skip to main content

Attention:

The NSF Public Access Repository (PAR) system and access will be unavailable from 11:00 PM ET on Friday, December 13 until 2:00 AM ET on Saturday, December 14 due to maintenance. We apologize for the inconvenience.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


  • NSF Public Access
  • Search Results
  • Efficient Global Sensitivity Analysis of Model-Based Ultrasonic Nondestructive Testing Systems Using Machine Learning and Sobol’ Indices
Title: Efficient Global Sensitivity Analysis of Model-Based Ultrasonic Nondestructive Testing Systems Using Machine Learning and Sobol’ Indices
Abstract The objective of this work is to reduce the cost of performing model-based sensitivity analysis for ultrasonic nondestructive testing systems by replacing the accurate physics-based model with machine learning (ML) algorithms and quickly compute Sobol’ indices. The ML algorithms considered in this work are neural networks (NNs), convolutional NN (CNNs), and deep Gaussian processes (DGPs). The performance of these algorithms is measured by the root mean-squared error on a fixed number of testing points and by the number of high-fidelity samples required to reach a target accuracy. The algorithms are compared on three ultrasonic testing benchmark cases with three uncertainty parameters, namely, spherically void defect under a focused and a planar transducer and spherical-inclusion defect under a focused transducer. The results show that NNs required 35, 100, and 35 samples for the three cases, respectively. CNNs required 35, 100, and 56, respectively, while DGPs required 84, 84, and 56, respectively.  more » « less
Award ID(s):
1846862
PAR ID:
10273113
Author(s) / Creator(s):
;
Date Published:
Journal Name:
Journal of Nondestructive Evaluation, Diagnostics and Prognostics of Engineering Systems
Volume:
4
Issue:
4
ISSN:
2572-3901
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ( , Computational Science – ICCS 2021. ICCS 2021. Lecture Notes in Computer Science)
    Paszynski, M. ; Kranzlmüller, D. ; Krzhizhanovskaya, V.V. ; Dongarra, J.J. ; Sloot, P.M. (Ed.)
    Global sensitivity analysis (GSA) is a method to quantify the effect of the input parameters on outputs of physics-based systems. Performing GSA can be challenging due to the combined effect of the high computational cost of each individual physics-based model, a large number of input parameters, and the need to perform repetitive model evaluations. To reduce this cost, neural networks (NNs) are used to replace the expensive physics-based model in this work. This introduces the additional challenge of finding the minimum number of training data samples required to train the NNs accurately. In this work, a new method is introduced to accurately quantify the GSA values by iterating over both the number of samples required to train the NNs, terminated using an outer-loop sensitivity convergence criteria, and the number of model responses required to calculate the GSA, terminated with an inner-loop sensitivity convergence criteria. The iterative surrogate-based GSA guarantees converged values for the Sobol’ indices and, at the same time, alleviates the specification of arbitrary accuracy metrics for the surrogate model. The proposed method is demonstrated in two cases, namely, an eight-variable borehole function and a three-variable nondestructive testing (NDT) case. For the borehole function, both the first- and total-order Sobol’ indices required 200 and 105 data points to terminate on the outer- and inner-loop sensitivity convergence criteria, respectively. For the NDT case, these values were 100 for both first- and total-order indices for the outer-loop sensitivity convergence, and 106 and 103 for the inner-loop sensitivity convergence, respectively, for the first- and total-order indices, on the inner-loop sensitivity convergence. The differences of the proposed method with GSA on the true functions are less than 3% in the analytical case and less than 10% in the physics-based case (where the large error comes from small Sobol’ indices). 
    more » « less
  2. ; ; ; ( , International Journal for Numerical Methods in Engineering)
    Summary

    We show that deep convolutional neural networks (CNNs) can massively outperform traditional densely connected neural networks (NNs) (both deep or shallow) in predicting eigenvalue problems in mechanics. In this sense, we strike out in a new direction in mechanics computations with strongly predictive NNs whose success depends not only on architectures being deep but also being fundamentally different from the widely used to date. We consider a model problem: predicting the eigenvalues of one‐dimensional (1D) and two‐dimensional (2D) phononic crystals. For the 1D case, the optimal CNN architecture reaches 98% accuracy level on unseen data when trained with just 20 000 samples, compared to 85% accuracy even with 100 000 samples for the typical network of choice in mechanics research. We show that, with relatively high data efficiency, CNNs have the capability to generalize well and automatically learn deep symmetry operations, easily extending to higher dimensions and our 2D case. Most importantly, we show how CNNs can naturally represent mechanical material tensors, with its convolution kernels serving as local receptive fields, which is a natural representation of mechanical response. Strategies proposed are applicable to other mechanics' problems and may, in the future, be used to sidestep cumbersome algorithms with purely data‐driven approaches based upon modern deep architectures.

     
    more » « less
  3. ; ( , Optics Letters)

    Multi-parametric photoacoustic microscopy (PAM) is uniquely capable of simultaneous high-resolution mapping of blood oxygenation and flowin vivo. However, its speed has been limited by the dense sampling required for blood flow quantification. To overcome this limitation, we have developed a high-speed multi-parametric PAM system, which enables simultaneous acquisition of ∼500 densely sampled B-scans by superposing the rapid optical scanning across the line-shaped focus of a cylindrically focused ultrasonic transducer over the conventional mechanical scan of the optical-acoustic dual foci. A novel, to the best of our knowledge, optical-acoustic combiner (OAC) is designed and implemented to accommodate the short working distance of the transducer, enabling convenient confocal alignment of the dual foci in reflection mode. A resonant galvanometer (GM) provides stabilized high-speed large-angle scanning. This new system can continuously monitor microvascular blood oxygenation (sO2) and flow over a 4.5 × 3 mm2area in the awake mouse brain with high spatial and temporal resolutions (6.9 µm and 0.3 Hz, respectively).

     
    more » « less
  4. ; ; ; ; ; ; ; ; ; ; et al ( , ACS ES&T Water)
    To evaluate the use of wastewater-based surveillance and epidemiology to monitor and predict SARS-CoV-2 virus trends, over the 2020–2021 academic year we collected wastewater samples twice weekly from 17 manholes across Virginia Tech’s main campus. We used data from external door swipe card readers and student isolation/quarantine status to estimate building-specific occupancy and COVID-19 case counts at a daily resolution. After analyzing 673 wastewater samples using reverse transcription quantitative polymerase chain reaction (RT-qPCR), we reanalyzed 329 samples from isolation and nonisolation dormitories and the campus sewage outflow using reverse transcription digital droplet polymerase chain reaction (RT-ddPCR). Population-adjusted viral copy means from isolation dormitory wastewater were 48% and 66% higher than unadjusted viral copy means for N and E genes (1846/100 mL to 2733/100 mL/100 people and 2312/100 mL to 3828/100 mL/100 people, respectively; n = 46). Prespecified analyses with random-effects Poisson regression and dormitory/cluster-robust standard errors showed that the detection of N and E genes were associated with increases of 85% and 99% in the likelihood of COVID-19 cases 8 days later (incident–rate ratio (IRR) = 1.845, p = 0.013 and IRR = 1.994, p = 0.007, respectively; n = 215), and one-log increases in swipe card normalized viral copies (copies/100 mL/100 people) for N and E were associated with increases of 21% and 27% in the likelihood of observing COVID-19 cases 8 days following sample collection (IRR = 1.206, p < 0.001, n = 211 for N; IRR = 1.265, p < 0.001, n = 211 for E). One-log increases in swipe normalized copies were also associated with 40% and 43% increases in the likelihood of observing COVID-19 cases 5 days after sample collection (IRR = 1.403, p = 0.002, n = 212 for N; IRR = 1.426, p < 0.001, n = 212 for E). Our findings highlight the use of building-specific occupancy data and add to the evidence for the potential of wastewater-based epidemiology to predict COVID-19 trends at subsewershed scales. 
    more » « less
  5. ; ; ; ; ; ; ; ( , The Analyst)
    The COVID-19 pandemic has highlighted the urgent need for sensitive, affordable, and widely accessible testing at the point of care. Here we demonstrate a new, universal LFA platform technology using M13 phage conjugated with antibodies and HRP enzymes that offers high analytical sensitivity and excellent performance in a complex clinical matrix. We also report its complete integration into a sensitive chemiluminescence-based smartphone-readable lateral flow assay for the detection of SARS-CoV-2 nucleoprotein. We screened 84 anti-nucleoprotein monoclonal antibody pairs in phage LFA and identified an antibody pair that gave an LoD of 25 pg mL −1 nucleoprotein in nasal swab extract using a FluorChem gel documentation system and 100 pg mL −1 when the test was imaged and analyzed by an in-house-developed smartphone reader. The smartphone-read LFA signals for positive clinical samples tested ( N = 15, with known Ct) were statistically different ( p < 0.001) from signals for negative clinical samples ( N = 11). The phage LFA technology combined with smartphone chemiluminescence imaging can enable the timely development of ultrasensitive, affordable point-of-care testing platforms for SARS-CoV-2 and beyond. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email