Predictive Maintenance and

Machinery Condition Monitoring

Predictive Testing and Inspection
in Plant Condition Assessment
 Condition Monitoring

Goal of condition monitoring (CM) is proactive,

predictive maintenance implemented in order to
provide early warning of an impending asset
failure.

CM systems use analytical engines to

➢ Evaluate an asset’s condition,
➢ Isolate and analyze the potential for failure,
➢ Identify the failure mode, and
➢ Send alerts to maintenance when threshold
tolerances are exceeded.
 Potential Savings

Condition Based Maintenance

Increased Machine Availability

Savings:
- Reduced maintenance expenditure
- Increased production uptime
- Increased product quality
- Increased operator safety
- Decreased spare parts inventory
Vibration monitoring is the most commonly used method used in machinery inspection
 ISO for Condition Monitoring and Diagnostics of Machines

ISO 13373 Vibration condition monitoring

!SO 13374 Data processing, communication and presentation

ISO 13379 General guidelines on data interpretation and diagnostic

techniques
ISO 13380 General guidelines on using performance parameters

ISO 13381 Prognostics

ISO 14830 Tribology-based monitoring and diagnostics

ISO 18434 Thermography

ISO 22096 Acoustic Emission

ISO 18436 Requirements for training and certification of personnel

 Potential Failure (P-F Curve)

PdM Predictive Maintenance

CBM Condition Base Maintenance
CM Condition Monitoring
 What to Measure

• Gives very early warning

Vibration • Most faults, Unbalance, Misalignment, Bearing faults, etc..

• Inexpensive transducers
Temperature • Thrust bearing problems.

• Reciprocating machinery
Oil Analysis • Combined with Vibration

• Process monitoring
Fluid Pressure • No early warning
ISO 17359
ISO 17359 : Faults Matched to Techniques
ISO 17359 Example: Faults Matched to Techniques
 Defects Analysis Matrix

➢ A refinement of procedures recommended in ISO 17359

➢ Assign points to sensitivity of measurement techniques

to faults.

➢ Sensitivity 0, 1, 2

➢ Add up points and select techniques with highest points

Example 1
Example 2
 Five Common Techniques

Vibration
Thermography

Oil Analysis
& Tribology

High
Frequency
&
Ultrasonic

Motor
Current
 Vibration
Measurement for force response…
 Vibration Monitoring

Periodic assessment
for changes
 Human Analogy
Vibration analysis in some ways are similar to an ECG for humans.
Tribology based Monitoring
Wear and oil analysis
 Wear Particles
Rotating and contact usually results in wear
Example: Gear Wear
Locomotive engine indicate a
severely scored liner and piston
wear
 Bearing Failure in Journal Bearing

GEF9 Gas Turbine Journal Bearing

 Oil & Wear Analysis
Qualitative and Quantitative Ferrography
 Human Analogy
Oil and wear debris is similar to urine and blood test in humans.
Metallurgy Identification
Size
 Thermography

Thermal imaging and

temperature readings
for fault detection
 Thermography

For electrical faults, and other static faults

Thermal Imaging
 Applications

Mechanical
• Pump bearing in need
of maintenance
Thermograph of a Misalignment
 Applications

Refractory
Crack-like defect
in riser refractory
lining
 Applications
Low Voltage Electricity
• Hot neutral wire in fuse
panel
Motor Current Analysis
 Motor Current Analysis

This is similar to vibration analysis, except that it is

based on electrical current.
 High Frequency Detection

Faults emit HF signals which are picked up by

acoustic emission techniques.
Acoustics Emission: Compressor valves monitoring
Typical Valves Failures
 Machine Condition Monitoring Program

Monitoring Technique Equipment Monitored Short Term Benefits Long Term Benefits

Identifies deteriorating or defective Identifies improper alignment or

Rotating machinery: pumps,
bearings and gears looseness, balancing procedures, improper bearing
Vibration Analysis compressors, turbines, gear-
rubs, misalignment, imprecise or gear installation practices, excessive
boxes diesel/gasoline engines
balancing operating conditions

Rotating machinery with Identifies improper bearing installation or

Spectral Energy Identifies deteriorating or defective
antifriction bearings, motors, repair practices, improper oil handling or
Emission/Shock Pulse bearings, excessive wear, lack of
large pumps, turbines, dispensing practices, improper machine
Method lubrication
compressors operation

Identifies worn or damaged Identifies improper bearing or gear

Motors and generators gear
Motor Current Signature gear/bearing seal packing, wear, installation practices, improper
boxes, pumps, air conditioning
Analysis misalignment, imprecise seal/packing replacement practices,
equipment blowers
balancing poor alignment or balancing practices
Identifies improper seal/packing
Steam systems plant air or Identifies compressed air vacuum
installation practices, poor quality
Ultrasonic Analysis pneumatic systems, hydraulic or steam seal leaks, hydraulic
replacement parts, poor equipment
systems, vacuum systems leaks, bearing wear
operation or design

Roofs, boilers, furnaces, steam Identifies roof and insulation leaks, Identifies poor electrical, roofing steam
systems, electrical switching, poor electrical contacts or system or insulation repair procedures
Thermographic Analysis
motor control centers, high connections, leaking steam trips, excessive operating conditions, poor
voltage lines refractory deterioration equipment design
 Machine Condition Monitoring Program (cont.)

Identifies improper motor installation or

Induction of synchronous Identifies turn-to-turn and phase-to-
Electrical Surge repair, excessive operating conditions,
motors, DC amartures phase shorting, shorts circuits,
Comparison design differences between motor
synchronous field poles reversed or open coils
manufacturers

Lubrication/hydraulic systems, Identifies excessive wear of Identifies improper fluid or lubricant

compressor, turbines, large lubricated parts, deterioration of selection, improper oil handling or
Oil Analysis
pumps, diesel/gasoline fluid or lubricant condition, the dispensing practices, excessive
engines presence of contamination operating conditions

Lubrication/hydraulic systems, Identifies improper bearing or gear

Identifies excessive wear of
compressor, turbines, large installation practices, improper oil
Wear Particle Analysis lubricated parts, the presence of
pumps, diesel/gasoline handling and dispensing practices,
contamination
engines excessive operating conditions

Lubrication/hydraulic systems, Identifies improper fluid or lubricant

compressor, turbines, large Identifies excessive gas, liquid or handling or filtration practices, improper
Contamination Analysis
pumps, diesel/gasoline particle contamination installation of seals or packing, poor
engines filter system design

Heat exchanges, pumps, Identifies losses in pumping Identifies improper maintenance or

motors, compressors, efficiency, flow rates and heat repair practices, poor machine design or
Performance Trending
refrigeration equipment, transfer due to deteriorating installation, excessive equipment
diesel/gasoline engines components loading or operating procedures

Identifies loose or visibly worn

Identifies improper maintenance or
Visual Observations Virtually any machine shown parts, oil leaks, excessive
repair practices, poor quality parts,
Listening, Touching above noise/vibration, hot
excessive operating procedures
bearings/couplings
Why Vibration Analysis is Useful for
Machinery Health Monitoring?
 Where are vibrations coming from?

Vibrations

Internal
Forces

Structural
Vibration =Force Mobility Mobility
 Different Faults

Vibration
amplitude

Frequency
 Create a reference for each measurement point

first measurement
= reference

alert
times
10

danger
times
2.5

limit
alert
limit
Trending
 Making analysis gives earlier warning

L
Overall Level

f
 Overall Levels from RMS Summation of
Discrete Components

L
Overall Level
Overall Level
……read as a
single number
For example
8.9 mm/s, or
f 250 mils, etc.
 The Problem with Overall Levels
Without Fault 11.0 mm/s

(9.525 MM/SEC)
With Fault 11.4 mm/s
1780 RPM
(FMOT )2 + (2FMOT )2 + (3FMOT )2 + (FBRG )2 + (FGEAR )2
FMOTOR
OA =

AT
BASELINE OA = (9.525) + (3.810) + (1.270) + (0.762) + (3.810) = 11.049 MM/SEC
2 2 2 2 2

8.89 4 WEEKS OA = (9.525)2 + (3.810)2 + (1.270)2 + (3.048)2 + (3.810)2 = 11.43 MM/SEC

LATER

FREQ. 59.33 Hz (3.81

2FMOT = 2 x MOTOR

FGEAR = 40T X FINPUT

7.62 APPROXIMATE FORMULA

1,186.67 Hz (3.81
DETERIORATING
3FMOT = 3 X MOTOR FREQ.
6.35
MM/SEC)

MM/SEC)
PEAK VELOCITY

BEARING
5.08 89 Hz (1.27 MM/SEC)
(MM/SEC)

3.81
FBRG (4 WEEKS
LATER)
2.54
145.33 Hz (3.048
MM/SEC)
1.27 FBRG (AT BASELINE)
145.33 Hz (0.762 MM/SEC)

FREQUENCY (Hz)
Vibration as a detection tool of impending failure

CASE STUDY
 Case Study: Gas Turbine Example

The following presentation shows how

vibration monitoring and analysis can be used
to detect changes in vibration forcing
mechanism in a gas turbine, with changes in
the vibration signatures.
 Gas Turbine Rotor
Stages of Compressor Blades
Turbine Section

Combustion
Chamber

Compressor Section

Turbine Blades
 Case Study 1: Plant Information

• 4 unit gas turbines used in power

generation. 2 other steam turbines used
in combined cycle mode.
• Operating speed 3000 rpm (50Hz).
• Peaking plant.
 Some background…

• OEM wrote to the Power Plant that GT6 is to be

shut down immediately since it had missed a
major inspection.
• Because of this, the Insurance company withdrew
insurance cover.
• Plant decided that unit has to run as power
dispatch was critical, and outage not scheduled.
• Head I win, tail you lose.
• UTM was requested to assist in vibration
assessment, and help plant manage the risk of
operations beyond maintenance schedules.
 Blade Passing Frequencies for GTs

Compressor Blades Turbine Blades

Rotor Blades Stator Blades Rotating Stationary
Stage No. of Blades BPF (Hz) No. of Blades BPF (Hz) No. of Blades BPF (Hz) No. of Blades BPF (Hz)
1 25 1250 26 1300 63 3150 40 2000
2 31 1550 32 1600 63 3150 60 3000
3 31 1550 32 1600 63 3150 60 3000
4 39 1950 40 2000 63 3150 60 3000
5 45 2250 40 2000 63 3150 50 2500
6 39 1950 40 2000
7 39 1950 40 2000
8 49 2450 40 2000
9 51 2550 54 2000
10 49 2450 50 2700
11 61 3050 64 2500
12 61 3050 64 3200
13 65 3250 64 3200
14 61 3050 78 3900
15 71 3850 78 3900
16 71 3850 78 3900
17 71 3850 78 3900
18 71 3850 78 3900
19 71 3850 78 3900
20 71 3850 78 3900
21 71 3850 78 3900
 0.16

Row 11,12,14 - 3050Hz

0.14

Turbine Blades - 3150Hz

0.12
Row 10 - 2450Hz

0.1
Row 13 - 3250Hz
mm/s

0.08
Row 9 - 2550Hz

0.06 Row 15 to 21 - 3850Hz

0.04

0.02

0
2000 2250 2500 2750 3000 3250 3500 3750 4000
Frequency (Hz)
 Comparison of Spectrum for all GTs.

1.80

1.60

1.40

1.20
Amplitute, Gs

GT3
1.00 GT4
0.80 GT5
GT6
0.60

0.40

0.20

0.00
1150
1300
1500
1650
1950
2000
2300
2500
2600
3100
3200
3100
3200
3800
3800
3950
4550
4700
4900
Frequency, Hz
Trending of Blade Pass Frequencies
 A Happy Ending ?….

• Unit was operated for another 2 months

(instead of originally planned 2 weeks
extension).
• Major inspection undertaken with no
signs of damaged blades.
• Plant reported savings of US$ 8 million
on extended operations.

(Ng Boon Hee, et.al 2000)

 A Year Later with another Unit….

• 1 months after overhaul on GT4,

peculiarities were observed on unit.
• Audible noise from compressor section
during unit shut down. Not audible
during operations. More intense by the
day.
• Coast down time reduced significantly
from 2 hours to 20 minutes.
 Chronology of Events

(1st set of data)

8 Dec 2 Feb
4 May
28 May (Last set of data)
No major
damage on
blades
Before Overhaul
Overhaul
Post Overhaul

Dec Jan Feb Mar Apr May Jun July Aug Sep Oct Nov
Emergency Outage

2000 2001 GT4 seized Severe

24 – 26 July rubbing
Operation Resume
1 set of Vibration 4 Sep
data was collected
 3 Months and 1 Day Before Failure: Bearing Vertical

3 months before failure 1 day before failure

 3 Months and 1 Day Before Failure: Bearing Casing Axial

3 months before failure 1 day before failure

 Trending of Blade Pass Frequencies

Bearing Casing Vertical

1.6
1.4 (Row 11,12,14)
1.2
Amplitute, Gs

Freq 3050
1 Freq 3850
Freq 3900
0.8
(Stator 15-20)
0.6
0.4
0.2
(Row 15-21)
0

Date

BPF 3050 Hz (Stage 11, 12, 14) = Mean + 13x Std Deviation
BPF 3250 Hz (Stage 13) = Mean + 28x Std Deviation
BPF 3850 Hz (Stage 15-21) = Mean + 7x Std Deviation
BPF 3900 Hz (Stator 15-20) = Mean + 7x Std Deviation
 Trending of Blade Pass Frequencies

Bearing Casing Axial

3.5
3
2.5
Amplitute, Gs

Freq 3050

2 Freq 3850
Freq 3900

1.5
1
0.5
0
1/5/2001

2/2/2001

3/2/2001

6/8/2001

7/6/2001
12/8/2000

12/22/2000

1/19/2001

2/16/2001

3/16/2001

3/30/2001

4/13/2001

4/27/2001

5/11/2001

5/25/2001

6/22/2001

7/20/2001
Date

BPF 3050 Hz (Stage 11, 12, 14) = Mean + 13x Std Deviation
BPF 3250 Hz (Stage 13) = Mean + 28x Std Deviation
BPF 3850 Hz (Stage 15-21) = Mean + 7x Std Deviation
BPF 3900 Hz (Stator 15-20) = Mean + 7x Std Deviation
 Facts available…

Extremely high BPF of rotor stage 11 to 21

➢ clearance reduced.
Coast down time from 2 hours to 20 minutes
➢ friction built up.
Rubbing noise during coast down
➢rubbing of components.

Recommendations for outage but not heeded.

Ummh… ?
 What happen next….

• Unit operated the following day, and

then…
• Rotor seized after coast down… cannot
barr. Rotor bow. Upon cooling, rotor
relieved itself the following day.
• Rotor run up again for load dispatch….
• Seized up again after coast down.
 Visual Inspection

• Compressor blades left traces of rub

marks at tips, particularly at stage 20
and 21.
• Further inspections confirmed more
severe and underlying rubs by labyrinth
glands (~ 3 mm in depth).
• Rotor eccentricity was thought to be the
root cause, which could be attributed to
several reasons.
 Findings

• Suggestion by OEM that root cause could be

due to compressor casing distortion (uneven
insulation).
• Root cause not conclusively identified,
although unit back in service after re-
assembly and remedial works to insulation.
• Plant overall vibration alarm limits not a
good indicator of impending rubs and shaft
seizure.
 MORE CASE STUDY …

• Refinery
– Coke Structure Vibration Investigation

72
 Coke Structure Vibration Investigation

• Coke Structure
73
 Unit 32

2 2
2 2
Measureme
nt Setup 2

2 2

2 2

X -sea 2
Y-Lift 2
2
2
Legend: 2 2

No. of Accelerometers
Measureme
2 2
2
nt Setup 1
2 1
Level 4 (6th floor) 2

2 2
Level 3 (5th floor)/ 8th Floor
2
Level 2 (4th floor)/ 7th Floor

Level 1 (3th floor)

 Drilling @ Drum 2-Measurement Setup 2

X direction Y direction
Level 1

7th Floor

8th Floor

Level 4
Long Term Full Cycles Data

OPERATIONAL DEFLECTION SHAPES

Observations and Findings

FINDINGS
 Summary of Results

• Vibration amplitude average on structure in

order of 0.8mm/s. Highest 5.4 mm/s (Overall in
frequency range 0-250Hz).
• Highest spectral frequencies at 2.54 mm/s (1.87
Hz), 3.67mm/s (2.25 Hz), 2.66mm/s (2.44 Hz),
1.91mm/s (2.62Hz), 0.7 mm/s (4.3 Hz).
• Within limits for cosmetic damage based on ISO
4866* and BS 7385. Limit 15 mm/s (0.2 to 30
Hz).

*Mechanical vibration and shock — Vibration of fixed structures — Guidelines for the
measurement of vibrations and evaluation of their effects on structures
 Observations 1

• During Drum 2 operations (Drum 1

online feeding) be it during Drilling and
Normal stages of the full cycle, coke
structure and drum 2 body consistently
had higher vibration response.
• This implied higher forcing excitation
from Drum 2; which may be due to lose
hold down bolts (issues at drum ring) or
drilling tower of Drum 2.
 Observations 2

• Structural support members for Lift at

Level 2 and Level 3 had higher vibration
levels as compared to main structure.
• This is consistent with site reports of
more violent motion at Level 2 and
Level 3.
• Structural members may be relatively
more flexible (corresponding to a
natural mode?).
 Recommendations

• To inspect integrity of all hold down

bolts, drum ring; and check securing of
fixtures and connections for Drum 2
(including pipes to Drum 2).
• To inspect connections / fasteners for
structural members which reported
vibrations higher than normal (as
compared to main structure). Refer to
list attached.
• To undertake re-tightening of all bolts
for entire structure.
83