Efficient Model-Assisted Probability of Detection and Sensitivity Analysis for Ultrasonic Testing Simulations Using Stochastic Metamodeling

Leifsson, Leifur; Meeker, William Q.; Gurrala, Praveen; Song, Jiming; Roberts, R. A.

doi:10.1115/1.4044446

Cited by 11 publications

(5 citation statements)

References 86 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This serves as the baseline to compare the metamodeling results. In the case of the PCE metamodel, its coefficients can be used to provide the 1 st and total order Sobol' indices [27]. For the remaining metamodels, 75,000 MCS points were used to provide satisfactory results for the SA for each of the cases.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Model-Based Sensitivity Analysis of Nondestructive Testing Systems Using Machine Learning Algorithms

Nagawkar

Leifsson

Miorelli

et al. 2020

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Model-based sensitivity analysis is crucial in quantifying which input variability parameter is important for nondestructive testing (NDT) systems. In this work, neural networks (NN) and convolutional NN (CNN) are shown to be computationally efficient at making model prediction for NDT systems, when compared to models such as polynomial chaos expansions, Kriging and polynomial chaos Kriging (PC-Kriging). Three different ultrasonic benchmark cases are considered. NN outperform these three models for all the cases, while CNN outperformed these three models for two of the three cases. For the third case, it performed as well as PC-Kriging. NN required 48, 56 and 35 high-fidelity model evaluations, respectively, for the three cases to reach within 1% accuracy of the physics model. CNN required 35, 56 and 56 high-fidelity model evaluations, respectively, for the same three cases.

show abstract

Section: Resultsmentioning

confidence: 99%

“…The fused quartz block has a density of 2,000 kg/m 3 , a longitudinal wave speed of 5,969.4 m/s and a shear wave speed of 3,774.1 m/s. For more information about the models, refer to Du et al [27].…”

Section: Problem Setupmentioning

confidence: 99%

Model-Based Sensitivity Analysis of Nondestructive Testing Systems Using Machine Learning Algorithms

Nagawkar

Leifsson

Miorelli

et al. 2020

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…The true labels are obtained by solved the two physical equations in Eq. (10). Note that in real applications, the true labels are not available.…”

Section: Table 1 Distributions Associated With Input Random Variablesmentioning

confidence: 99%

“…As indicated by many studies, including our own, machine learning is particularly useful for UQ problems encountered in engineering design. Machine learning techniques can significantly reduce the computational burden in the following aspects: Reduce the dimension of uncertain input variables [9] and create accurate but inexpensive surrogate models [10][11][12][13], both for higher computational efficiency. Specific machine learning techniques have increasingly used.…”

Section: Introductionmentioning

confidence: 99%

Label Free Uncertainty Quantification

Yin

2022

AIAA SCITECH 2022 Forum

View full text Add to dashboard Cite

Uncertainty quantification (UQ) is essential in scientific computation since it can provide the estimate of the uncertainty in the model prediction. Intensive computation is required for UQ as it calls the deterministic simulation repeatedly. This study discusses a physics-based label-free deep learning UQ method that does not need predictions at training points or labels. It satisfies the physical equations from which labels could be generated without solving the equations during the training process. Then inexpensive surrogate models are built with respect to model inputs. The surrogate models are used for UQ with a much lower computational cost. Two examples demonstrate that the label-free method can efficiently produce probability distributions of model outputs for given distributions of random input variables.

show abstract

“…Surrogate modeling methods are useful for reducing the computational cost in various problems involving optimum design [1,2], uncertainty quantification [3,4], and global sensitivity analysis [5,6]. In these methods, a computationally efficient surrogate model replaces an expensive physics-based model, reducing the overall cost involved in solving such problems.…”

Section: Introductionmentioning

confidence: 99%

Gradient-enhanced multifidelity neural networks for high-dimensional function approximation

Nagawkar¹,

Leifsson²

2021

Preprint

Self Cite

View full text Add to dashboard Cite

In this work, a novel multifidelity machine learning (ML) model, the gradient-enhanced multifidelity neural networks (GEMFNNs), is proposed. This model is a multifidelity version of gradient-enhanced neural networks (GENNs) as it uses both function and gradient information available at multiple levels of fidelity to make function approximations. Its construction is similar to multifidelity neural networks (MFNNs). This model is tested on three analytical function, a one, two, and a 20 variable function. It is also compared to neural networks (NNs), GENNs, and MFNNs, and the number of samples required to reach a global accuracy of 0.99 coefficient of determination (R 2 ) is measured. GEMFNNs required 18, 120, and 600 high-fidelity samples for the one, two, and 20 dimensional cases, respectively, to meet the target accuracy. NNs performed best on the one variable case, requiring only ten samples, while GENNs worked best on the two variable case, requiring 120 samples. GEMFNNs worked best for the 20 variable case, while requiring nearly eight times fewer samples than its nearest competitor, GENNs. For this case, NNs and MFNNs did not reach the target global accuracy even after using 10,000 high-fidelity samples. This work demonstrates the benefits of using gradient as well as multifidelity information in NNs for high-dimensional problems.

show abstract

Efficient Model-Assisted Probability of Detection and Sensitivity Analysis for Ultrasonic Testing Simulations Using Stochastic Metamodeling

Cited by 11 publications

References 86 publications

Model-Based Sensitivity Analysis of Nondestructive Testing Systems Using Machine Learning Algorithms

Model-Based Sensitivity Analysis of Nondestructive Testing Systems Using Machine Learning Algorithms

Label Free Uncertainty Quantification

Gradient-enhanced multifidelity neural networks for high-dimensional function approximation

Contact Info

Product

Resources

About