Upload
others
View
6
Download
3
Embed Size (px)
Citation preview
Composite Thermal Damage Measurement with Handheld FTIR
November 14, 2013 B.D. Flinn, T. Howie, A.C. Tracey, D. Pate,
J.T. Morasch, E. Roberts University of Washington
Composite Thermal Damage Measurement with Handheld FTIR
§ Motivation and Key Issues § Damage detection in composites requires different
techniques than metals § Incipient thermal damage occurs below traditional NDE
detection limits
§ Objective § Determine if handheld FTIR can detect thermal damage and
guide repair § Approach
§ Characterize panels with controlled thermal damage and perform repair based on FTIR inspection
1
FAA Sponsored Project Information
§ Principal Investigators & Researchers § Brian Flinn (PI) § Ashley Tracey (PhD student, UW-MSE) § Tucker Howie (PhD student, UW-MSE)
§ FAA Technical Monitor § David Galella (year 3) § Paul Swindell (year 1 & 2)
§ Industry Participation § The Boeing Company (Paul Shelley, Paul Vahey) § Sandia National Lab (Dennis Roach) § Agilent (formerly A2 Technologies)
2
Background § Continuation of existing project (year 3 of 3) § Years 1 and 2 (A2 Technologies, Boeing and U of DE)
§ Characterization of homogeneous thermal damage § Ultrasound § Short beam shear (SBS) § Microscopy § Handheld FTIR (ExoScan)
§ Calibration curve for FTIR detection of thermal damage (SBS data)
§ Mapped surface of localized thermal damage § Year 3 (UW and Boeing)
§ 3-D characterization of localized thermal damage § Contact angle and fluorescence spectroscopy § FTIR guided scarf repair § Test repair
3
Thermal Damage vs. Detection Method
§ SBS, ultrasound, and microscopic analysis of composites with thermal damage § Properties degrade before detection possible à need
method to detect incipient thermal damage (ITD)
Short Beam Shear Strength Retention vs. Temp./Time – Epoxy 1
Onset of crack
development visible in
micrographs,
No cracks visible in micrographs
Damage becomes visible in C-Scans
4
Thermal Damage vs. Detection Method
§ SBS, ultrasound, and microscopic analysis of composites with thermal damage § Properties degrade before detection possible à need
method to detect incipient thermal damage (ITD)
Short Beam Shear Strength Retention vs. Temp./Time – Epoxy 1
Onset of crack
development visible in
micrographs,
No cracks visible in micrographs
Damage becomes visible in C-Scans
4
Summary of Work Completed
FTIR Contact Angle
Fluor-escence
Thermal Damage ✔ ✗ ✗ Resin rich (tooled) surface oxidation peaks characterized
Fiber rich (sanded) surface
oxidation peaks absent à multivariate analysis (MVA)
Fiber orientation signal varies with orientation
Surface finish
can only compare surfaces with MVA when surface finish
is same
5
Summary of Work Completed
FTIR Contact Angle
Fluor-escence
Thermal Damage ✔ ✗ ✗ Resin rich (tooled) surface oxidation peaks characterized
Fiber rich (sanded) surface
oxidation peaks absent à multivariate analysis (MVA)
Fiber orientation signal varies with orientation
Surface finish
can only compare surfaces with MVA when surface
finish is same
5
Experimental Overview
§ Thermally Damaged SBS samples § FTIR measurements on SBS samples § Develop calibration curve for FTIR from
SBS values § Predict evaluation set to validate model § Composite panel locally damaged § Panel Mapped using FTIR § Panel cut up for mechanical testing
§ SBS § Tg (DMA)
6
Materials and Process
§ Toray T800/3900 composites with various levels of thermal damage § SBS samples provided from Year 1 & 2 research § SBS samples thermally exposed in air § Locally damaged panels heated in air – UW/Boeing
§ Sand SBS surfaces with 180 grit Al2O3 sanding pads § Measure sanded surface with diffuse reflectance FTIR
§ 3 measurements per sample § 3 samples per time/temp
§ Use MVA to develop calibration curve for thermal damage § GRAMS IQ software
9 measurements
7
Materials and Process – FTIR
§ Mid-IR data region: 4000 cm-1 to 650 cm-1 § Diffuse reflectance sampling interface § Data collection: 90 coadded scans with 16 cm-1
resolution for background and specimen
An infrared beam path for diffuse reflectance
Mirror
Substrate Reflected
Beam
IR beam from interferometer
θi = 0
θr
To detector
ExoScan FTIR
8
Resin Rich vs. Fiber Rich Surfaces
§ Resin rich surfaces: oxidation peaks increase with damage
§ Fiber rich surfaces: oxidation removed by sanding Ø Need MVA to determine differences in spectra and
correlate to SBS data
0
0.1
0.2
0.3
0.4
0.5
0.6
1200 1300 1400 1500 1600 1700 1800 1900 2000
Abs
orba
nce
Wavenumber (cm-1)
High Damage
No Damage
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
1200 1300 1400 1500 1600 1700 1800 1900 2000
Abs
orba
nce
Wavenumber (cm-1)
Fiber Rich Surface Resin Rich Surface
9
Spectral Analysis
§ FTIR spectra of CFRP surfaces complex § Multiple constituents à many spectral peaks
§ How to analyze spectra with confidence? § Multivariate analysis!
§ Principal Component Analysis (PCA) § Exploratory to
identify trends § Peak locations and
intensities § Used to develop
models
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
650 1150 1650 2150 2650 3150 3650
Abs
orba
nce
Wavenumber (cm-1)
10
Effect of Sanding Variables on FTIR
§ Variables: temperature, method (hand vs. orbital), direction of sanding, grit size
-0.01
-0.005
0
0.005
0.01
0.015
-‐0.015 -‐0.01 -‐0.005 0 0.005 0.01
PC
2
PC1
1 hr @ 400 °F
1 hr @ 500 °F
Temperature Grit size
11
-0.006
-0.004
-0.002
0
0.002
0.004
0.006
0.008
-‐0.01 -‐0.005 0 0.005 0.01
PC
2
PC1
180 grit
240 grit
320 grit
Effect of Sanding Variables on FTIR
Ø FTIR Results influenced by sanding technique Ø Measure consistent surfaces and develop model
-0.005
-0.0025
0
0.0025
0.005
-‐0.015 -‐0.01 -‐0.005 0 0.005 0.01 0.015 0.02
PC
2
PC1
Parallel to Fibers
Perpendicular to Fibers
Direction of Sanding Sanding Method
12
-0.004
-0.002
0
0.002
0.004
-‐0.015 -‐0.01 -‐0.005 0 0.005 0.01 0.015 0.02 P
C2
PC1
Hand sanding
Orbital sanding
Developing FTIR Model
§ SBS samples sanded and measured with FTIR § FTIR spectra processed to remove baseline
effects § 1st derivative with Savitzky-Golay 7pt smoothing
§ Partial Least Squares model developed using MVA on processed spectra
Raw Spectra Processed Spectra
13
FTIR Model
ü Data fits model
R² = 0.9244
50
60
70
80
90
100
110
50 60 70 80 90 100
Pre
dict
ed S
BS
Ret
entio
n Va
lue
(%)
Actual SBS Rentention Value (%)
14
Model Validation
§ Model used to predict SBS retention in independent evaluation set
§ Error in model determined
%100*%actual
actualpredictedError −=
146
14 11
< 5 % 5 % ≤ x ≤ 10 % > 10 %
Nu
mb
er o
f S
amp
les
% Error
15
Localized Heat Damage – Process
16
Localized Heat-Damage Mapping
§ Different levels of thermal damage detected by FTIR Ø Cut panels into SBS and dynamic mechanical analysis
(DMA) samples for mechanical testing
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
650 850 1050 1250 1450 1650 1850
Abs
orba
nce
Wavenumber (cm-1)
Outside Heated Region Edge of Heated Region Center of Heated Region
0 50
100 150 200 250 300 350 400 450 500
0 25 50 75 100 125
Tem
pera
ture
(°F)
Time (min)
17
SBS and DMA Testing
§ Damaged panel cut into SBS and DMA coupons
§ Testing in progress § Compare SBS and Tg to
determine best method to correlate to FTIR
18
Summary
§ FTIR measurements sensitive to surface finish § Need to test samples with consistently
sanded surfaces § Calibration model developed from SBS
samples § Model predicted evaluation set well
§ Panels created with localized thermal damage and surface mapped with FTIR
§ SBS and DMA testing in progress to correlate mechanical damage to FTIR spectra
19
Future Work
§ Map thermally damage panels provided from Years 1 & 2 with FTIR
§ Determine mechanical test to correlate damage to spectra
§ Characterize thermally damage of panels provided from Years 1 & 2
§ Perform scarfed repair guided by FTIR § Test scarfed repair
20
Looking Forward
§ Benefit to Aviation § Improved damage detection § Greater confidence in repairs
§ Future needs § Application to other composite
systems § Other applications of handheld FTIR
§ Chemical damage § Surface prep for bonding
21
Fluorescent Thermal Damage Probe
1 inch
Thermal exposure on composite Fluorescence emission
Fluorescence inspection
Funded by: The Boeing Co.
Probe-doped coating 1 hr @ 450 °F
22
Acknowledgements
FAA, JAMS, AMTAS Boeing Company
Ø Paul Vahey, Paul Shelly, Greg Werner, Megan Watson, Jim Chanes
Sandia National Labs Agilent Technologies UW MSE
23