ML18012A720
| ML18012A720 | |
| Person / Time | |
|---|---|
| Issue date: | 01/16/2018 |
| From: | Diaz A, Dib G, Holmes A, Jacob R, Morales R, Moran T, Carol Nove, Prowant L, Ramuhalli P Office of Nuclear Reactor Regulation, NRC/RES/DE/CIB, Pacific Northwest National Laboratory |
| To: | |
| Rezai A, NRR-DMLR 415-1328 | |
| Shared Package | |
| ML18002A510 | List: |
| References | |
| Download: ML18012A720 (32) | |
Text
PNNL-SA-131574 Ultrasonic Modeling and Simulation GERGES DIB, MATT PROWANT, RICHARD JACOB, AIMEE HOLMES, ROMARIE MORALES, TRACI MORAN, PRADEEP RAMUHALLI, AARON DIAZ Pacific Northwest National Laboratory CAROL NOVE, NRC COR Industry-NRC NDE Technical Information Exchange Meeting Rockville, Maryland January 16-18, 2018 0
Outline Modeling and simulation Objectives Approach and examples of results to date Key takeaways Next steps Limited coverage Objectives Approach Key takeaways Next steps 1
Outline Modeling and simulation Objectives Approach and examples of results to date Key takeaways Next steps Limited coverage Objectives Approach Key takeaways Next steps 2
Modeling and Simulation: Objectives Quantify the effectiveness of beam models with respect to coverage and flaw detection capability Quantify effects of component geometries, material properties, probe specifications, and other parameters on sound field characteristics and detection capability Quantify effect of input parameter uncertainties Coverage area for beam within 6 dB in simulation results Identify gaps in common software tools for 2D and 3D simulation of ultrasonic inspection.
Develop guidance for use and interpretation of simulation models for conveying information on Coverage area for beam within 20 dB effectiveness of inspections.
Foundation for confirming solid technical basis for conducting, interpreting, and applying ultrasonic modeling to assess the effectiveness of inspections of NPP components 3
Approach Leveraging, when possible, prior work at PNNL and elsewhere Verification and validation of simulation tools Variability of calculated flaw amplitudes; initial methods for noise incorporation in simulations Sensitivity of calculated flaw amplitudes to parameter uncertainties Metrics for comparing simulation and experimental results This phase of research Compare and contrast ultrasonic beam calculations from commonly used ultrasonic NDE simulation tools Determine if correlations between sound fields and flaw amplitudes may be inferred Quantify sound field and flaw amplitudes as a function of several parameters that are hypothesized to influence ultrasonic amplitudes Define techniques to include noise in simulation results - SNR Quantify uncertainty in simulation results and its impact on interpretability 4
Key Takeaways to Date from Modeling Studies Some software tools have limited ability to simulate beam propagation through complex materials (welds, coarse grained, or anisotropic base material)
Software tools evaluated for ultrasonic beam Simulated Ultrasonic modeling show some differences in beam Beam in Isotropic Material coverage estimates in isotropic materials Comparisons of simulation results and interpretation will require calibration (normalization) of calculated ultrasonic flaw amplitudes Beam models show limited correlation with Simulated Ultrasonic calculated and measured amplitudes from flaws Beam Through a Weld Flaw amplitudes are a complex function of beam energy as well as a number of other parameters 5
Assumptions and Limitations of Research Industry-accepted standards and guidance for UT examinations used to develop bounding cases (geometry, probe, flaw parameters)
Limit conditions and inform test matrix development Focus is on developing guidance for:
Interpreting model results for applicability to coverage and detection capability Identifying and documenting variables impacting coverage and detection calculations Quantifying uncertainty in calculated values of sound field coverage and detection ability Assessing the value of 2D and 3D simulations Focus of effort is NOT on assessing ability to discriminate between responses from flaws and non-flaws 6
Parameters Influencing Ultrasonic Simulation Results Parameters include probe, coupling, wedge, material, flaw, and data acquisition system parameters Many of these parameters cannot be specified with sufficient precision for modeling purposes Add uncertainty in the Aleatory Epistemic Bias simulation results
- Specimen moduli
- Probe squint
- Element coupling
- Specimen density
- Probe orientation
- Wedge coupling
- Specimen
- Specimen surface
- Material thickness
- Grain noise attenuation
- Flaw height
- Grain boundaries
- Electronics
- Flaw length
- Flaw morphology
- Piezoelectric
- Flaw tilt element transfer
- Flaw skew function
- Probe shape
- Wedge dims
- Wedge moduli
- Wedge density 7
A Brief Summary of Prior Results Prior (Phase I) efforts: assessing variability in simulated flaw amplitudes Sensitivity of calculated amplitudes to key parameters Metrics for comparing simulation and experiment, in the presence of noise Results documented in ML17082A190:
Validation of Ultrasonic Nondestructive Examination (NDE) Computational Models -
Phase 1, PNNL-26336 (2017)
Histogram of 358 different angle Differences attributed to beam inspections Inability to precisely quantify experimental factors (uncertainty)
Approximations inherent in models Flaw amplitude from simulation results and experiments appear to compare reasonably well under ideal conditions (isotropic materials), although noise and uncertainty in input parameters can limit simulation fidelity.
8
Normalization of Beam Amplitudes Enable Comparison Across Probes and Software Tools 2.25 MHz SV, 45o, 6.35 mm 2.25 MHz SV, 60o, 6.35 mm 5 MHz SV, 45o, 6.35 mm 5 MHz SV, 60o, 6.35 mm Unnormalized Beam Amplitude 2.25 MHz L, 45o, 12.7 mm 5 MHz L, 45o, 12.7 mm vs. Distance from Probe Beam Simulations for Six Ultrasonic Probes Normalized Beam Amplitude Note: Beam amplitude profiles were obtained along the vs. Distance from Probe 9 probe axis.
Sound Field Simulations Examples Shear wave probes Refracted Shear Wave Inspection Refracted Longitudinal Wave Using Single Element Probe Inspection Using TRL Probe 5 MHz, 60o, 6.35 mm (1/4) 2 MHz, 45o, 15x25 mm 5 MHz, 45o, 12.7 mm (1/2) 2 MHz, 60o, 15x25 mm Software tools evaluated for ultrasonic beam modeling show some differences in beam coverage estimates in isotropic materials.
10 Note: Figures show beam shapes with a -12 dB cutoff.
Beam Simulations Summary Probe # Type Element Shape Frequency (MHz) Size (mm) Angle Wave Mode UltraVision CIVA Difference
(%) (%) (%)
1 Single Circular 1.00 6.35 45 shear 1.2 10.0 8.8 2 Single Circular 1.50 6.35 45 shear 2.0 12.1 10.1 3 Single Circular 2.25 6.35 45 shear 3.1 15.9 12.8 4 Single Circular 5.00 6.35 45 shear 12.2 29.8 17.6 5 Single Circular 1.00 6.35 60 shear 0.0 6.4 6.4 6 Single Circular 1.50 6.35 60 shear 0.0 7.4 7.4 7 Single Circular 2.25 6.35 60 shear 1.2 9.2 8.0 8 Single Circular 5.00 6.35 60 shear 5.5 18.3 12.8 9 Single Circular 1.00 12.7 45 shear 9.8 26.3 16.5 10 Single Circular 1.50 12.7 45 shear 32.4 11 Both simulation Single tools2.25 Circular appear to be capable45of providing 12.7 shear similar14.9ultrasonic beam 17.5 27.1 44.7 17.6 12 amplitudes Single in isotropic Circular materials.
5.00 Experimental 12.7 45 data is needed shear to confirm 78.8 observed 81.9 3.1 13 Single Circular 1.00 trends from 12.7 simulation 60 results.
shear 4.7 16.4 11.7 14 Single Circular 1.50 12.7 60 shear 7.5 20.0 12.5 15 Single Circular 2.25 12.7 60 shear 14.1 28.6 14.5 16 Single Circular 5.00 12.7 60 shear 52.9 60.0 7.1 101 Dual Rectangular 2.00 20x34 45 Long 69.4 79.0 9.6 102 Dual Rectangular 2.00 20x34 60 Long 82.0 73.6 8.4 103 Dual Rectangular 1.00 20x34 45 Long 54.1 57.3 3.2 104 Dual Rectangular 1.00 20x34 60 Long 51.0 26.4 24.6 105 Dual Rectangular 2.00 15x25 45 Long 79.2 74.1 5.1 106 Dual Rectangular 2.00 15x25 60 Long 52.5 31.6 20.9 107 Dual Rectangular 1.00 15x25 45 Long 44.3 37.8 6.5 108 Dual Rectangular 1.00 15x25 60 Long 28.6 9.3 19.3 11 Note: Beam amplitude computed at 50 mm depth relative to maximum overall beam amplitude.
How Does Beam Coverage Relate to Flaw Detection?
Software: CIVA Commonly used in nuclear industry for simulating ultrasonic NDE processes Capable of simulating the interaction of ultrasonic energy with a flaw Flaws: Reference flaw (10 mm x 20 mm, rectangular flaw) in a 50 mm thick stainless steel specimen Metric: Flaw amplitude Generally used for detection Maximum amplitude in B-scan used as metric in this study Complex function of flaw parameters (depth, length, orientation, tilt, morphology) and other parameters that influence incident sound field Beam metric: Maximum beam amplitude at the ID of the specimen One of many possible metrics for representing sound fields Scan direction Ultrasonic Probe Flaw 12 End view Side view
Flaw Amplitudes vs. Beam Amplitudes Inferring flaw amplitudes from ultrasonic beam amplitudes alone is challenging. Knowledge of additional parameters and physics of ultrasound is critical.
13
For a Given Flaw, Flaw Amplitudes Are a Function of Several Parameters 1 MHz 1.5 MHz 5 MHz 2.25 MHz Normalized Beam Amplitude (Simulation) at Specimen ID (dB)
All data in plots are from refracted shear wave inspection simulations, with a 20 mm x 10 mm rectangular flaw.
14
Flaw Amplitudes vs. Beam Amplitudes Refracted Angle Blue: Simulation Green: Empirical Probe Aperture Inferring flaw amplitudes from ultrasonic beam amplitudes alone is challenging. Knowledge of additional parameters and physics of ultrasound is critical.
15
Key Takeaways to Date from Modeling Studies Some software tools have limited ability to simulate beam propagation through complex materials (welds, coarse grained, or anisotropic base material)
Software tools evaluated for ultrasonic beam Simulated Ultrasonic modeling show some differences in beam Beam in Isotropic Material coverage estimates in isotropic materials Comparisons of simulation results and interpretation will require calibration (normalization) of calculated ultrasonic flaw amplitudes Beam models show limited correlation with Simulated Ultrasonic calculated and measured amplitudes from flaws Beam Through a Weld Flaw amplitudes are a complex function of beam energy as well as a number of other parameters 16
Next Steps: Modeling and Simulation Modeling beam propagation through welds Appropriate techniques to model welds due to anisotropy and weld dendrites.
Investigate extents of beam redirection/change in expected coverage when compared to homogeneous isotropic assumptions.
Approaches for incorporating noise (material noise, measurement noise) in simulation results SNR estimates, flaw detection ability Quantification of uncertainty in simulation input parameters Uncertainty propagation Empirical measurements Evaluate effects of limited insonification of flaw (no corner response, flaw at periphery of beam) Effects of Noise on Fidelity Limited coverage of Simulation Results 17
Simulation and Modeling: Activities under NRC RES-EPRI Memorandum of Understanding (MOU)
Efforts at PNNL parallel efforts in modeling and simulation at EPRI To date, focus under MOU activities have been on Reviewing technical reports Sharing key findings Future coordinated research plan under NRC RES-EPRI MOU is currently under development Sharing data (empirical and simulation)
Sharing simulation models Identifying and sharing best practices for modeling and simulation Exchanging specimens for independent data acquisition Looking for other ideas, where research activities can be coordinated, data from the activities can be independently assessed, and conclusions can be independently drawn 18
Outline Modeling and simulation Objectives Approach and examples of results to date Key takeaways Next steps Limited coverage Objectives Approach Key takeaways Next steps 19
Modeling and Simulation Also Supports Assessment of Limited Coverage Establish a basis to determine how far a flaw would have to emanate from an uninspectable region into an inspectable region before it can be reliably detected with UT Austenitic welds Less than 100% coverage due to access limits Weld tapers or transitions Permanent item blocking coverage Material such as CASS with no qualified single-sided procedures 20
Limited Coverage: Objectives Develop methods for ascertaining the largest flaw that may be missed given a specified beam coverage (less than 100%) in austenitic welds Assess the impact of materials in Example of SS Elbow to SS Pipe where Dual-Sided limited coverage examinations (ferritic Access was Obtained and Partially Limited due to Weld steel, stainless steel, cast stainless Geometry steel)
Identify important factors that govern flaw detectability and sizing in limited coverage examinations Quantify flaw detection rates in limited Example of SS Pipe to SS (or CASS) Valve where coverage examinations as a function of Only Single-Sided Access was Obtained due to important factors Limitation Caused by Valve Taper and SS/CASS Material 21
Examples of Limited Coverage Welds (PWRs)
Based on Relief Requests Proximity Percent Range of No Single Number CASS Weld Nozzle Valve Component to adjacent Safe-end Weld Component Description Examination Description Coverage Diameters, Sided of Welds Material taper taper taper taper component/ width geometry Claimed inches Procedure structure RCP nozzle CASS safe-end-to-CS 16 Single-sided from CS elbow side (a) 30 16 16 elbow(b)
RCP nozzle CASS safe-end-to-CS 8 Single-sided from CS elbow side 36 8 8 elbow(b)
RCP nozzle CASS safe-end-to-CS 16 Single-sided from CS pipe side (a) 30 16 16 6 pipe(b)
RCP nozzle CASS safe-end-to-CS 8 Single-sided from CS pipe side 36 8 8 pipe(b)
CASS safe-end-to-CS nozzle(b) 18 Single-sided from CS nozzle side (a) 12 18 17 CS nozzle-to-SS safe-end with SS weld 1 Single-sided from SS safe-end side 26 to 75 4 1 1 CS nozzle-to-SS safe-end with SS weld 1 Single-sided from SS safe-end side 14 1 1 CS nozzle-to-SS safe-end with Alloy 4 Single-sided from SS safe-end side 24 to 26 (c) 4 4 52 weld Axial scan from nozzle side.
CS nozzle-to-SS safe-end with Alloy 40.5 and 2 Circumferential scan from safe-end 38 2 2 52/152 42 side.
CS nozzle-to-CASS elbow with SS 3 Single-sided from CASS elbow side 49 to 65 34 3 3 3 weld CS nozzle-to-CASS elbow with SS 3 Single-sided from CASS elbow side 38 3 3 3 weld CS nozzle-to-SS safe-end with Inconel 4 Single-sided from SS safe-end side 75 42 4 4 weld CASS safe-end-to-CS nozzle(b) 4 Two sided (a) 12 4 4 CS nozzle to SS safe-end with Alloy 1 Two sided (a) 29 1 82/182 weld with Alloy 52 inlay(b)
CS nozzle to SS safe-end with Alloy 1 Two sided 31 1 82/182 weld with Alloy 52 inlay(b)
CS nozzle-to-SS safe-end with SS weld 3 Two sided 37 to 86 6 3 1 2 CS nozzle-to-SS safe-end with SS weld 1 Two sided 15 1 1 1 22
Limited Coverage Examples in PWRs RCP CASS Nozzle-to-CASS Safe-End-to (top) CS Elbow (bottom) CS Pipe From Sartain 2014, ML14051A109 23
Limited Coverage Examples in PWRs RCP CASS Pump-to-CASS Safe-End-to with Single Sided Access Only From Katzman 2009, ML090430304 24
Overall Approach to Addressing Limited Coverage Identify the most common and problematic weld configurations Develop a statistics-based design of experiments (DoE) to inform an optimized test matrix Data gaps will be filled in by modeling and simulation activities Develop metrics for assessing flaw detection capability under limited coverage conditions Assess detection capability for flaws in limited coverage situations.
25
Limited Coverage Assessments are Using Design of Experiments Approach Construct Design of Experiments (DoE) to identify the set of experiments (empirical and simulation) necessary to quantify the effectiveness of UT under limited coverage conditions Define the parameters likely to affect flaw detection Design specimens that can provide the necessary data for the evaluation Develop the test matrix (parameter combinations) for the experiments Inputs to DoE Insights from physics of ultrasonic testing and modeling results to date ML17318A118: Summary of Literature Search of Relief Requests on ASME Code,Section XI, Volumetric Examination Coverage Requirements for Piping Butt Welds, PNNL-26157, Rev. 1 (2017)
ML12011A130: An Assessment of Ultrasonic Techniques for Far-Side Examinations of Austenitic Stainless Steel Piping Welds, NUREG/CR-7113, PNNL-19353 (2011)
Generic inspection procedures ASME Boiler and Pressure Vessel Section XI and Code Case N770-1 26
Design of Experiments:
Factors and Levels Factor levels based on Factors Number List of levels Notes of levels typical conditions in the CASS-CS field Materials 3 CASS -SS SS - SS Thin, Medium, Assumed correlation between wall Wall Thickness 3 Conditions limiting Thick thickness and pipe diameter Weld Root Using best case scenario of no weld coverage include taper Condition 1 None root.
(weld and/or Probe Aperture 2 Small, Large Single Element, component), physical Probe Type 3 Phased Array, TRL All angles for Phased Array, access restrictions, Refracted Angle 4 30 O ,45O ,60O ,70O Conventional 45O ,60O ,70O ,
weld geometry, and TRL - 30O ,45O ,60O ,70O Shear is only applicable for material microstructure Wave Mode 2 Shear, Longitudinal conventional probe (CASS) Conventional - 1 MHz, 2.25 MHz, Probe 1 MHz, 2 MHz, 2.25 5 MHz 4
Frequency MHz, 5 MHz Phased Array - 1 MHz, 2 MHz TRL - 1 MHz, 2.25 MHz, 5 MHz Metrics for quantifying Flaw Ongoing assessment with respect to size distributions, aspect ratio, location, orientation, and tilt. Other factors may also be included as coverage along with Parameters assessment progresses.
uncertainty bounds are being developed 27
Key Takeaways to Date (Limited Coverage Research)
DoE significantly reduces the number of specimens and experiments needed Recommended number of experiments reduced from 1836 (full factorial design) to 118 The actual numbers may change a little based on allowable flaw parameters, and allowable probe-frequency combinations Specimens under consideration are welds between CS - CASS SS - CASS SS - SS Specimen fabrication needs can be identified from DoE results Fabrication needs based on available specimens and flaw dimensions, and configurations that can be evaluated in simulation Number of flaws per specimen expected to range from 6-16 depending on size of the specimen, and length, orientation, and depth of the flaw. 28
Mockup Designs Leverage available specimens Mockups of primary coolant headers in a BWR recirculation system, austenitic stainless steel welds Flaws include implanted thermal fatigue cracks, saw cuts, and notches Flaw depths range from 10% to 64% TW deep Other dissimilar weld mockups available Planned 1-2 additional mockups to accommodate combinations of factors and flaws Limited coverage can be real as determined from weld crown or root presence, or simulated by restricting axial scan position Flaw size and location are being finalized 29
Next Steps Complete mockup design and 0.272 (6.9 mm) deep notch 45%
TW fabrication Develop metrics for assessing flaw Empirical detection capability under limited coverage conditions Simulated Complete flaw response measurements and simulations Empirical measurements from 2.25 MHz, 45° shear 5.0 MHz, 45° shear mockups Simulations of flaws of interest in test matrix that are not part of mockups Assess detection capability for flaws in limited coverage situations. 30
Summary Ongoing research on limited coverage examinations will include both simulation and empirical studies Design of experiments can assist in targeted evaluation of configurations, significantly reducing the test matrix size Specimens that may be applicable to evaluating limited coverage examinations Insights into limited coverage constraints in a field setting
Contact:
Pradeep Ramuhalli 509-375-2763 pradeep.ramuhalli@pnnl.gov 31