 7 – Peakvue on Fan Bearingg • The data shown below is taken using conventional vibration methods on the inboard bearing of the fan – 1x peak is highlighted showing amplitudes of 4mm/sec 1x peak is highlighted showing amplitudes of 4mm/sec – Waveform is showing less than 1G of force both +/‐ 40 - Preheater Fan M4425 -F1H Fan Inboard Horizontal

R M S Ve lo c it y in m m /S e c

5

Route Spectrum 29-Oct-02 11:19:26 OVERALL= 4.18 V-DG RMS = 4.18 LOAD = 100.0 RPM= 825. (13.75 Hz)

4 3 2 1 0 0

10

20

30 40 50 Frequencyin Orders

60

70

80

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

1.5

• There There are indications of are indications of bearing frequencies showing high frequency Route Waveform 29-Oct-02 11:19:26 RMS = .3837 PK(+/-) = 1.19/1.05 CRESTF= 3.11

10 1.0 0.5 0 -0.5 -1.0 10 -1.5 0

1


4

5

Ordr: 1.000 Freq: 13.75 Spec: 3.721

– These may be missed due to y the amplitude of the 1x peak

y Case Study 7 – Peakvue on Fan Bearingg • The Peakvue data above is taken from the same point as the previous data.

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

– This particular reading is using a 1000 Hz High Pass filter. This particular reading is using a 1000 Hz High Pass filter. 40 - Preheater Fan M4425 -F1P Fan Inboard Horz Peakvue

0.7 0.6

F

F

F

Route Spectrum 29-Oct-02 29 Oct 02 11:15:59 (PkVue-HP 1000 Hz) OVERALL= 1.10 A-DG RMS = 1.10 LOAD = 100.0 RPM= 830. (13.84 Hz) >SKF 22240CC F=BPFO -IO

F

0.5 0.4 0.3 0.2 01 0.1

•

– This is not non‐synchronous d data and the frequency d h f matches that of the BPFO for the bearing.

0 0

5

10

15 20 25 Frequency in Orders

30

35

40

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

8 7

Route Waveform 29-Oct-02 11:15:59 (PkVue-HP 1000 Hz) RMS = 3.31 PK(+) = 7.47 CRESTF= 2.25 DCoff = -3.08

6 5 4 3 2 1 0 0

10


40

50

Ordr: 8.176 Freq: 113.14 Spec: .194

Here the data is showing there H th d t i h i th is stress wave activity at 8.176 orders.

•

The waveform data is measuring over 7 G‐s of force i 7G ff as oppose to the 1G from the previous data.

Case Study 7 – Peakvue on Fan Bearing • Conclusion • There is significant bearing damage relating the outer race of There is significant bearing damage relating the outer race of the bearing. • As the machine was critical to the process, the bearing was changed on the next available opportunity that tied in with process requirements.

Electrical Defects Electrical Defects

Electrical Defects Electrical Defects • A motor can be simply broken down into two key components – Rotor – Stator

•

The stator is stationary The stator is stationary – Consists of wire wound in coils and placed in slots of an iron core. – The stator produces a rotating magnetic field.



The rotor is not stationary – Consists laminations with solid conductors called rotor bars – A circular flow of current through these rotor bars causes the rotor to become an electromagnet which will rotate in a magnetic filed.

Electrical Defects Electrical Defects – Spectral Data Spectral Data • The most common electrical frequency that materialises in the spectral data is the 2 x Line Frequency. – For For most industrial applications the line frequency used to supply most industrial applications the line frequency used to supply motors is 50Hz (Europe). – Therefore the frequency of concern for most electrical faults would be 100Hz (2xLf [Lf=line frequency]) Ex7

0.6

Ex7 - Example 7 -M1H Motor Outboard Horizontal Route Spectrum 08-Nov-00 14:27:35 OVERALL= .5613 V-DG RMS = .5607 LOAD = 100.0 RPM== 2967 RPM 2967. (49 (49.44 44 Hz)

RMS V Ve loc it y in mm /Se c

0.5

0.4

• The spectral plot is showing a peak at 100Hz showing a peak at 100Hz (6000cpm) – 2xLf – This can be mistaken for misalignment

0.3

0.2

0.1

0 0

500

1000 Frequency in Hz

1500

2000

Freq: 100.00 Ordr: 2.023 Spec: .386

Electrical Defects Electrical Defects – Waveform Data Waveform Data • The waveform data from a 100Hz peak will show a sinusoidal pattern like the waveform shown below Ex7

1.5

Ex7 - Example 7 -M1H Motor Outboard Horizontal

10 1.0

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

0.5 Ve loc it y in m m /Se c

• Again this type of pattern can be associated with can be associated with misalignment.

Route Waveform 08-Nov-00 14:27:35

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

0

-0.5

-1.0

-1.5

-2.0 0

1

2

3 Revolution Number

4

5

6

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

Electrical Defects Electrical Defects ‐ Causes • Common fault types that can produce the 2xLf peak are as follows: • Dynamic Eccentricity – Usually Rotor Related • Static Eccentricity – Usually Stator Related • Loose Iron or Slot Defect – Rotor or Stator • Open or Shorted Windings • Insulation Breakdown or Imbalanced Phase I l i B kd I b l d Ph • Loose Connectors

Electrical Defects ‐ Peakvue • Peakvue data also shows electrical defects at the 2xLf peak. – This may be due to the rotor or stator bowing; due to heat build up. y g p

• The spectral plot below is indicating a 100Hz peak using Peakvue with a 1000Hz filter.

Case Study Case Study – Electrical Defect Electrical Defect • The following case study was taken from a glass manufacturer. The data was from the ‘Electric Front Wall Cooling Fan’. – This This fan unit is a critical fan to the process and has no standby unit. fan unit is a critical fan to the process and has no standby unit. – In this particular instance the motor failed shortly after the data was collected.

•

The Peakvue data taken on the motor non‐drive end is showing a dominant 100Hz showing a dominant 100Hz peak. – This frequency is at 2xLf and is associated with electrical problems

Case Study Case Study – Electrical Defect Electrical Defect • The multi‐plot above shows the same measurement point going back over the last 5 route readings. – This particular plot is useful for determining rate of change. This particular plot is useful for determining rate of change. – It is quite clear how this particular frequency suddenly appeared

•

Conclusion – As As the motor failed shortly after the motor failed shortly after data collection no action was taken to prevent failure. – The investigation in the motor showed one of the connectors had come loose causing the motor to burn out.

Belt Defects Belt Defects V‐Belts V Belts Timing Belts

Belt Defects Belt Defects •

Belts are the most common low cost way to transmit power from one shaft to another. – Belt drives rely on friction between the belt and pulley to transmit power between drive and driven shafts

•

The ability of belt to transmit power depends upon The ability of belt to transmit power depends upon 1. 2. 3. 4.

Belt Tension (tension on the belt holds it tightly against the sheave) Friction between the belt and sheave The arc of contact between the belt and sheave (Wrap) The speed of the belt

• However, belts can be easily damaged by heat, oil and grease and since belts slip with in the sheaves they can not b be used where exact speed changes are required (except for d h t d h i d( tf timing belts)

Belt Defects Belt Defects • Belt defects can be considered non‐critical faults by many maintenance groups due to the relative ease of replacement requiring minimum downtime. equ g u do t e – But belt defects are a major contributor to the overall vibration of the machine resulting in premature failure of other machine components.

Sources of belt drive defects Poor Maintenance Enviromental Factors Poor Installation Poor Design g Other Defects

Belt Defects Belt Defects – Belt Types Belt Types • There are many different types of belt drive systems. This section covers the most commonly used types of belt in industry today. dust y today • •

V‐Belts V‐belts are the most common type of belts used. They are ‘V’ shaped in cross‐section cross section, this allowing the belt to wedge against the side of the this allowing the belt to wedge against the side of the sheave. – This design allows the belt to be run faster than most other type of belt applications with power transmission efficiencies as high as 95%

Belt Defects • Timing Belts • These are flat belts with equally spaced teeth that mesh These are flat belts with equally spaced teeth that mesh with notches on the pulley. Timing belts are different from other belt drives as they do not induce any slip. – M Most commonly used where constant velocity and strict timing l d h l i d i i i application is required.

Belt Defects Belt Defects – Fault Characteristics Fault Characteristics • Belt defects, such as cracks, broken or missing pieces, hard and soft spots can generate vibration at the turning speed of the belt (1xbelt) and harmonics – Due to the length of the belt in relation to the pulleys (sheaves) the 1xbelt frequency is sub‐synchronous 1xbelt frequency is sub synchronous and very often the 2xbelt and very often the 2xbelt frequency may be sub‐synchronous as well

• The predominant harmonic is typically the 2xBelt frequency and can be seen in the radial plain in‐line with the belts. – Severity is judged by the number and amplitude of the harmonics seen in the spectral data

Belt Defects Belt Defects – Fault Characteristics Fault Characteristics • Just like two mating shafts, belt drive systems can also be misaligned in both angular and offset directions. – When When misalignment is induced into a belt drive system then the life of misalignment is induced into a belt drive system then the life of the belt is significantly reduced as well as the overall vibration of the system increases.

Offset Misalignment Angular Misalignment

• Pulley misalignment results in high axial vibration at the shaft turning speed. – If If the belt is also defected then 1xbelt frequency and harmonics may the belt is also defected then 1xbelt frequency and harmonics may also show in the axial direction

Belt Defects Belt Defects – Calculations Calculations • The fundamental belt frequency can be calculated using the following equation: Belt Freq. = (3.142 * Pulley Ts * Pulley PCD) Belt (Length) – Where: • Ts = Turning Speed • PCD = Pitch Circle Diameter • Note: The PCD and belt length must be in the same units

• A timing will belt will also have a specific frequency related to the number of teeth on the pulley Timing Belt Freq. = (Pulley Ts) * (# Pulley Teeth)

Belt Defects Belt Defects – Calculation Example Calculation Example • • • •

Belt Frequency Calculation Belt Frequency = (3.142 Belt Frequency (3.142 * 1480 1480 * 300) / (2000) 300) / (2000) Belt Frequency = (1395048) / (2000) Belt Frequency = 697.524 CPM – This is sub‐synchronous to the 1xTs of the pulley Motor RPM Pulley Diameter Belt Length

= 1480 RPM = 300 mm = 2000mm

Belt Defects Belt Defects – Spectral Data Spectral Data • The spectral data above is data taken of a motor from an Air Handling Unit. – The The frequency highlighted by the primary cursor is showing the 1xTs of frequency highlighted by the primary cursor is showing the 1xTs of the motor (1 Order) •

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

There are a lot of sub‐ synchronous peaks showing in this data this data. – The first peak is the fundamental frequency of the belt rotation. – The second peak is the 2xbelt The second peak is the 2xbelt frequency suggesting there is damage to the belt – As the harmonics of the belt increase in number they surpass the 1xTs of the motor surpass the 1xTs of the motor and in this case the third harmonic becomes non‐ synchronous data.

Case Study 9 Case Study 9 – Belt Defect Belt Defect • The following data was taken on an Air Handling Unit. The Air Handling Unit is a supply fan from shared services. This is a sta d a o e u t t o sta d by capab ty stand alone unit with no stand by capability BL31 - 559 AHU Supply -M2H Motor Inboard Horizontal

559S

0.5 J

J

J

J

J

J

J

J

J

Route Spectrum* 22-Feb-05 13:53:33

J

OVERALL= 1.22 V-DG RMS = .7701 LOAD = 100.0 RPM = 1272. (21.21 Hz)

0.4

>Belt Freqs J=Belt 1 Freq

0.2

0.1

x - Fa n spee d

0.3

X - M ot or s p e e d

RM S Ve loc ity in mm /Sec

•

•

0 0

•

4000

8000 Frequency in CPM Label: Belt defect/worn belts & sheaves

12000

16000

Freq: q 835.69 Ordr: .657 Spec: .04393

The data shows the motor turning speed t t i d along with a sub‐ synchronous peak of the belt frequency. The primary cursor is The primary cursor is highlighting the 1xbelt with several harmonics. The 2xbelt is very The 2xbelt is very dominant suggesting there is damage to the belts.

Case Study 9 Case Study 9 – Belt Defect Belt Defect • As this is a critical machine it was recommended on the next available opportunity that the belts needed to be checked for da age a d e a g ed damage and re‐aligned.

• •

The machine was stopped and the belts were inspected based upon the recommendation. Significant damage was found to several of the belts during this inspection as well as worn pulleys. Both the belts and pulleys were replaced and correctly aligned before re‐starting the machine.

Resonance

Resonance • Resonance is defined as: An excitation of a natural frequency by a periodic forcing excitation of a natural frequency by a periodic forcing function. • All assets contain natural frequencies that vary depending upon the stiffness and mass. – Resonance Resonance can be considered to be a vibration amplifier, that takes the can be considered to be a vibration amplifier that takes the force level of the periodic forcing function and amplifies it; which significantly increases the movement of the asset.

If Vibration is a Fire The Resonance is a Fuel If Vibration is a Fire, The Resonance is a Fuel

Example of Resonance Example of Resonance •

The example shown represents the effect on amplitude of the forcing function when in resonance. – In plot 1 the 1xts is running below the natural frequency (Fn). – Fn can be seen in plot 2. – Plot 3 shows the increase in amplitude of the forcing function when run at the natural frequency – t lf thi i this is resonance Before Excitation

1 Frequency

Resonance Curve

2

Frequency

Amplified Signal

3 Frequency

Resonance •

There are two factors that determine the natural frequency of an asset these are; 1. Mass – The heavier an object the lower the natural frequency 2. Stiffness – The more rigid a structure the higher the natural frequency

• Resonance is becoming more of a problem in industry in ece t yea s due to recent years due to: – Older equipment having to run faster to meet current production demands (often above what it was designed for) – Equipment is being built cheaper and lighter

• This is resulting in amplification of the forcing function creating excessive machine movement resulting premature machine failure.

Effects of Resonance • The ODS data is showing a steel frame structure deflecting at one corner in the vertical direction due to a resonant co d t o condition.

Characteristics of Resonance Characteristics of Resonance • Characteristics of Resonance – Resonance is very directional in nature (Movement may be greater in y ( y g one plain than the other) – Vastly different amplitudes of the forcing function from one direction to the other (between Horizontal and Vertical – Rule of thumb ratio is 3:1 difference) – Resonance is very speed sensitive (small changes in speed can show large differences in amplitude of the forcing function) – Resonance can occur at any frequency but most commonly associated with the 1xTs

Resolving a Resonance Resolving a Resonance • There are a number of alterations to the system that can be made to resolve a resonance condition. – However if structural changes are to be made you need to be careful you don’t excite another natural frequency once the change has been made?

• Once you are sure you have a resonant condition it can be corrected by one of the following methods: – Change the Mass Ch th M – Change the Stiffness – Remove the forcing function – Dampen the structure  Dampening is a method used to convert mechanical energy into thermal energy. It does not remove the resonant condition only gy y controls the amount of movement.

Resonance Resonance – Spectral Data Spectral Data The spectrum is showing the 1xTs peak of the motor with amplitudes reaching 19mm/sec.

•

– This is high for the 1xTs.

Very often this type of data can be mistaken for Imbalance as this defect can also produce a high 1xTs peak.

•

– However Imbalance is a centrifugal force and should show similar amplitudes in both radial plains where as resonance is very directional. 40 - No 1 GCTCompressor M4551 -M2H Motor Inboard Horizontal

27

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

24 OVERALL= 19.95 V-DG RMS = 19.85 LOAD = 100.0 RPM= 1484. (24.73 Hz)

R M S Ve loc it y in m m /Se c

21

18

15

In order to help resolve this issue we need to check the amplitude of the 1xTs 90 degrees to this point (horizontal to vertical) – This can easily be done by using the ‘multi point plot’ in the software

12

9

6

3

0 0

•

500


1500

2000

Freq: 24.72 Ordr: 1.000 Spec: 19.50

Resonance Resonance – Multi Plot Multi Plot The multi point plot allows the analyst to display several measurement points on the same plot. Here we are showing all the radial points from the motor. motor – It is very clear that the amplitudes of the 1xTs peak are excessive in the horizontal direction when compared to the vertical. This is a characteristic of a resonant condition. 40 - No 1 GCTCompressor GCTCompressor M4551

- Multiple Points (13-Feb-03)

24 20

Max Amp 22.0

16 12 8

R M S Vee lo c it y in m m /S e c

•

4 0 M2V 10:15

M2H 10:14

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

500


1500

2000

Freq: Ordr: Sp 3:

25.00 1.011 19.35

Case Study 10 Case Study 10 – Resonance • The following case study is taken from a motor and a reciprocating compressor. The unit is mounted on a steel frame which, in turn sits on spring mounts designed for dampening c , tu s ts o sp g ou ts des g ed o da pe g • Recently the motor had been replaced due to bearing defect; however the new motor was smaller and lighter but defect; however the new motor was smaller and lighter but delivered the same power as the previous motor. • When the compressor was put back into service it was noted there was excessive vibration coming from the unit. The unit was left to run like this for several months until the vibration became to excessive vibration became to excessive.

Case Study 10 Case Study 10 – Resonance • Data was taken across the unit using route based data collection. CP1

60

SL - Compressor -M1H Motor Outboard Horizontal Route Spectrum 02-Feb-04 15:09:54 OVERALL= 45.58 V-DG PK = 45.32 LOAD = 100.0 RPM= 1490. (24.83 Hz)

PK Velocity in mm/Sec P

50

40

30

20

10

0 0

300

600

900 Frequency in Hz

1200

1500

1800

Freq: Ordr: Spec:

24.83 1.000 45.19

• The plot above is taken from the motor showing a 1xTs peak in excess of 40mm/sec in excess of 40mm/sec.

Case Study 10 Case Study 10 – Resonance • This data is very high in amplitude. • The data was then displayed in a multi plot format to show The data was then displayed in a multi plot format to show how the amplitude was across the radial plains. • Due to the vastly different amplitudes at the 1xTs frequency the defect on this motor was Resonance. CP1

SL - Compressor - Multiple Points (02-Feb-04)

50

Max Amp 44.1

40

30

PK V elo ccity in m m /Sec

Amplitude differences between radial plains

20

10

0 M2H 15:26

M2V 15:26 0

4000


12000

16000

Case Study 10 Case Study 10 – Resonance • Recommendation • It was determined that the change in motor size may be the It was determined that the change in motor size may be the cause of the resonance as the mass had been altered. A visual inspection of the frame work also revealed that one of the support beams had cracked along the weld this altering the support beams had cracked along the weld – this altering the stiffness of the structure. The support was welded and strengthened and more data was acquired to determine if any effect on the resonance had occurred.

Case Study 10 Case Study 10 – Resonance • The spectra, shows the ‘Before’ and ‘After’ plot of the motor inboard horizontal. It shows a significant drop in amplitude of the 1xTs peak. – By stiffening the structure the natural frequency had increased moving it away from the 1xTs peak thus resulting in a significant drop in it away from the 1xTs peak thus resulting in a significant drop in amplitude. CP1

SL - Compressor -M2H Motor Inboard Horizontal

50

Max Amp 44.1

40

PK Ve lo c it y in m m /Se c

30

20

10

0 07-May-04 10:08:05

02-Feb-04 15:26:38 0

1000


3000

4000

Summary of Faults Summary of Faults

Belt Frequency

Misalignmeent

Electrical Imbalance

Resonance

Looseness Electrical

Advanced Bearing Wear

Lower Gearmesh G Severe Frequencies Misalignment Severe Looseness

c c

c

c

Early Bearing Wear Gearmesh Frequency Electrical Slot Pass Frequency

c c c

F re q u e n c y In T e rm s Of RPM

M o s t L ik e ly C a u s e s

1 x RPM

U n b a la n c e

2 x RPM

M e c h a n ic a l Loosenes s

3 x RPM

M is a lig n m e n t

Le s s th a n 1 x RPM

O il W h irl (le s s t h a n 1/ 2 R P M

S y n c h ro n o u s (A . C . L in e F re q u e n c y ) 2 x S ynch. F re q u e n c y M a n y T im i es RP M (H a rm o n ic a lly R e la t e d F re q . )

E le c t ric a l P ro b le m s T o rq u e P u ls e s B a d G e a rs A e ro d y n a m ic F o rc e s H y d ra u lic F o rc e s M e c h a n ic a l L o o s e n e s s

R e c ip ro c a t in g F o rc e s H ig h F re q u e n c y B a d A n t ii-F F ric t io n (N o t H a rm o n ic a lly B e a rin g s R e la t e d )

O t h e r P o s s ib le C a u s e s & R e m a rk s 1 ) E c c e n t ric jo u rn a ls , g e a rs o r p u lle y s 2 ) M is a lig n m e n t o r b e n t s h a ft - If h ig h a x ia l vib ra t io n 3 ) B a d B e lt s - If R P M o f b e lt 4) Res onanc e p ro c a t in g fo rc e s 5 ) R e c ip 6 ) E le c t ric a l p ro b le m s 7) Loosenes s 8 ) D is t o rt io n - s o ft fe e t o r p ip in g s t ra in 1 ) M is a lig n m e n t - if h ig h a x ia l vib ra t io n 2 ) R e c ip ro c a t in g fo rc e s 3) Res onanc e 4 ) B a d b e lt s - if 2 x R P M o f b e lt U s u a lly a c o m b in a t io n o f m is a lig n m e n t a n d e x c e s s ive a x ia l c le a ra n c e s (lo o s e n e s s ). 1 ) B a d d rive b e lt s 2 ) B a c k g ro u n d vib ra t io n 3 ) S u b -h a rm o n ic re s o n a n c e 4 ) " B e a t " V ib ra t io n C o m m o n e le c t ric a l p ro b le m s in c lu d e b ro k e n ro t o r b a rs , e c c e n t ric ro t o r u n b a la n c e d p h a s e s in p o ly -p h a s e s y s t e m s , u n e q u a l a ir g a p . R a re a s a p ro b le m u n le s s re s o n a n c e is e x c it e d G e a r t e e t h t im i es RP M of bad gear N u m b e r o f fa n b la d e s t im e s R P M N u m b e r o f im p e lle r va n e s t im e s R P M M a y o c c u r a t 2 , 3 , 4 a n d s o m e t im e s h ig h e r h a rm o n ic s if s e ve re lo o s e n e s s 1 ) B e a rin g vib ra t io n m a y b e u n s t e a d y - a m p lit u d e a n d fre q u e n c y 2 ) C a vit a t io n , re c irc u la t io n a n d flo w t u rb u le n c e c a u s e ra n d o m , h ig h fre q u e n c y vib ra t io n 3 ) Im p ro p e r lu b ric a t io n o f jo u rn a l b e a rin g s (F ric t io n e x c it e d vib ra t io n ) 4 ) R u b b in g

Useful References • • • •

Simplified Handbook of Vibration Analysis Volume 1 – Arthur R. Crawford Simplified Handbook of Vibration Analysis Volume 2 – Arthur R. Crawford BS ISO 13373‐1 2002 – Condition Monitoring and Diagnostics of Machines – General Procedures BS ISO 13373‐2 BS ISO 13373 2 – Condition Monitoring and Diagnostics of Condition Monitoring and Diagnostics of Machines – Processing, Presentation and Analysis of Vibration Data

Vibration Analysis Level- 1 Updated [Compatibility Mode]

Recommend Documents