Nondestructive Testing of Structurally Inhomogeneous Composite Materials by the Method of Elastic-Wave Velocity Hodograph

Potapov, A. I.; Kondrat’ev, A. V.; Smorodinskii, Ya. G.

doi:10.1134/s106183091906007x

Cited by 9 publications

(5 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In most cases, developers of industrial automation systems do not have at their disposal the necessary amount of production data sufficient to implement machine learning algorithms. This is due to the company not recording the necessary parameters beforehand nor doing it properly; automatic markup of production data being difficult, and manual markup requiring a high level of specialist competence; data collection must be performed in long time intervals (months and years) [4,5]. As a consequence, these limitations, taken together, significantly complicate the implementation of machine learning algorithms in automated control systems for technological processes [6].…”

Section: Introductionmentioning

confidence: 99%

Synthetic Data Generation for Steel Defect Detection and Classification Using Deep Learning

et al. 2021

View full text Add to dashboard Cite

The paper presents a methodology for training neural networks for vision tasks on synthesized data on the example of steel defect recognition in automated production control systems. The article describes the process of dataset procedural generation of steel slab defects with a symmetrical distribution. The results of training two neural networks Unet and Xception on a generated data grid and testing them on real data are presented. The performance of these neural networks was assessed using real data from the Severstal: Steel Defect Detection set. In both cases, the neural networks showed good results in the classification and segmentation of surface defects of steel workpieces in the image. Dice score on synthetic data reaches 0.62, and accuracy—0.81.

show abstract

Section: Introductionmentioning

confidence: 99%

Synthetic Data Generation for Steel Defect Detection and Classification Using Deep Learning

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Some of these singlephase arrangements may be those that are unstable on mechanical deformation, thermal processing and other factors [18][19][20][21][22]. Ultrasonic attenuation (UA) is the main physical parameters to describe a compound, which appreciates the specific relationship between the anisotropic behaviour of the adjacent atomic planes and affinity and the structural motion, some physical quantities such as specific heat, thermal energy density and thermal conductivity is well associated with higherorder elastic constants [23,24].…”

Section: Introductionmentioning

confidence: 99%

Elastic, Mechanical and Thermophysical properties of Single-Phase Quaternary ScTiZrHf High-Entropy Alloy

Rai

Chaurasiya

Yadawa

2021

Phys. Chem. Solid St.

View full text Add to dashboard Cite

Consequent to the interaction potential model, the high-order elastic constants at high entropy alloys in single-phase quaternary ScTiZrHf have been calculated at different temperatures. Elastic constants of second order (SOECs) helps to determine other ultrasonic parameters. With the help of SOECs other elastic moduli, bulk modulus, shear modulus, Young’s modulus, Pugh’s ratio, elastic stiffness constants and Poisson’s ratio are estimated at room temperature for elastic and mechanical characterization. The other ultrasonic parameters are calculated at room temperature for elastic and mechanical characterization. The temperature variation of ultrasonic velocities along the crystal's z-axis is evaluated using SOECs. The temperature variation of the average debye velocity and the thermal relaxation time (τ) are also estimated along this orientation axis. The ultrasonic properties correlated with elastic, thermal and mechanical properties which is temperature dependent is also discussed. The ultrasonic attenuation due to phonon – phonon (p-p) interactions is also calculated at different temperatures. In the study of ultrasonic attenuation such as a function of temperature, thermal conductivity appears to be main contributor and p- p interactions are the responsible reason of attenuation and found that the mechanical properties of the high entropy alloy ScTiZrHf are superior at room temperature.

show abstract

“…It also takes a lot of time and effort to manually annotate the acquired data 3 and doing so frequently necessitates competence in spotting image flaws and requires data collection under long-time intervals. 4,5 Hence, the creation of comprehensive datasets is often challenging. As a result, these limitations taken together complicate establishing a supervised defect detection method.…”

Section: Introductionmentioning

confidence: 99%

“…This makes them less useful as training data. It also takes a lot of time and effort to manually annotate the acquired data 3 and doing so frequently necessitates competence in spotting image flaws and requires data collection under long-time intervals 4 , 5 . Hence, the creation of comprehensive datasets is often challenging.…”

Section: Introductionmentioning

confidence: 99%

Enhancing titanium spacer defect detection through reinforcement learning-optimized digital twin and synthetic data generation

Mohanty,

Su,

2024

J. Electron. Imag.

View full text Add to dashboard Cite

In the field of automatic defect detection, a major challenge in training accurate classifiers using supervised learning is the insufficient and limited diversity of datasets. Obtaining an adequate amount of image data depicting defective surfaces in an industrial setting can be costly and time-consuming. Furthermore, the collected dataset may suffer from selection bias, resulting in an underrepresentation of certain defect classes. Our research aims to address surface defect detection in titanium metal spacer rings by introducing an approach leveraging a digital twin framework. The behavior of the digital twin is optimized using a reinforcement learning (RL) algorithm. The optimized digital twin is then used to generate synthetic data, which is employed to train a spacer defect detection classifier. The classifier's performance is evaluated using real-world data. The results show that the model trained with synthetic data outperforms the one trained on a limited amount of real data. Our work highlights the potential of digital twin-based synthetic data generation and RL optimization in enhancing spacer surface defect detection and addressing the data scarcity challenge in the field. When the inspection network is trained solely using generated synthetic data, it achieves an inspection precision of 96.10%, with a recall of 85.77%. Incorporating real data alongside synthetic data for training further enhances performance, yielding an inspection precision of 93.07% and a recall of 94.20%. This surpasses the defect detection capabilities observed when training the inspection network exclusively with real data.

show abstract

Nondestructive Testing of Structurally Inhomogeneous Composite Materials by the Method of Elastic-Wave Velocity Hodograph

Cited by 9 publications

References 11 publications

Synthetic Data Generation for Steel Defect Detection and Classification Using Deep Learning

Synthetic Data Generation for Steel Defect Detection and Classification Using Deep Learning

Elastic, Mechanical and Thermophysical properties of Single-Phase Quaternary ScTiZrHf High-Entropy Alloy

Enhancing titanium spacer defect detection through reinforcement learning-optimized digital twin and synthetic data generation

Contact Info

Product

Resources

About