bearing life time

BEARING TM

With MULTIPLE DISCRIMINANT ANALYSIS TM

“Everything should be as simple as possible, but no simpler.” AE

A new approach to rolling element bearing life estimation, and extension.

John E. JuddDynamic Measurement Consultants, LLCjejvibes@aol.com•US PATENT 6,763,312 B1

Introduction

The useful operating life of a rolling element bearing is influenced by a number of factors.Some of the factors are controlled by the designer, others are controlled the user. Bearing LIFEGUARDTM is a metrics based system for monitoring and optimizing key factors under user control.

Ockham’s Razor for PdM!

There is an ongoing need for simpler methods to assess machinery bearing condition.

This presentation describes a process developed in a three year effort to find and test a simple but effective approach to the condition assessment of rolling element bearings!

RELIABILITY

RELIABILITY –A DEFINITION The probability that a component part, equipment, or system will perform its intended function, under specified conditions of environment, and satisfactory maintenance, for a specified period of time. [Ref 9]

A SIMPLE LOOK ATBEARING LIFE ASSESSMENT: key factors!

• OUTSIDE DYNAMIC FACTORS ACT TO REDUCE BEARING LIFE.

• INTERNAL BEARING EMMISSIONS COMMUNICATE ACTUAL CONDITION AND ALLOW ESTIMATES OF PROBABLE REMAINING LIFE.

ARE YOU LISTENING?

LIFE REDUCING FORCES

HELP!

This is what the emissions from the bearing look like! All the information you need is there!

g units vs time.

Time>

Acceleration frequency spectrum.

Frequency >

BEARING LIFE-KEY FACTORS:

• Select proper bearing.• Install it properly.• Minimize lubricant contamination.• Control & Minimize the Forces that act

to shorten bearing life.• Monitor the actual condition of

bearing.

BEARING FAILURE

L10 Bearing life is defined as the number of cycles that 90 % of an apparently identical group of bearings will run before spalling defect reaches 0.01 inch2 (6mm2).

Timken specifies L10 for 90x106cycles.ISO 381 specifies for 1x106 cycles.

BEARING LIFEGUARDtm

PROVIDES:

A MEANS TO MEASURE/CONTROL:• LIFE REDUCING FORCES.• ACTUAL CONDITION OF BEARING.• LIFE EXPECTANCY BASED ON THESE

FACTORS.• ESTIMATED PROBABILITY OF NEAR

TERM FAILURE.

The DF METRIC – A measure of Dynamic Forces that Reduce life!

DF (DYNAMIC FORCES) RANGE 1-10

1-2=Optimum near L10 life.

2-4=Slightly High (Monitor)

4-7=Excessively High (Action)

7-10=Danger! (Shut down)

4.5

IF DF Over 4-Check for Imbalance, Misalignment or other low frequency problems!

The BD METRIC provides information on actual BEARING CONDITION

BD-BEARING DEGRADE RANGE 1-10

1-2= Optimum L10 life

2-4= Early degrade state

4-7 = Second Degrade State (Monitor)

7-10 = Final Degrade State (Replace)

1

IF BD EQUALS 10-PROBABILITY OF FAILURE IN 90 DAYS = 63%.

The LE METRIC Estimates effects on BEARING LIFE= C1DF + C2BD

LE -LIFE EXPECTANCY ESTIMATE 1-10

1-2= Optimum L10 life

2-4= 10 to 30% life reduction.

4-7 = 30 to 70% life reduction

7-10 =70 to 80% life reduction

4

IF LE 4-7, CHECK DF OR BD FOR PROBLEM!

IN ONE QUICK GLANCE:

• The tech knew that the machine bearing was fine but its expected life is dropping.

• DF indicates that dynamic forces are causing the reduction.

• The machine required further checking for imbalance, misalignment or other low frequency dynamic problem.

The tech needed only three numbers-and did not require:

• Frequency spectra, or data analysis. • A sophisticated expert analysis.• High level expertise in mechanical

engineering or signal processing.• The tech had enough actionable

information to make a decision!

NEWARK POWER PLANTCogen Plant - 10.5 MW 376,000 BTU/HR - Cascade Heate

474,000 LBS/HR - Steam 20,000 TONS - Refrigeration

2,200 KVA - Emergency Generator

Fig 3

Facility Power Plant

A sample TFM/PdM managers report:

SAMPLE

Main Campus

MAINTENANCE GAP= $2,200

SAMPLEMain Campus

AVOIDED COST = $22,000

MAIN CAMPUS LIFE ESTIMATE DISTRIBUTION

0

5

10

15

20

25

30

35

100 75 50 25 FAIL

AHUPUMPSMOTORS

L-FACTOR

MACHINES

CLICK HERE FOR MEAN TREND

ILLUSTRATION

PERCENT LIFE EXPECTANCY

144 MACHINES HAVE REDUCED BEARING LIFE!

NUMBER OF MACHINES BY LIFE FACTOR

LIFE ESTIMATE DISTRIBUTION

0

10

20

30

40

50

60

1 ALERT BAD

AHUPUMPSMOTORSBAD MOTORS

BAD

1-3 3-7 7-10L-FACTOR

MACHINES

SAMPLE

NUMBER OF MACHINES BY LIFE FACTOR

MACHINE DEGRADATION FACTOR DISTRIBUTION

Things that indicate machine is in failure state

05

101520253035404550

3 10 15

AHU-SAMPLE

100

MACHINES OF SAME TYPE

GOOD- ALERT- ACTION

D-FACTOR NUMBER

D- FACTOR NUMBER[Sample]

FACILITY LIFE ESTIMATE TREND

0

2

4

6

8

10

12

JAN FEB MAR APRIL MAY JUNE

ILLUSTRATIONMEAN LIFE FACTOR TREND[100 MACHINES]

350 HORSEPOWER GAS COMPRESSORBDF Reading on shaft idler bearing =12 -Probability of

bearing failure in 90 days 63%

DEGADE FACTOR=12Detailed acceleration spectrum taken after bearing failure alert.

Top-before bearing replacement. BD =12

Lower-after replacement. BD =2

3kHz

1.4 g

BD=12 Near Failure Bearing removed from compressor.

How is that possible?Lets take a closer look.

• What are the factors that influence bearing life?

•How many of these factors does Maintenance control?

FACTORS1) ROTATIONAL SPEED

2) RATIO OF RATED LOAD/APPLIED LOA

3) ENVIRONMENT

4)BEARING MATERIAL

5) TIME AT LOAD

6) ASSEMBLY

7) LUBRICATION

Items 2,6 & 7 ?

L10 BEARING LIFE EQUATIONManufacturers rating on new bearing.

•L 10 = (K 1* a1 * a2 * a3 ) [ fa * CE /P ]10/3 (hours)

N K1 = 16667

• L 10 is estimated life of 90% of sample test bearings under specified operating conditions.

• K1, a 1, 2, 3 and fa, are manufacturer’s constants related to material, environment, reliability %. (ie-a3 = 0.2( For 99% ) and fa= number of parallel bearings.

• CE/P = ratio of rated load to actual load.• N = rotational speed in rpm

Ref: Timken Bearing Manual

IMPORTANT POINTSto note in L 10 equation:

LIFE VS BEARING LOAD• 2 X INCREASE RPM -

DECREASE BEARING LIFE factor 2• 2X INCREASE BEARING LOAD -

DECREASE BEARING LIFE factor (C/PL)3.3!• INCREASE BOTH X 2-

DECREASE BEARING LIFE factor 20!• Drop bearing load from 50 to 40% -double bearing

life!

How Bearing LifeGuard tmLE Factor Changes with Machine Speed.

DROP IN LE Life Expectancy factor VS. SPEED

0102030405060708090

1 2 31080 1800 3600 RPM

% L

10

WB,K2

BEARING FAILURE is difficult to predict!

• Years of experience has shown that bearing failure is probabilistic and very difficult to predict accurately.

• Failure data indicates that characteristics follow a Weibull probability distribution.

• The Variance on this distribution extends from < 0.5 to >15 times the mfgs. L10 life.

• It is easy to see why failure prediction is difficult!

FAILURE CHARACTERISTICS

STUDIES BY FAA, NASA AND OTHERS HAVE CONCLUDED:

‘MOST BEARING FAILURES ARE RANDOM AND ‘SCHEDULED’PREVENTIVE MAINTENANCE ALONE IS NOT THE MOST COST EFFECTIVE MAINTENANCE STRATEGY !

CONDITIONAL PROBABILITY OF AGE RELATED FAILURES

UAL BROMBERG U.S. NAVY

4% 3% 3%

2% 1% 17%

5% 4% 3%

A

B

C

AGE RELATED FAILURES [CONT.]

UAL BROMBERG U.S. NAVY

7% 11% 6%

14% 15% 42%

68% 66% 29%

D

E

F

REF 4

AIRCRAFT COMPONENT FAILURE CHARACTERISTICS

Fe(t) VS. OPERATING TIME

4% BATHTUB CURVE

2% CONSTANT TO EXPONENTIAL

4% LINEAR INCREASE WITH TIME

89% RANDOM FAILURES

REF: FAA STUDY MSG-1 BOEING 747

[See also Ref.4, Appendix A]

FAILURE CHARACTERISTICS Veridian Engineering- Overman.

Type A-- Bathtub Curve ( 4%)

Type B-- Constant 1/mtbf then exp. Increase (2%)

Type C-- Prob. of failure linear increase w/time (5%)

Type D-- Low prob. when new then constant (7%)

Type E-- Constant 1/mtbf [ Same old or new!] (14%)

Type F-- Exp. decrease then constant 1/MTTF [68%]

High Number of Excessively High Stress Cycles Lead to Subsurface Material Fatique.

CYCLES TO FATIQUE FAILURECARBON STEEL (0.37 QUENCHED)

Ref:Shock&Vibration Handbook 3rd Ed.

0

20

40

60

80

100

120

10^4 10^5 10^6 10^7 10^8

NUMBER OF CYCLES

PER

CEN

T LI

FE E

XPEC

TAN

C

STRESS/1000psi

Cycles to Fatigue FailureCarbon Steel (quenched 0.37)

100

80

60

40

20

M

A

X

S

T

R

E

S

S

10

00

psi

10^4 10^5 10^6 10^7 10^8CYCLES(STRESS REVERSALS)

One year = 18 x 10^8 @ 3600 cpm

Two impacts per rev@ 3600 = 3.6 x10^9 /year

Ref; Shock & Vibration Handbook 3rd Ed.

L10 GENERAL FATIQUE LIFE VS LOAD

1

10

100

1000

10000

100000

1000000

MinLoad

2500 lb 5000 lb 10000lb

15000lb

20000lb

Series1

•L10 Life is reduced at higher stress levels.

•GENERAL FATIGUE LIFE VS LOAD

• MIN 2,500 5,000 10,000 15,000 20,000BEARING LOAD ( Lbs.)

•%100

10

0.1

0.01

0.001

0.0001

•FATIGUE CURVE

L10 LIFE VS. DEGREE MISALIGNMENT

0

50

100

150

200

0 D 5 D 10 D 15 D 20 D

DEGREE OF MISALIGNMENT

IDEALCROWNIS = 7.7mm

SHAFT RPM VS BEARING LIFE(ANSI STD)

1

10

100

1000

10000

100000

900 RPM 2500 RPM 5000 RPM 10000 RPM

EXAMPLE ONLY

SHAFT RPM VS BEARING LIFE(ANSI STD)

•TOTAL BEARING LOAD

•Rotor weight+

•Belt tension= xlbs

•Load zone force = vector sum of forces.

•Dynamic force= Imbalance + Misalignment + ImpactsImbalance =1/2 rotor mass x (2∏ RPS)2 *mass cg offset

Von Mises & Hertzian stress loads on Bearing surface

•Surface point contact stresses can reach 200,000 to 500,000 psi!Ref: Harris Rolling bearing

analysis

Sub surface fatique defects migrate to surface.

Timken exponential failure distribution

Ref: Timken Bearing Manual

WEIBULL EQUATIONS:Re(t) = Prob. of survival= e-(t-λ/θ-λ)k

Fe(t) = Prob. of failure = (1-Re(t) ) Fe(t) = 1- e-(t-λ/θ-λ)k dFe(t)/dt = f(t)= Rate of change of Fe(t)

f(t) = k θ -k t (k-1) e-(t-λ/θ-λ) for k = 1, λ= 0

= 1/θ e t/θ

Where ; k= shape dispersion factor, λ= location, θ=MTTF, t = time period

Timken Bearing uses λ = 0, k = 1.5 for L10

WHY BEARING FAILURE IS HARD TO PREDICT.

FAILURE DISTRIBUTIONS VS DISPERSION FACTOR

05

10152025303540

0.5 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0

L10 (@ k= 1.5) LIFE MULTIPLESL10 life (1-e -(t/MTTF))^k = 0,1% failure prob.

% U

NIT

FA

ILU

RES

K= 0.5K=1.0K=1.5K=2.0

L10 TIMKEN STANDARD

MTTF(4.81 L10)

10% FAIL

63% FAIL

RELIABILITY ESTIMATE

Rs = Probability of system functioning for time t with elements connected in series:

Rs = System Reliability = e-(t-λ/θ-λ)k

Rs = P1*P2*Pn [Ex: 0.8*0.7*0.9= 0.504]

Fs = Probability of failure = 1- Rs = 0.496

How does MDA work?

• Let’s look at how DF is derived.• Let’s look at how BD is derived.• Finally, let’s look at how probability of

failure is estimated.

SOURCES OF LIFE SHORTENING DYNAMIC FORCE USED FOR DF

•Imbalance

•Misalignment

•Eccentric Shaft

•Belt Resonance

•Other sources of low frequency motion.

•High frequency impacts & resonance.ALL OF THESE CREATE STRESS REVERSALS.

They are indicated and measured as low frequency bearing housing accelerations and impact energy, adjusted per ISO standards, for flexible or rigid mounting.

WHAT IS THE NATURE OF BEARING FAULTS ?

FATIQUE PROBABLE CAUSESpalling-subsurface fatique Excessive loadPeeling -surface fatique LubricationWEARFretting /Surface Corrosion Vibration/movementAbrasion ContaminationScoring Abrasion Check SealsCorrosions SealsBrinneling/Localized Fretting VibrationSmearing Sliding friction;lubricationPitting/Fluting Electrical DischargePLASTIC FLOWBrinnelling Excessive loadDenting Excessive Point loadMaterial Failure Hard/Cold workingFRACTURE (Catastrophic) Latent Defect

Most faults cause impacting & high frequency energy!

MDA CONVERTS THESE ELEMENTS TO METRIC FACTORS

DYNAMIC FORCES

VIBRATION DATA

IMBALANCE COUPLING GEAR MISALIGNMENT WARPED SHAFT ECCENTRICITY

BELT DEFECT BELT RESONANCE

PULLEY ALIGN PULLEY BALANCE

BLADE PASS BEARING- CAGE BEARING INNER

BEARING OUTER BALL

PITTING FRETTING SCORING

SPALLING

PROCESS

FLEX/RIGID

PROCESS

DYNAMIC FORCES FACTOR

BEARING DEGRADEFACTOR

ACTION/LIFE REDUCED ALERT

OPTIMUM

NEAR FAIL ALERT OPTIMUM

INFORMATION

BEARING CONDITION

10

10

1

1

DATA INTEGRATION

US PATENT # 6,762,312 OTHER PATENTS PENDING

HOW MULTIPLE DISCRIMINANT ANALYSIS WORKS?

PEAK CAPTURE

DFDYNAMIC FORCE

LE LIFEEXPECTENCY

BDBEARINGCONDITION

SPLITTER 3

2

1

4

LFD ADDER

ADDER

SIGNAL FFT

INPUT SIGNAL

PROCESSORANALYSIS

& SUMMING DISPLAY

BEARING LIFE FACTOR

•Adder

•DF [20%]

•BD [80%]

•LE [100%]

• LE A composite of DF & BD which indicates overall machine condition.

1= Optimum L10 Life10 = Minimum Life Expectancy.

•DF- A measure of dynamic forces on bearing [20-40% contribution- user selectable]

•BD- A measure of actual bearing condition. [80-60% contribution.]

LE is a forecast of the expected bearing life!

BD - DF & BEARING LIFE RELATIONSHIP- Cr/Ca [- 20%]

1 3 5 7 9

S1

0

2

4

6

8

10

EXPECTED LIFE

10 =100%

BDF-BEARING DEGRADATION

FACTOR

DYNAMIC FORCES

EXPECTED L10 BEARING LIFE

8.00-10.006.00-8.004.00-6.002.00-4.000.00-2.00

DF DYNAMIC

FORCE FACTOR1

•ASSUME RATIO OF RATED TO APPLIED FORCE DROPS BY 20% AS DF GOES FROM 1 TO 10

DF INCREASE ONLY

BD INCREASE ONLY

10

L10 LIFE

SAMPLE

10

DF max =50% reduction.

A new bearing can have low Life Expectancy! It may be affected by such factors as rotational speed and imbalance.

WHAT IS MDA BDBEARING CONDITION?

• MDA USES A COMBINATION OF POWERFUL BEARING ANALYSIS TECHNIQUES

• Crest Factor, Kurtosis, High Frequency energy and Envelope Demodulation, or others.

• Each analysis technique used is based on acceleration and is converted to a 1-10 metric.

• The metrics are combined to provide BD = 1-10.• BD is then related to estimated MTTF of bearing.• BD =1, MTTF ≈L10, when BD =10, MTTF = 2160

hrs = 63% probability of failure.

ILLUSTRATION OF IMPACTS CAUSED BY BEARING DEFECTS

Courtesy: DLI instruments, WA.

MACHINERY VIBRATIONTIME WAVEFORM

Courtesy: Condition Monitoring, LLC , NJ

Peak = 0.4 in/sec

RMS = 0.17 in/sec.

CF Vel= 0.4/0.17 = 2.46 BD=4.5, LE=3.9, DF=3.9 HF=7.7 CF=14 KF=.55, ED=4.27

BRG 5, OUT, rms=2.22567

BD=11, [HF=12.7, CF=14, KF=10, ED=12.1]

0 20.0m 40.0m 60.0m 80.0m 100.0m 120.0m 140.0m 160.0m-10.0

-5.0

0

5.0

10.0

se

Re

RMS: 2.2

Live X1 X: 0.0799805 Y: 0.209865

G’s

ACTUAL TIME HISTORY SHOWING EXPONENTIAL DECAY

Peak--- rms.CF = P/rms

= 1-10K =

(P-rms)^4/rms=1-10

Courtesy of JLF Analysis, Schenectady, NY

1 2 3 4 5 6 7S1

S3

S50.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

KURTOSIS FACTOR

CREST FACTOR

G(RMS)

KURTOSIS FACTOR VS CF & G(RMS) FACTOR

12.0-14.0

10.0-12.0

8.0-10.0

6.0-8.0

4.0-6.0

2.0-4.0

0.0-2.0

K is quadratic expression sensitive to both peak value and rms g value.

SAMPLE SHOWING LOW FREQUENCY ENVELOPE

Rectified Low frequency envelope.

High frequency stripped off.

ED = rms value of envelope= 1-10

Courtesy of JLF Analysis, Schenectady, NY

BRG 5, OUT, rms=2.22567

BD=11, [HF=12.7, CF=14, KF=10, ED=12.1]

-5.0

0

5.0

10.0

Re

RMS: 2.2

Live X1 X: 0.0799805 Y: 0.209865

G’s

High frequency ring down 20kHz

Bearing impact frequency

Low frequency demodenvl.

ED = Rms value of enveloped bearing impact energy!

BD CLOSELY FOLLOWS BEARING DEFECT SEVERITY

INCREASING BDF WITH INCREASING DEFECT SEVERITY

TWELVE SKF 6205

02468

7 4 11 3 1 9 10 8 12 2 5 6

WBK2- ARR

BDF-wbk2

0

2

4

6

8

10

12

14

Vol

ts

BDF-wbk2 0.64 0.75 1 1.75 1.89 3.2 5.47 5.6 6.9 10 12.2 12.95

7REF

4ABR1

113BAL

3IN

1IN/OU

9IN2

123SCR

83SCR

10OUT2

2BALL

6ABR3

5OUT

7 REFERENCE- GOOD BEARING

4 LIGHT ABRASION/ GRINDING COMPOUND

11 LIGHT SCORING ON THREE BALLS

3 LIGHT SCORING CONDITION

1 MILD SCORING ON INNER/OUTER RACE

9 HEAVY SCORING ON INNER RACE

12 MED SCORING INNER/OUTER AND BALL

8 HEAVY SCORING ON INNER RACE/BALL

10 HEAVY SCORING ON OUTER RACE

2 HEAVY SCORING ON BALLS

6 SEVERE ABRAS- HVY GRINDING COMPOUND

5 HEAVY SCORING ON OUTER RACE

BEARING TYPE- SKF 6205

RELATION OF BD AND PROBABILITY OF FAILURE RATIO OF t/θ=1.0 USING K=1.5

FAILURE PROBABILITY VS DISPERSION FACTOR(1-e ̂t / MTTF)^k

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5

RATIO t/THETA

PRO

BA

BIL

ITY

K = 1.0

K = 0.5

K=1.5

K=0.75

BDF 1-10

F(t) = (1-e –(t/θ)3/2)

Assume t/MTTF =1 when BD =10

BD =10

MTTF HAS DROPPED TO 90 DAYS (21SUBSTITUTE FOR BDF READINGS AB

PROBABILITY OF FAILURE FOR MDA-3t/MTTF (1-e -(t/mttf) 1̂.5)- MTTF

1 0.1 0.311280057 4.81*L102 0.28 1.3770865094 0.46 2.6800908716 0.64 4.0070421228 0.82 5.240971686

10 1 6.321205588 2161.18 7.224649641.36 7.952609251.54 8.520808949 PROBABILITY X

20 1.72 8.95206009622 1.9 9.271220579

PROBABILITY OF FAILURE VS. BD

0

1

2

3

4

5

6

7

1 2 4 6 8 10BD VALUE

PRO

BA

BIL

ITY

X10

t/MTTF

FAILPROB

63.2% In 2160hrs

BRG 7, REF, rms=0.223764

0 20.0m 40.0m 60.0m 80.0m 100.0m 120.0m 140.0m 160.0m-10.0

-5.0

0

5.0

10.0

se

Re

RMS: 0.2

Live X1 X: 0.0799609 Y: 0.175307

G’s

BD=0.6, HF=0.67, CF=5, KF=0.02, ED=0.16

NEW REFERENCE BEARING

B7-REF (run135) TEST 1-25-04

0 5.0K 10.0K 15.0K 20.0K0

50.0m

100.0m

150.0m

200.0m

250.0m

300.0m

350.0m

400.0m

450.0m

500.0m

Hz

Ma

S1 X: 6625 Y: 0.00297311

HFD= 0.6 CFD= 3.2

BD = 0.64HFD = 0.6CFD = 3.2KFD= 0.04EDD = 0.15

NEW REFERENCE BEARING

B4-ABR15 (137)

0 5.0K 10.0K 15.0K 20.0K0

100.0m

200.0m

300.0m

400.0m

500.0m

Hz

Ma

S1 X: 2725 Y: 0.00214685

BD = 0.75HFD = 0.87CFD = 3.5KFD= 0.02EDD = 0.22

VERY LIGHT ABRASION

B11-3BALLS (run 136)

0 5.0K 10.0K 15.0K 20.0K0

50.0m

100.0m

150.0m

200.0m

250.0m

300.0m

350.0m

400.0m

450.0m

500.0m

Hz

Ma

S1 X: 5662.5 Y: 0.00749745

BD = 1.0HFD = 1.45CFD = 4KFD= 0.03EDD = 0.37

LIGHT SCORING ON BALLS

B3-IN (138)

0 5.0K 10.0K 15.0K 20.0K0

100.0m

200.0m

300.0m

400.0m

500.0m

Hz

Ma

S1 X: 1337.5 Y: 0.027393

BD = 1.75HFD = 2.5CFD = 7KFD= .03EDD = 0.85

LIGHT/MODERATE SCORING ON BALLS

B1-IN/OUT (run 134)

0 5.0K 10.0K 15.0K 20.0K0

100.0m

200.0m

300.0m

400.0m

500.0m

Hz

Ma

S1 X: 9037.5 Y: 0.101844

BD = 1.89HFD = 3.24CFD = 6.6KFD= 0.03EDD = 0.95

MILD SCORING

B9-IN2 (run 140)

0 5.0K 10.0K 15.0K 20.0K0

100.0m

200.0m

300.0m

400.0m

500.0m

Hz

Ma

S1 X: 6000 Y: 0.172742

BD = 3.2HFD = 6.26CFD = 6KFD= 0.04EDD = 2.5

HEAVY SCORING ON INNER RACE

B12-3SCORE2 (run 144)

0 5.0K 10.0K 15.0K 20.0K0

100.0m

200.0m

300.0m

400.0m

500.0m

Hz

Ma

S1 X: 5987.5 Y: 0.315009

BD = 5.45HFD = 8.59CFD = 14KFD= 0.6EDD = 4.2

MED SCORING ON INNER/OUTER RACE

AND BALLS.

B10-OUT2 (143)

0 5.0K 10.0K 15.0K 20.0K0

100.0m

200.0m

300.0m

400.0m

500.0m

Hz

Ma

S1 X: 5875 Y: 0.264455

BD = 6.9HFD = 5.5CFD = 14KFD= 1.5EDD = 10

HEAVY SCORING ON OUTER RACE

B8-3SCORE (run 145)

0 5.0K 10.0K 15.0K 20.0K0

100.0m

200.0m

300.0m

400.0m

500.0m

Hz

Ma

S1 X: 6387.5 Y: 0.307514

BD = 5.6HFD = 10.5CFD = 14KFD= 0.24EDD = 3.36

HEAVY SCORING ON INNER RACE

And BALL

BRG 12, 3SCR2, RMS = 1.40289

0 20.0m 40.0m 60.0m 80.0m 100.0m 120.0m 140.0m 160.0m-10.0

-5.0

0

5.0

10.0

se

Re

RMS: 1.4

Live X1 X: 0.0799805 Y: 1.11501

G’s

BD=6.3, HF=10, CF=14, KF=1.5, ED=5.2

B2-BALL (run 142)

0 5.0K 10.0K 15.0K 20.0K0

100.0m

200.0m

300.0m

400.0m

500.0m

Hz

Ma

S1 X: 5925 Y: 0.276644

BD = 10HFD = 9.2CFD = 9.2KFD= 14EDD = 7.5

HEAVY SCORING ON BALLS

B5-OUT (run 139)

0 5.0K 10.0K 15.0K 20.0K0

100.0m

200.0m

300.0m

400.0m

500.0m

Hz

Ma

S1 X: 6637.5 Y: 0.156745

BD = 12.95HFD = 13.3CFD = 14KFD= 14.2EDD = 12.95 HEAVY SCORING ON OUTER RACE

AccelerationBRG 5, OUT, rms=2.22567

BD=11, [HF=12.7, CF=14, KF=10, ED=12.1]

0 20.0m 40.0m 60.0m 80.0m 100.0m 120.0m 140.0m 160.0m-10.0

-5.0

0

5.0

10.0

se

Re

RMS: 2.2

Live X1 X: 0.0799805 Y: 0.209865

G’s

B6-ABR30 (run 146)

0 5.0K 10.0K 15.0K 20.0K0

100.0m

200.0m

300.0m

400.0m

500.0m

Hz

Ma

S1 X: 187.5 Y: 0.671253

BD=12.1, HF= 14,

CF=14, KF=14, ED=7.7

HEAVY GENERAL ABRASION

BRG 6, ABR30, rms=2.21398

0 20.0m 40.0m 60.0m 80.0m 100.0m 120.0m 140.0m 160.0m-10.0

-5.0

0

5.0

10.0

se

Re

RMS: 2.2

Live X1 X: 0.0799805 Y: 0.415329

G’S

BD=12.1, HF=14, CF=14, KF=14, ED=7.7

FAILURE PROBABILITY SAMPLE CALCULATION

• Forecast period = t = one year = 8760 hours.

• L10 = 3000 at 500 rpm, MTTF = 14430 hrs.

• R(T) = Probability of survival = exp -(t/mttf)3/2

• F(T) = Probability of new bearing failure

• = 1- R(T) = [1- exp -(8760/14,400) 3/2] ≈ 38%

• If BD indicates MTTF drops to 8760 t/θ = 1

• Probability of failure in one year. F(T)= 63%

CONCLUSION- A new way to look at bearing monitoring and fault analysis.

Bearing Lifeguard TM provides three simple metrics for maintenance technicians.

The three metrics provide information on forces acting to reduce bearing life, actual bearing condition and estimated remaining life.

Using these factors the system provides an estimated probability of failure within the next 90 days.

The system also makes available acceleration signals and demodulated envelope signals for detailed analysis if required.

BEARING TM

US PATENT #6,763,312 B1

Information in this presentation is provided for illustration of LIFEGUARD TECHNOLOGY & MDA principles only. Use for other purposes without express permission of DMC, LLC is strictly prohibited.

Reference material used in this presentation.• Shock & Vibration Handbook, Cycil Harris, 3rd Edition• Rolling Element Bearings-Tedric Harris, 3rd Edition• RCM, Condition Monitoring or both? Richard Overman,

Veridian Engineering, Maintenance Technology, Jan. 02.• NASA-Reliability Centered Maint. & Commissioning.

[Appendix A], Feb. 2002• The McGraw-Hill Dictionary of Scientific & Technical

Terms-5th Edition.• Mil Handbook 217E• SAE JA 1011 Surface Vehicle/Aerospace Std.-Evaluation

Criteria for Reliability Centered Maintenance.• Vibra-Metrics Inc. Vibration Reference Guide.