You are on page 1of 42

An Introduction

to
Vibration Analysis
Theory and
Practice
An overview of…
Various Maintenance Methods
¾ Breakdown
¾ Preventive
¾ Predictive
¾ Reliability centered (Proactive)

Vibration analysis
¾ What is machine vibration
¾ Measuring and analyzing vibration
¾ The BIG 4
Why do machines stop running?

Component failures:
¾ Failed bearings
¾ Broken fan blades
¾ Seized couplings
¾ And the list goes on…
Why Question Existing
Maintenance Practices?
Minimize failures:
¾ Balance and align machines
¾ Improved maintenance practices
¾ Clean lubricants

Reduce the impact:


¾ Avoid unscheduled repairs
¾ Stop “secondary damage”
¾ Save $$$$$
Breakdown Maintenance
Just let it fail
¾ Also known as “run to failure”
¾ Remains common practice in many places
¾ Budgeted and accepted cost of operation

Disadvantages:
¾ Secondary costs of failure (10X$)
¾ High downtime
¾ Large spare parts inventory
¾ Worker safety issues
Preventive Maintenance

Fix it before it breaks


¾ Also known as “planned maintenance”,
¾ Historical maintenance
¾ Periodic Maintenance
¾ Calendar-based maintenance
¾ Most common maintenance practice today
¾ Assumes that all machines will fail in time
¾ You perform maintenance before it fails
¾ But WHEN will it fail?
¾ WHY will it fail?
Preventive Maintenance
Disadvantages
¾ Machines fail before planned shutdown
¾ Perfectly good machines are “over-repaired”
¾ Overhauls often introduce problems due to defects
¾ Unnecessary costly downtime
¾Excessive spare parts Inventory
Predictive Maintenance
“If it ain’t broke, don’t fix it!”
¾ Also known as “condition based maintenance”
¾ Predict when a machine will fail
¾ Repair it when most convenient
¾ Repair/replace ONLY the components that are required
¾ Intelligently assess the “risk”
Predictive Maintenance

¾ Perform “condition monitoring”


¾ Determine health status
¾ Predict failure mode
¾ Act accordingly

Advantages:
¾ No surprise downtime
¾ No unexpected failures
¾ No secondary damage
¾ All maintenance is planned
¾ Sounds great!
Proactive Maintenance
“Fix it once, fix it right!”
¾ Also known as “reliability centered maintenance” and
“precision maintenance”
¾ Change machine design, purchasing and maintenance procedures to
reduce failures and increase machine reliability
¾ Precision balancing, laser alignment, etc.
What technology is available
Condition monitoring:
¾ Vibration analysis
¾ Oil analysis
¾ Wear particle analysis
¾ Thermography
¾ Ultrasound
¾ Steam Trap
Which Technology to use?
Infrared Thermography

15%

Ultrasonic 10%

5% 45%
Motor Current
Vibration Analysis
10%

15% %

Steam Trap

Oil

Percent of total PAYBACK when adopting a predictive maintenance


program plant wide
Vibration Analysis

¾ All machines vibrate

¾ The vibration ‘signature’ changes


as the condition changes.

¾ What you can hear is only part of


the story.

¾ Vibration analysis can help you


detect a wide variety of fault
conditions.
What Causes Vibration?

As the shaft turns, there are


frictional and rotational forces.

That vibration created by those


forces is transferred via the bearings
to the machine housing.
The BASIC Vibration Signal

¾ The fan rotates five times every


second.

¾ Add weight which creates an


unbalance force.

One second of time


The ‘Frequency?’
¾ Hertz = Hz = Cycles per second
RPM = Revolutions per minute
CPM = Cycles per minute
CPM = RPM = Hz x 60

¾ Fan speed = 5 Hz or 300 RPM

Period = 1/Frequency
Fan speed = 5 Hz or 300 RPM
Increase the Frequency
¾ The fan is now going twice as
fast.

¾ Cycles of the waveform are


closer together.

¾ Fan speed = 10 Hz or
600 RPM
The ‘Amplitude’

¾ The height of the wave is the


“amplitude”.

¾ Because of the weight on one


blade, the vibration level
increases as the fan speeds up.

Displacement: mils or microns


Velocity: in/sec or mm/sec
Acceleration: g
Add a second source of vibration

¾ The rub introduces a


new source of vibration.

Rub Imbalance

New vibration = 10 x 8 = 80 Hz
8 blades x 10 revolutions/second
The ‘Spectrum’
Spectrum - Examples
5 Hz = 300 RPM
5 Hz

FFT
10 Hz = 600 RPM
10 Hz

FFT
A More Complex Spectrum

Rub Imbalance

10 Hz

FFT
80 Hz

600 CPM 4,800 CPM


Resolution
¾ Data collector samples
High resolution
the electrical signal from the
sensor.

¾ The sampling rate,


number of samples, and the
Low resolution
length of the time record
determine “resolution” and
“Fmax”.
Forcing Frequencies
Special calculations are used to indicate where to
look in the spectrum – called “forcing frequencies”
Vibration tells an interesting story
Imbalance

¾ Detailed vibration Misalignment


analysis can reveal far
more information.

¾ We can detect different


fault conditions, and
assess the severity.

Bearing faults
Looseness
The Big 4

¾ Imbalance

¾ Misalignment

¾ Looseness

¾ Faulty Bearings
Imbalance
What causes “Imbalance”?
¾ A heavy spot along the shaft
¾ Causes high vibration and premature bearing failure
¾ Your vbSeries data collector can correct imbalance
Misalignment
What is “Misalignment”?
¾ Definition: “The shaft center-lines are not collinear”
¾ Can be detected in vibration signature
¾ Corrected with dial indicators and lasers
¾ Also cause of high vibration, and thus bearing damage
Looseness

¾ Rotating looseness
- excessive clearance
between rotation &
stationary parts

¾ Non Rotating looseness


- between two normally
stationary parts. ie
between foot & foundation
Bearing Faults
¾ Monitor the vibration at the bearings
¾ Amplitude levels indicate severity of the problem
¾ Frequency patterns indicate nature of the problem
¾ Many different ways to analyze the data
Demodulation

¾ Ball/roller strikes defect and Demodulated spectrum


creates a “shock wave”.

6204 BPFO
Aux Comp C7 - Mntg Base #1 - Vertical - Acc Time 800 ms
17/08/1999 14:02:15
O/All 2.235 m/s/s rms

1.6

¾ Bearing then “rings like a bell” or 1.4

1.2

resonates. 1

m/s/s rms
0.8

6204 BPFO
0.6

6204 BPFO

6204 BPFO

6204 BPFO
0.4

0.2

0 5,000 10,000 15,000 20,000 25,000 30,000


CPM
BASIC Vibration Analysis
How do we Monitor Vibration? February 25

¾ In practice, we watch how the


patterns and levels change over time.

¾ We relate the changes to what we March 17


know about the machine.

April 26
Where do I mount the Sensor?

¾ The sensor converts


the vibration into an
electronic signal.

¾ The most common


sensor is an
accelerometer.

¾ The sensor is
commonly attached
using a magnet.
Mounting the sensor

¾ Proper mounting
is very important.

¾ “Repeatability” is
essential.

¾ Good “mechanical
transmission path”
from the bearing.
Repeatability

¾ Vibration changes
when the speed and
load change.

¾ The machine must


operate in the same
state during every test.

¾ Check the speed


and load with each
measurement.
Repeatability
¾ Tests are typically
performed every 30 days.
¾ Test a machine at
2 or 3 bearings.
¾ Collect vertical, horizontal
and axial data.
C annery Motor - OE (left) - Horizontal - Acc Freq 120000 C P M
18/02/2000 8:43:18
O/A ll 3.169 m /s /s rm s

1.2

0.8
m/s/s rms

0.6

0.4

0.2

0 20,000 40,000 60,000 80,000 100,000 120,0


CPM
Look for patterns and changes
Cannery Motor - DE (right) - Vertical - Acc Time 100 ms
6/03/2000 12:27:05
O/All 5.982 m/s/s rms

15

10

-5

-10
m/s/s rms

-15

0 10 20 30 40 50 60 70 80 90 100
ms
6/03/2000 12:27:05 O/All 5.982 m/s/s rms <set RPM> <add note>

¾ The vibration pattern is important.

¾ How the pattern changes


is equally important.

Cannery Motor - DE (right) - Vertical - Dis Time 200 ms


6/03/2000 12:26:29
O/All 0.008 mm rms

0.01

0.005

peak-peak
-0.005

mm peak-
-0.01

-0.015

0 20 40 60 80 100 120 140 160 180 200


ms
6/03/2000 12:26:29 O/All 0.008 mm rms <set RPM> <add note>
Trend vibration levels
An Overall RMS trend can provide useful information

Aux Comp C7 - Mntg Base #1 - Vertical - Overall Trend


Velocity
4.5

3.5

2.5
mm/s rms

1.5

0.5

0
12/06/1998 1:28:20 10/07/1998 1:32:32 14/08/1998 1:23:53 11/09/1998 1:25:20

Velocity
What does it mean…..?

¾ How do you know when to take


action?

¾ Standards are available.

®
¾ ASCENT® removes the
guesswork.
Reality Check!

Predictive maintenance:
¾ Monitoring machines regularly with
repeatable results requires discipline
¾ Not all machines can be monitored
¾ Some machines cannot be monitored
frequently enough
¾ Technologies are not perfect
¾ Recommendations are not always followed
¾ Some machines will still fail until analysis
experience grows
Normal Fatigue
Infant
Service Life & Failure
Probability of Failure Mortality

Time

The result is an increase in machine


availability with a decrease in total costs

You might also like