International Conference on Engineering Vibration … of Fatigue Cycle Identification Methods •Simple range count vs. range pair count vs. rainflow count •ASTM E1049 − 85 (2011)

International Conference onEngineering Vibration

Comparison of Fatigue Cycle Identification Methods

Ljubljana, Slovenia

7-10 September 2015

Curtis E. Larsen, Ph.D., P.E.

NASA Technical Fellow for Loads and Dynamics

[email protected]

Tom Irvine

Dynamic Concepts, Inc.,

Industry Member, NESC Loads and Dynamics Technical Discipline Team

[email protected]

https://ntrs.nasa.gov/search.jsp?R=20160006936 2018-05-10T16:59:14+00:00Z

Comparison of Fatigue Cycle Identification Methods • Simple range count vs. range pair count vs. rainflow count

• ASTM E1049 − 85 (2011) “Standard Practices for Cycle Counting in Fatigue Analysis

• “The rain flow cycle counting method allows satisfactory predictions of the effects of different block sizes, different sequences of applying the same strain peaks, and superimposed loadings.”

• “The range pair counting method is nearly identical to the rain flow method, but the use of any of the other well known cycle counting methods, such as peak counting, level crossing counting, or range counting, can result in large differences between predicted and actual fatigue lives.”

• “The use of any method of cycle counting other than the range pair or rain flow methods can result in inconsistencies and gross differences between predicted and actual fatigue lives.”• Dowling, N. E., “Fatigue Failure Predictions for Complicated Stress-Strain Histories”, Department of Theoretical

and Applied Mechanics, University of Illinois, Urbana, Illinois, T. & A. M. REPORT NO. 337, January 1971. (Sponsored by Naval Air Development Center, Warminster, Pennsylvania 18974, Contract No. N00156-70-C-1256)

Comparison of Fatigue Cycle Identification Methods • Practical range of fatigue exponent: m = 3.324 and m = 7.3

• Power Spectral Densities Studied• Single Block

• Unimodal

• Two-Block

• Bimodal

Damage Rate ComparisonSingle Block PSDs

1.00

10.00

1.00 10.00

Damage Rate vs. RainflowSingle Block PSDs

m = 3.324

Simple Range Count

Range Pair Count

Rainflow Count

30

300

30.00 300.00

Damage Rate vs RainflowSingle Block PSDs

m = 7.3

Simple Range Count

Range Pair Count

Rainflow Count

Damage Rate ComparisonUnimodal PSDs

3.00

3.00

Damage Rate vs. RainflowUnimodal PSDs

m = 3.324

Simple Range Count

Range Pair Count

Rainflow Count

100

100.00


m = 7.3

Simple Range Count

Range Pair Count

Rainflow Count


3

3.00


m = 7.3

Simple Range Count

Range Pair Count

Rainflow Count

0.10

0.10


m = 3.324

Simple Range Count

Range Pair Count

Rainflow count

Damage Rate ComparisonTwo-Block PSDs

0.1

1

10

100

0.10 1.00 10.00 100.00

Damage Rate vs. RainflowTwo-block PSDs

m = 3.324

Simple Range Count

Range Pair Count

Rainflow Count

1

10

100

1000

1.00 10.00 100.00 1000.00

Damage Rate vs. RainflowTwo-Block PSDs

m = 7.3

Simple Range Count

Range Pair Count

Rainflow Count

Damage Rate ComparisonBimodal PSDs

0.5

5

0.50 5.00

Damage Rate vs. RainflowBimodal PSDs

m = 3.324

Simple Range Count

Range Pair Count

Rainflow Count

10

100

10.00 100.00

Damage Rate vs. RainflowBimodal PSDs

m = 7.3

Simpe Range Count

Range Pair Count

Rainflow Count

Conclusion

• Reiterate conclusions of Dowling and others• Rainflow cycle identification method preferred (conservative)

• Range-pair method nearly equivalent

• Simple range counting can be very unconservative• Small cycles can do “negative damage” as superimposed on large cycles that are

consequently not counted

• If a specific usage of simple range counting must be validated by rainflow, why not use rainflow in the first place?

Appendix

PSD and Time History Example PlotsSingle Block

PSD and Time History Example PlotsUnimodal

PSD and Time History Example PlotsUnimodal

PSD and Time History Example PlotsTwo-block

PSD and Time History Example PlotsBimodal

Comparison of Spectral Fatigue Methods

• Various published spectral methods compared vs. rainflow cycle identification• Narrow band or Rayleigh approximation (NB)

• Dirlik (DK)

• Alpha 0.75 (AL)

• Ortiz and Chen (OC)

• Zhao and Baker (ZB)

• Single Moment (SM)

• Wirsching and Light (WL)

• Benasciutti and Tovo (BT)

Damage Rate ComparisonSingle Block PSDs

1

10

1 10

Spectral Methods vs. Rainflow Damage RateSingle Block PSDs

m = 3.324

rainflow

DNB

DDK

DAL

DOC

DZB

DSM

DWL

DBT

1

10

100

1000

1 10 100 1000

Spectral Methods vs Rainflow Damage RateSingle Block PSDs

m = 7.3

rainflow

DNB

DDK

DAL

DOC

DZB

DSM

DWL

DBT


3

3

Spectral Methods vs. Rainflow Damage RateUnimodal PSDs

m = 3.324

rainflow

DNB

DDK

DAL

DOC

DZB

DSM

DWL

DBT

90

90

Spectral Methods vs Rainflow Damage RateUnimodal PSDs

m = 7.3

rainflow

DNB

DDK

DAL

DOC

DZB

DSM

DWL

DBT


0.1

0.1

Spectral Methods vs. Rainflow Damage RateUnimodal PSDs

m = 3.324

rainflow

DNB

DDK

DAL

DOC

DZB

DSM

DWL

DBT

4

4

Spectral Methods vs Rainflow Damage RateUnimodal PSDs

m = 7.3

rainflow

DNB

DDK

DAL

DOC

DZB

DSM

DWL

DBT

Damage Rate ComparisonTwo-Block PSDs

0.1

1

10

100

0.1 1 10 100

Spectral Methods vs. Rainflow Damage RateTwo-Block PSDs

m = 3.324

rainflow

DNB

DDK

DAL

DOC

DZB

DSM

DWL

DBT

10

100

1000

10 100 1000

Spectral Methods vs Rainflow Damage RateTwo-Block PSDs

m = 7.3

rainflow

DNB

DDK

DAL

DOC

DZB

DSM

DWL

DBT

Damage Rate ComparisonBimodal PSDs

0.5

5

0.5 5

Spectral Methods vs. Rainflow Damage RateBimodal PSDs

m = 3.324

rainflow

DNB

DDK

DAL

DOC

DZB

DSM

DWL

DBT

20

200

20 200

Spectral Methods vs Rainflow Damage RateBimodal PSDs

m = 7.3

rainflow

DNB

DDK

DAL

DOC

DZB

DSM

DWL

DBT

Questions and Discussion

International Conference on Engineering Vibration … of Fatigue Cycle Identification Methods •Simple range count vs. range pair count vs. rainflow count •ASTM E1049 − 85 (2011)

Documents