Evaluating Low Frequency Eddy Current Based Technologies for - PowerPoint PPT Presentation

Evaluating Low Frequency Eddy Current Based Technologies for Detecting Fatigue Cracks in Multi-Layer Metallic Structures 2014 ASIP Conference Dec 3, 2014 Gary Steffes AFRL/RXCA Integrity  Service  Excellence

Acknowledgements • Co-authors: – USAF – David Campbell, Charles Buynak – Texas Research Institute-Austin (TRI) - Doyle Motes, David Forsyth, Mark Keiser, Michael Mazurek – Computational Tools – John Aldrin • AFRL Sustainment Office (sponsors) – Stephan Wolanczyk, Stephanie Flanagan Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Outline • Introduction • Program Overview • Fatigue Crack Specimens • Sensor Systems • Preliminary Results • Accounting for Changes in Structural Geometries • Summary • Conclusions 3 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Introduction • Background – Low frequency eddy current (EC) methods used to inspect sub-surface aircraft structures – Increasing requirements for detection as fleets age • Smaller subsurface defects • Larger areas of coverage • Motivation – Reached the limits of current Low Frequency EC – Need to interrogate thick, complex structures using NDE – Avoid a/c structure disassembly • Impacts schedule and cost • Possibility of maintenance induced damage Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Program Overview Program targeted • Program goals were: specific inspections on – Collect data from candidate aircraft structure NDI sensor systems for detecting sub-surface fatigue cracks in complex structures • Fabricate samples – Results evaluated by AFRL NDI inspections of thick, complex wing structures can be time consuming and difficult 5 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Fatigue Crack Specimen Sets • Simulate second layer cracking 0.500” • Four 7075-T6 Al 0.375” specimen sets – Thicknesses (0.500”, 0.375”, 0.250”, 0.100”) – 4 - ¼” holes per specimen • Sixty (60) specimens tested for each 0.250” thickness – 240 holes per set 0.100” – 60 cracks per set 6 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Fully Assembled Inspection Targets • Specimen Setups Included: • Boeing (Cd plated) steel fasteners (BACB30LU) and Al nuts 1 st layer consisted of 0.170” and 0.250” 7075-T6 Al top sheet • 2 nd layer contained 60 individual 7075-T6 Al specimens at • specimen set thicknesses 7 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Other Sample Set Information Fatigue Crack 0.170” Top Sheet • Crack content information • Cracks only in holes • Crack types • Toward adjacent fastener holes 0.500” • Toward edges Specimen Set Setup #1 • Cracks orientations occur at 12 and 6 o’clock only 0.250” Top Sheet Fatigue Crack • DO NOT CONTAIN • Double cracked holes • Complete hole-to-edge or hole- to-hole cracks 0.500” Specimen Set Setup #2 8 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Sensor Systems Examined 2. Low Frequency EC 1. Giant Probe Magnetoresistive (GMR) Array 3. Conventional EC Ring Probe 5. Remote Field EC Ring Probe on Scanner 4. Remote Field EC Ring Probe 9 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

1. GMR Linked to MAUS • Scans completed by TRI’s NDI Level III inspector • Data interpretations – Raw c-scans – Automated Defect Analysis (ADA) processing codes • Scan frequency range 150-300Hz • High false call rates for both scans • Issues for interpreting C-scan data included: – Target edge effects – Sensor sensitivity bias – Inconsistencies between scans – Effects from cracks at adjacent inspection sites 10 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Example Raw C-scan Produced by GMR w/MAUS Edge effect GMR C-scan – circles indicates EDM notches (from individual specimen) Large edge effect (end of assembled specimens) Sensitivity Bias (appears Fastener site Magnetized as data striping) signal fastener 11 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

2. LFEC Sensor Mounted to the MAUS Called crack “Missed crack” Called crack • Completed by AFRL personnel • Scan frequency – 550 Hz used for thin top sheet (0.170”) – 225 Hz used for thicker top sheet (0.250”) • Visual interpretation difficult due to effects from adjacent holes (green arrows) 12 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

3. EC Ring Probe (Current method) • Scans completed by TRI’s NDI Level III inspector • Method provided a baseline for comparison • Inspections adapted from AF tech order procedures – Frequency set by deepest area to be inspected – Full Screen Height deflection (%FSH) recorded for each site – Detect threshold set at a ≥ 30% FSH • Low false call rates and good results with Cd-plated steel fasteners 13 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

4. RFEC Ring Probe • Tests conducted by TRI’s NDI Level III inspector • NDI procedures developed specifically for test • Scan frequency - 100 Hz (exception - 0.170” top and 0.100” specimen set) • Effects from cracked adjacent holes seen for longer cracks 14 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

5. RFEC Ring Probe Mounted to the MAUS • Conducted by AFRL personnel • Same procedures used as RFEC ring probe (100 Hz frequency) • Edge effects and adjacent cracks were found to be less pronounced at 100Hz 15 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Caveats to Preliminary Results • Specimens – aircraft not typically constructed with 2.4”x 2.4” squares (edge effects) • Procedures were developed specifically for test • Mean crack lengths Specimen Set Mean Crack Thickness (in) Length (in) • Definition of a false call 0.100 0.175 • POD calculations – 0.250 0.200 – Results collected by two people 0.375 0.252 • Insufficient number 0.500 0.352 • Non-aircraft inspectors – Fastener head diameter (0.469”) • Warning: Extrapolating results for similar geometries can be bad (i.e. part specific) Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Results Table (0.250” top x 0.500” bottom) Inspection Method Correct (%) False Call (%) a90 (in) a90/95 (in) 1a. GMR ImagIn 13% 5% > 1.000 > 1.500 1a. GMR ImagIn Adj 20% 3% > 1.000 > 1.500 1b. GMR ADA 80% 17% 0.370 0.471 1b. GMR ADA 62% 12% 0.480 0.584 2. LFEC 52% 6% NA NA 3. Conventional EC ring 75% 0% 0.310 0.336 4. RFEC Ring Probe 90% 0% 0.212 0.220 5. RFEC Ring Probe MAUS 93% 1% 0.206 0.217 Best Results RFEC MAUS Tie RFEC MAUS RFEC MAUS 17 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Summary of Results Best a90 Results 0.170” Top Sheet 0.250” Top Sheet 0.100” Specimen Set RFEC Ring Probe RFEC Ring Probe 0.250” Specimen Set RFEC Ring Probe MAUS RFEC Ring Probe 0.375” Specimen Set Conventional EC Ring RFEC Ring Probe MAUS Probe / RFEC Ring Probe 0.500” Specimen Set RFEC Ring Probe MAUS RFEC Ring Probe MAUS • Sensitivities drop off with increasing depth for all sensor systems. • EC ring probe results are better than the GMR Array MAUS and the LFEC MAUS for all inspections described in this work. • RFEC averaged a better correct call rate than EC ring probe by 17% 18 Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Other Evaluations • Motivation – Account for changes in structural geometry present in aircraft inspections – Tapers and steps – Changes in fastener material • Evaluated signal response due to ET frequency per thickness • Difference in signal response for structures using steel vs. aluminum fasteners • RFEC Ring Probe system used • Setup – Configuration 1: 0.170” top x 0.100” bottom – Configuration 2: 0.170” top x 0.500” bottom – Same inspection procedures used in all cases Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.

Signal Effects Due to Frequency 0.170” Top x 0.100” Bottom 100 Signal Response Screen Height (%FSH) 90 80 70 60 50 40 Crack Steel Fasteners 50Hz 30 No crack Steel Fasteners 100Hz 20 Steel Fasteners 200Hz Steel Fasteners 300Hz 10 Steel Fasteners 400Hz 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 Crack Length (in) Distribution Statement A: Approved for public release; distribution is unlimited. Case Number: 88ABW 2014-5495.