Planning meeting, 18 -19 November 2003, Austin TX

							                          MAPOD
• Planning meeting, 18 - 19 November 2003, Austin TX
• After considerable discussion on defining and prioritizing
  goals, the following prioritization/focus list for the
  proposed consortium was arrived at:

1. Validation of models (flaw response)
2. Totem pole of important factors
3. Design a multi-point calibration
4. POD transfer function; boundary conditions are important
5. NDE engineering tools
6. Model benchmarks
7. Demonstration with existing data
      MAPOD Demonstration
• PURPOSE FOR TODAY
  – develop more specific program plan
     • scope, level of effort
     • actionable items


• charge
  – develop demonstration project which deals with
    ALL the issues
         MAPOD purpose
• PROBLEM
 – MIL-HDBK-1823 is good, but too expensive
 – want quantified NDT reliability, cheaper
            MAPOD demo
• CONSENSUS TARGET PROBLEM
 – two layer structure à la wing skin + stringer
 – aluminum, thicknesses ~0.100 - 0.200”
 – discontinuity of interest: fatigue cracks
   nucleated at or near fastener holes

 – exact details TBD LATER
                     Objectives
• evaluate feasibility and limitations of using MAPOD for
  transferring POD from geom 1 to geom 2 of multilayer
  fastened structures
• evaluate feasibility and limitations of using MAPOD for
  transferring POD from notched samples to actual cracked
  multilayer fastened structures
• evaluate feasibility and limitations of using MAPOD for
  transferring POD from lab to depot inspection of
  multilayer fastened structures
• evaluate sample size requirements
• evaluate potential for providing additional inspection info
  such as ROC curves and crack size quantification
                     Approach
• team of experts including modeling, reliability,
  instrumentation, & users
• specific multilayer inspection problem based on biggest
  requirement (ROI), cross-platform commonality
• provide useful demonstration example (s) for expanding
  scope of 1823 to support MAPOD approaches
• transfer function emphasis
• eddy current and ultrasonics
• identify modeling needs to support MAPOD
• challenge to community: approximately 2 year effort
    Task 1: Identify and evaluate geometrical
    variables for multilayer fastener structures
•   Note: known key variables are:
     – hole diameter, fastener heads/material, thickness, number of layers, crack
       location


•   DOE to determine important factors and POD transfer functions (or
    penalty functions)
     – defines specimen and testing requirements
          • make specimens
          • conduct inspections
          • conduct data analyses
•    Evaluate modeling capabilities for the inspection variables. This
    should provide insight into formation of POD transfer functions
•    Blind validation of transfer functions (or penalty functions)
    Task 2: Identify and evaluate difference in
       response between EDM and cracks
•   Note: known key variables are:
     – crack location, number of cracks, residual stress, crack orientation and
       aspect ratio


•   DOE to determine important factors and POD transfer functions (or
    penalty functions)
     – defines specimen and testing requirements
          • make specimens
          • conduct inspections
          • conduct data analyses
•    Evaluate modeling capabilities for the inspection variables. This
    should provide insight into formation of POD transfer functions
•    Blind validation of transfer functions (or penalty functions)
   Task 3: Identify and evaluate factors that
effect variability in field and depot inspections
•   Note: known key variables are:
     – human, surface, hole quality, environment, sealant/bond/finish, speed,
       calibration


•   DOE to determine important factors and POD transfer functions (or
    penalty functions)
     – defines specimen and testing requirements
          • make specimens
          • conduct inspections
          • conduct data analyses
•    Evaluate modeling capabilities for the inspection variables. This
    should provide insight into formation of POD transfer functions
•    Blind validation of transfer functions (or penalty functions)
            Task 4 Factor Sensitivity Study

•   Evaluate significance of factors using data from Task 1-3 (statistical
    analyses)
     – factors and first order interactions
•   Evaluate sample sizes required to support XFN, FMA
•   Evaluate confidence bound methods for XFN, FMA
                    Totem Pole
•   system calibration (Rummel, Forsyth, Goldfine)
•   probe characterization (ISU, Gray, Patton, Broz, EWI)
•   model validation (Knopp, Vukelich, ISU, Gray, Todorov)
•   XFN validation (Smith, Hugo, Patton, ISU, Annis)
•   specimen design (Goldfine, Forsyth, Brausch, Annis,
    Spencer, Moore)
•   cracks versus notches (AMMTIAC, Thompson, Lindgren,
    Hugo, NRC)
•   number of specimens (Spencer, Annis)
•   noise (Goldfine, Annis, Spencer)
•   How do you know you are right??? (Malas, Vukelich,
    Thompson, Knopp)
•   Concurrent programs – data capture (ALL)
    Totem Pole - system, probe
• machine
  – drift, freq response, ...
• cabling/connectors
  – impedance
• probe
  – geometry, impedance
• Scan plan, increments, digitization,
  coverage
  Totem Pole - model validation
• Benchmark problem
• Error analysis
  – Uncertainty in inputs
  – Uncertainty in measurements used for
    validation
• Need target requirements
  – Need accurate estimate of POD – this will
    define requirements in terms of model accuracy
  Totem pole – XFN validation
• Generic range of validity
• Review existing protocol
• Factor interactions – how are they captured
  – interferences
• Error analysis
  – Uncertainties in inputs, uncertainties in
    measurements used for validation
  Totem pole - specimen design
• Based on goal of project
  – FMA, XFN validation
• material
  – vintage, heat treat, product form, directionality,
    em properties, coatings/sealants, ...
• machining
  – Tolerances, surface finish, residual stress, hole
    condition, ...
• crack
  – loading (spectrum?), single/multi axial,
    Totem pole - cracks vs notches
•   cracks from in-service
•   cracks from lab manufacture
•   EDM notches from manufacture
•   measure, understand difference
    – re
      Totem Pole - number of
           specimens
• Need guidelines on what is required to
  execute a FMA or XFN approach
  – complex, based on
     • what factors are included in study
     • what variances/uncertainties


• Remember false calls
          Totem pole – noise
• Need to characterize the response of actual
  (structure, engine, etc.) to understand
  variability due to in-service conditions
• Geometries, qualities, …(see specimen
  design)
 Totem pole – How do we know
         we are right
• Compare FMA, XFN to empirical
• Within what limit/confidence
   –   a hat vs a
   –   POD curve
   –   can model/xfr function accommodate key factors
   –   can model/xfr function accommodate complicating issues
         • add empirical corrections from specialized experiments if
           model/xfr function is not capable
        • hole quality
        • crack morphology
        • All noise issues
      Totem pole - others?
• ?
Source
              Available specimen sets
                            Type of              NDT technique            models available
                            Discontinuity
FAA-AANC lap splice         surface breaking     manual, pencil probe     no, but validation data
joint specimens             fatigue cracks       eddy current             available
(available now)
RFC data                    surface breaking     ECIS                     no, but validation data
                            fatigue cracks in                             available
                            engine rotating
                            components
UDRI TESI specimens         artificial           automated ultrasonics    ?
(available now)             metallurgical
                            discontinuities in
                            titanium
Royal Australian Air        cracks under         ultrasonics              ?
Force F-111 specimens       fasteners in
(available 2006?)           aluminum
Canadian Forces bolt hole   fatigue cracks in    manual, automated bolt   not yet, mid-2006 for
specimens                   holes in aluminum    hole eddy current        validated model (using
(available fall 2006)       plate                                         ECSIM)
AFRL-MLLP C-5               fatigue cracks       automated eddy current   VIC-3D®
(Lockheed) specimens
C-141 weep hole POD         EDM notches          automated ultrasonic     limited models:
study -                                                                   a) 2D BEM model,
WR-ALC / SAIC Ultra                                                       b) 3D hybrid model under
Image Int. /                                                              development
Computational Tools
C-130 hat section POD       EDM notches          automated ultrasonic     limited models:
study -                                                                   a) UTSIM (does not
WR-ALC / SAIC Ultra                                                       address guided wave
Image Int. /                                                              characteristics)
Computational Tools                                                       b) 3D hybrid model under
                                                                          development
  Partial Population of Potential
            Participants
• See word doc