University of Bristol Department of Aerospace Engineering FATIGUE LIFE PREDICTION FOR COMPOSITE MATE

1
University of BristolDepartment of Aerospace
EngineeringFATIGUE LIFE PREDICTION FOR
COMPOSITE MATERIALS - A NEW APPROACH

• Mostefa Bourchak,
• Dr Ian Farrow,
• Dr Ian Bond

2
Current Fatigue Assessment For Composites

• There is no generally applicable and reliable
  fatigue life prediction method available for
  fatigue of CFRP under spectrum loading.
• Instead use
• Conservative static design
• _at_ low stress/strain levels
• for static notch/impact allowables
• Conservative test validation
• hot/wet cycling
• damage tolerance demonstration

3
Classical Approach

• Three stages
• Load Cycle History Idealisation
• Aircraft load history monitoring and idealisation
• Counting methods
• Spectrum representation exceedence curves
• Standard load spectra
• Summary of load history idealisation
• Data Manipulation
• Cyclic data manipulation empirical constant
  life diagrams
• Damage Accumulation
• Cumulative damage rule (as before)
• OR
• Property degradation e.g. RESIDUAL STIFFNESS,
• RESIDUAL STRENGTH wear-out, PROBABILISTIC...

4
Classical Approach

• Typical Results

CEP-50/50-4BT YH
1.E09
CEP-50/50-4BC YH
CEF-0/91/9-FAT IB
CEF-0/91/9-FAT IB
CEF-0/91/9-2BT lm SD
1.E08
CEF-0/91/9-2BT hm SD
Conservative
CEF-0/91/9-FAT SD
CEF-0/91/9-FAC SD
CEF-0/91/9-FAT FV
CEF-0/91/9-SFT FV
1.E07
CEF-0/91/9-EFT FV
A
CEF-0/91/9-FAC FV
CEF-0/91/9-SFC FV
CEF-0/91/9-EFC FV
1.E06
CEF-0/91/9-2BT lm FV
Test
CEF-0/91/9-2BT hm FV
32
34
1.E05
36
CEP-25/50/25-SGC YD
CEF-25/50/25-FA PB
CEF-25/50/25-FAC SD
1.E04
CEF-25/50/25-FAC FV
CEF-25/50/25-SFC FV
CEF-25/50/25-EFC FV
CEP-25/50/25-TWC P
1.E03
CEP-25/50/25-TWC LOM P
Unconservative
CEP-25/50/25-TWC 2LOM P
CEP-25/20/25-TWC 4LOM P
CEP-25/50/25-TWC 4LOM P
1.E02
CEP-25/50/25-TWC 4H2LOM P
CEF-25/50/25-FAC IB
1.E02
1.E03
1.E04
1.E05
1.E06
1.E07
1.E08
1.E09
CEF-25/50/25-SFC IB
CEF-25/50/25-EFC IB
Prediction
Classical fatigue analysis prediction for
carbon-fibre composites.
5
New Approach

• At present
• no general consistent design approach for
  modelling cumulative damage predicting residual
  strength/fatigue life
• long term behaviour is currently unpredictable
• (probably!) not possible to derive cumulative
  damage law based on C/A data which will be
  superior to Palmgren/Miner
• developments MUST include effects of
  interactions between stresses at different levels

6
New Approach

• Load Cycle History Idealisation
• plus
• DAMAGE EVENT DEFINITION
• Data Manipulation
• Cyclic data manipulation empirical constant
  life diagrams
• Damage Accumulation
• Cumulative damage rule (as before)
• OR
• Property degradation e.g. RESIDUAL STIFFNESS,
• RESIDUAL STRENGTH wear-out, PROBABILISTIC...

7
Damage Event Definition
Mix Events
E.g. Significant increasing ve or decreasing -ve
events
8
Damage Event Definition

• Account for mix - methodology based on 2-block
  base data
• e.g. by R ratio, severity, sm sa

Investigation of significant damage events by
thermal transient monitoring
9
Mix Event
10
Effect of Cycle Mixing
11
Effect of Cycle Mixing
12
Data Manipulation
Test S-N data
Constant life fitting
?a
Mix S-N data
R

S


N
?m
?

13
Damage Accumulation
f
N ? dD ? f(?m, ?a, D)
dD f(?m, ?a, D) dN
e.g. D ?A(ni / Ni) B(ni/Ni)C
N total of applied cycles when D ?A(ni /
Ni) B(ni/Ni)C Df
14
Fatigue Life Prediction
15
Current UoB Activities

• Fatigue Life Prediction For Fibre Reinforced
  Plastics Under Complex Loading
• EPSRC/AWHL/DAP CASE Studentship 60k over 3 yrs
• 1. Damage event characterisation
• Mix counting algorithm (FALSTAFF,TWIST, HELIX
  etc.)
• Investigate damage event transient wear-in states
  by thermal/p.d./acoustic monitoring
• Flexural testing of damage feature specimens,
  modified lathe (gt10Hz)
• Laminate specimens with ply drop, fibre wrinkle,
  tow break
• 2. Prediction methodology verification
• Two-block testing to complement AWHL/DAP
  programmes
• Assess prediction method (FALSTAFF,TWIST, HELIX
  etc.)
• Computer controlled servo-hydraulic test machine
  (100kN, lt5Hz)
• Standard in-plane specimens (e.g. CRAG type) with
  ply drop, fibre wrinkle, tow break

16
Summary

• (Currently!) No reliable, generally applicable
  method of predicting fatigue response for any
  situation/application.
• Fatigue primary design issue
• More efficient designs_at_higher stress/strain
  levels ? Need for reliable fatigue life
  prediction
• UoB new approach
• Damage event definition within a load-time
  history
• ?
• Experimental evidence of mix effect on fatigue
  life
• ?
• Significant improvement in predictive accuracy