Computational Tools Inc.

Chatham, IL, United States

Computational Tools Inc.

Chatham, IL, United States
Time filter
Source Type

Aldrin J.C.,Computational Tools Inc. | Forsyth D.S.,TRI Austin | Welter J.T.,Air Force Research Lab
Proceedings of the American Society for Composites - 31st Technical Conference, ASC 2016 | Year: 2016

To address the data review burden and improve the reliability of the ultrasonic (UT) inspection of large composite structures, automated data analysis (ADA) algorithms have been developed to make calls on indications that satisfy the call criteria. Certain complex composite structures with varying shape, thickness transitions, front-wall and back-wall curvature, and the presence of bonds can greatly complicate this interpretation process and produce false calls. In this effort, enhancements to the automated data analysis algorithms are introduced to address these challenges for complex composite panels. First, the thickness of the part and presence of bonds is estimated by tracking potential backwall signals, detecting the presence of multiple signals and step changes which are indicators of bonded sections, and through the application of smart spatial filters for estimating the panel thickness and additional bonded sections with varying signal levels. Once part boundaries, thickness transitions and bonded regions are identified, feature extraction algorithms are applied to multiple sets of through-thickness and backwall C-scan images, for evaluation of both first layer through thickness and layers under bonds. To verify the algorithm performance, a set of test data was selected to challenge the ADA algorithms that includes a wide range of complex parts and artificial defects located both above and below bond lines. Results are presented for a variety of test specimens that include inserted materials and discontinuities produced under poor manufacturing conditions, demonstrating the desired detection capability while minimizing false indication calls to a manageable level.

McMahan J.A.,University of Dayton | Aldrin J.C.,Computational Tools Inc. | Shell E.,Wyle | Oneida E.,Wyle
AIP Conference Proceedings | Year: 2017

The Bayesian approach to inference from measurement data has the potential to provide highly reliable characterizations of flaw geometry by quantifying the confidence in the estimate results. The accuracy of these confidence estimates depends on the accuracy of the model for the measurement error. Eddy current measurements of electrically anisotropic metals, such as titanium, exhibit a phenomenon called grain noise in which the measurement error is spatially correlated even with no flaw present. We show that the most commonly used statistical model for the measurement error, which fails to account for this correlation, results in overconfidence in the flaw geometry estimates from eddy current data, thereby reducing the effectiveness of the Bayesian approach. We then describe a method of modeling the grain noise as a Gaussian process (GP) using spectral mixture kernels, a type of non-parametric model for the covariance kernel of a GP This provides a broadly applicable, data-driven way of modeling correlation in measurement error. Our results show that incorporation of this noise model results in a more reliable estimate of the flaw and better agreement with the available validation data. © 2017 Author(s).

Wertz J.,Air Force Research Lab | Wallentine S.,Air Force Research Lab | Welter J.,Air Force Research Lab | Dierken J.,University of Dayton | Aldrin J.,Computational Tools Inc.
AIP Conference Proceedings | Year: 2017

The volumetric characterization of delaminations necessarily precedes rigorous composite damage progression modeling. Yet, inspection of composite structures for subsurface damage remains largely focused on detection, resulting in a capability gap. In response to this need, angle longitudinal wave ultrasound was employed to characterize a composite surrogate containing a simulated three-dimensional delamination field with distinct regions of occluded features (shadow regions). Simple analytical models of the specimen were developed to guide subsequent experimentation through identification of optimal scanning parameters. The ensuing experiments provided visual evidence of the complete delamination field, including indications of features within the shadow regions. The results of this study demonstrate proof-of-principle for the use of angle longitudinal wave ultrasonic inspection for volumetric characterization of three-dimensional delamination fields. Furthermore, the techniques developed herein form the foundation of succeeding efforts to characterize impact delaminations within inhomogeneous laminar materials such as polymer matrix composites. © 2017 U.S. Government.

Oneida E.K.,KBRwyle | Shell E.B.,KBRwyle | Aldrin J.C.,Computational Tools Inc. | Sabbagh H.A.,Victor Technologies, LLC | And 2 more authors.
Materials Evaluation | Year: 2017

Initial results for model-based inversion of eddy current flaw response data are presented as an alternative or supplement to amplitude-based sizing. The approach provides an estimation of the flaw dimensions (length, depth, and width) and orientation that are independent of amplitudes and assumptions from a probability of detection (POD) analysis. The effects of probe construction characteristics on eddy current signals are considered. The inversion method performs well over a range of intermediate flaw sizes and orientations, representing an improvement for flaws that are dissimilar from those used in an amplitude-based POD model. Future research directions to improve inversion estimates associated with greater uncertainty are discussed.

Aldrin J.C.,Computational Tools Inc. | Coughlin C.,TRI Austin | Forsyth D.S.,TRI Austin | Welter J.T.,Air Force Research Lab
AIP Conference Proceedings | Year: 2014

Progress is presented on the development and implementation of automated data analysis (ADA) software to address the burden in interpreting ultrasonic inspection data for large composite structures. The automated data analysis algorithm is presented in detail, which follows standard procedures for analyzing signals for time-of-flight indications and backwall amplitude dropout. ADA processing results are presented for test specimens that include inserted materials and discontinuities produced under poor manufacturing conditions. © 2014 AIP Publishing LLC.

Aldrin J.C.,Computational Tools Inc. | Annis C.,Statistical Engineering | Sabbagh H.A.,Victor Technologies, LLC | Knopp J.S.,Air Force Research Lab | Lindgren E.A.,Air Force Research Lab
AIP Conference Proceedings | Year: 2014

A comprehensive approach to NDE characterization error evaluation is presented that follows the framework of the 'ahat-versus-a' model evaluation process for probability of detection (POD) assessment. Before characterization error model building is performed, an intermediate step must evaluate the presence and frequency of several possible classes of poor characterization results. A case study is introduced based on the estimation the length, depth and width of surface breaking cracks using bolt hole eddy current (BHEC) NDE. This study highlights the importance of engineering and statistical expertise in the model-building process to ensure all key effects and possible interactions are addressed. © 2014 AIP Publishing LLC.

Knopp J.S.,Air Force Research Lab | Grandhi R.,Wright State University | Aldrin J.C.,Computational Tools Inc. | Park I.,Wright State University
Journal of Nondestructive Evaluation | Year: 2013

Recent work on reliably detecting and characterizing cracks in multi-layer airframe structures has used modeling and simulation to extract features from raw eddy current data, and to assist in the evaluation of probability of detection (POD). This paper focuses on the statistical analysis of the data from these studies. Hit/miss, linear, and physics-inspired methods are employed to evaluate POD. The Box-Cox transformation is used as a remedy for violations of the constant variance assumption. In addition, a bootstrapping method is introduced for confidence bound calculation on a 2nd order linear model. The objective of this work is to provide on insight how different models and assumptions impact POD evaluation. © 2012 Springer Science+Business Media, LLC.

Aldrin J.C.,Computational Tools Inc. | Blodgett M.P.,Air Force Research Lab | Lindgren E.A.,Air Force Research Lab | Steffes G.J.,Air Force Research Lab | Knopp J.S.,Air Force Research Lab
Journal of the Acoustical Society of America | Year: 2011

Prior work has proposed the use of ultrasonic angle-beam shear wave techniques to detect cracks of varying angular location around fastener sites by generating and detecting creeping waves. To better understand the nature of the scattering problem and quantify the role of creeping waves in fastener site inspections, a 3D analytical model was developed for the propagation and scattering of an obliquely incident plane shear wave from a cylindrical cavity with arbitrary shear wave polarization. The generation and decay of the spiral creeping waves was found to be dependent on both the angle of incidence and polarization of the plane shear wave. A difference between the angle of displacement in 3D and the direction of propagation for the spiral creeping wave was observed and attributed to differences in the curvature of the cavity surface for the tangential and vertical (z) directions. Using the model, practical insight was presented on measuring the displacement response in the far-field from the hole. Both analytical and experimental results highlighted the value of the diffracted and leaky spiral creeping wave signals for nondestructive evaluation of a crack located on the cavity. Last, array and signal processing methods are discussed to improve the resolution of the weaker creeping wave signals in the presence of noise. © 2011 Acoustical Society of America.

Forsyth D.S.,TRI Austin | Aldrin J.C.,Computational Tools Inc. | Welter J.T.,U.S. Air force
International SAMPE Technical Conference | Year: 2013

Automated inspection systems are routinely deployed in manufacturing processes. Total inspection coverage is often required for flight safety-critical components. A long history of nondestructive testing (NDT) of metal components has given the industry confidence in these inspection systems. There is less corresponding history and confidence for fiber-reinforced polymer composites in primary structures, and the inspection burden is more significant. There are potentially multiple defect criteria, and the criteria will vary from part to part. The result is that highly trained, experienced, and relatively expensive NDT personnel are needed to review the enormous amounts of data collected by automated inspection systems. In this presentation, we describe an ongoing effort to automate the interpretation of ultrasonic inspection of composite structures. We consider defect types and acceptance criteria, design of Automated Defect Analysis (ADA) algorithms, validation, and the estimation of the return on investment obtained by implementing an on-line ADA software into the ultrasonic inspection system. Copyright 2013 by Aurora Flight Sciences.

Annis C.,Statistical Engineering | Aldrin J.C.,Computational Tools Inc. | Sabbagh H.A.,Victor Technologies, LLC
Materials Evaluation | Year: 2015

The statistical techniques in MIL-HDBK-1823A work for the large majority of situations, but not all situations. When the model (any mathematical model, not just a probability of detection [POD] model) does not have the same characteristics as the data, then it is the wrong model. Two, not uncommon, examples are a POD minimum value that never approaches zero, perhaps due to background noise, or a maximum POD that does not approach one, possibly a result of an occluded inspection site, while the MIL-HDBK-1823A requires asymptotes at both zero and one. The authors discuss a statistical procedure for adding an asymptote to the standard POD model, estimating its most likely value, and computing confidence bounds on that estimate. © 2015, American Society for Nondestructive Testing. All rights reserved.

Loading Computational Tools Inc. collaborators
Loading Computational Tools Inc. collaborators