Lecture notes in computer science, Journal Year: 2023, Volume and Issue: unknown, P. 133 - 147
Published: Jan. 1, 2023
Language: Английский
Materials Today Communications, Journal Year: 2022, Volume and Issue: 33, P. 104900 - 104900
Published: Nov. 9, 2022
Language: Английский
Citations
84
IEEE Transactions on Components Packaging and Manufacturing Technology, Journal Year: 2022, Volume and Issue: 12(3), P. 489 - 501
Published: Jan. 27, 2022
In the mass production of electronic products, in-circuit-test (ICT) and printed circuit board assembly (PCBA) quality tests are performed. ICT measures resistance values capacitance, but not only does it require use a fixture that is expensive requires frequent replacement also fixture's needles may cause PCBA defects. To overcome these limitations, various studies tried to replace using visual inspection methods; however, methods cannot be applied chip resistors capacitors do have externally visible characteristics. this article, we propose contactless method can detect defects without by comparison thermal images deep learning (DL) analysis. We review existing compare them with our proposed image analysis method. analyzed applying structural similarity index map as rule-based object detection method, used convolutional neural networks (CNNs), regions CNN features, an autoencoder DL methods. As result, achieved highly accurate defective component location in real time.
Language: Английский
Citations
44
Journal of the Mechanics and Physics of Solids, Journal Year: 2023, Volume and Issue: 173, P. 105231 - 105231
Published: Jan. 31, 2023
Language: Английский
Citations
43
Engineering Failure Analysis, Journal Year: 2023, Volume and Issue: 153, P. 107587 - 107587
Published: Sept. 3, 2023
Literature sources have been reviewed, various techniques, methods and software for investigating fracture of polycrystalline materials subjected to the systematic analysis, dimples ductile separation quantified. Regularities in description patterns inherent at macro-, meso- micro-levels addressed, their features parameters scrutinized. Algorithms recognizing calculating images obtained by electron microscopy analyzed terms functionality ability ensure highly informative content reliability fractographic control.
Language: Английский
Citations
32
Colloids and Interface Science Communications, Journal Year: 2022, Volume and Issue: 47, P. 100595 - 100595
Published: Feb. 17, 2022
Colloidal material design necessitates a collection of computer approaches ranging from quantum chemistry to molecular dynamics and continuum modeling. Machine learning (ML) other umbrella terminology for current optimization (requiring computation) have accelerated the predictability characteristics. materials include polymers, liquid crystals, colloids. Supervised unsupervised strategies come under scrutiny in this review. Other ways, such as combined simulation ML modeling procedures, are also available that not through present arsenal characterization tools. Such hybrid can improve our understanding protocols. In review, we accumulated expertise information over 300 sources.
Language: Английский
Citations
36
Energies, Journal Year: 2022, Volume and Issue: 15(5), P. 1845 - 1845
Published: March 2, 2022
Nuclear energy is presently the single major low-carbon electricity source in Europe and overall expected to maintain (perhaps eventually even increase) its current installed power from now 2045. Long-term operation (LTO) a reality essentially all nuclear European countries, when planning phase out. New builds are planned. Moreover, several including non-nuclear or phasing out ones, have interests next generation systems. In this framework, materials material science play crucial role towards safer, more efficient, economical sustainable energy. This paper proposes research agenda that combines modern digital technologies with practices pursue change of paradigm promotes innovation, equally serving different positions throughout Europe. chooses overview structural fuel used reactors, as well their wider spectrum for summarising relevant issues. Next, it describes approaches common any (including classes not addressed here, such concrete, polymers functional materials), identifying each them goal. It concluded among these goals development structured qualification test-beds acceleration platforms (MAPs) operate under harsh conditions. Another goal multi-parameter-based health monitoring based on non-destructive examination testing (NDE&T) techniques. Hybrid models suitably combine physics-based data-driven behaviour prediction can valuably support developments, together creation population centralised, “smart” database materials.
Language: Английский
Citations
34
Computer Methods in Applied Mechanics and Engineering, Journal Year: 2022, Volume and Issue: 395, P. 115021 - 115021
Published: May 1, 2022
Language: Английский
Citations
32
Mathematics, Journal Year: 2022, Volume and Issue: 10(16), P. 3019 - 3019
Published: Aug. 22, 2022
Deep learning has been widely used in different fields such as computer vision and speech processing. The performance of deep algorithms is greatly affected by their hyperparameters. For complex machine models neural networks, it difficult to determine In addition, existing hyperparameter optimization easily converge a local optimal solution. This paper proposes method for that combines the Sparrow Search Algorithm Particle Swarm Optimization, called Hybrid Algorithm. takes advantages avoiding solution search efficiency Optimization achieve global optimization. Experiments verified proposed algorithm simple networks. results show strong capability avoid solutions satisfactory both low high-dimensional spaces. provides new problems models.
Language: Английский
Citations
30
ACS Omega, Journal Year: 2024, Volume and Issue: 9(23), P. 24695 - 24706
Published: May 10, 2024
The morphological characteristics of tissue engineering scaffolds, such as pore and window diameters, are crucial, they directly impact cell-material interactions, attachment, spreading, infiltration the cells, degradation rate mechanical properties scaffolds. Scanning electron microscopy (SEM) is one most commonly used techniques for characterizing microarchitecture scaffolds due to its advantages, being easily accessible having a short examination time. However, SEM images provide qualitative data that need be manually measured using software ImageJ quantify features As it not practical measure each pore/window in requires extensive time effort, only number pores/windows assumed represent whole sample, which may cause user bias. Additionally, depending on samples groups, study require measuring thousands human error increase. To overcome problems, this study, deep learning model (Pore D2) was developed (such size size) open-porous automatically first algorithm tested emulsion-templated fabricated under different fabrication conditions, changing mixing speed, temperature, surfactant concentration, resulted with various morphologies. Along model, blind manual measurements were taken, results showed tool capable quantifying sizes high accuracy. Quantifying circumstances controlling these enable us engineer precisely specific applications. Pore D2, an open-source software, available everyone at following link: https://github.com/ilaydakaraca/PoreD2.
Language: Английский
Citations
7
Materials Today Communications, Journal Year: 2024, Volume and Issue: 39, P. 108710 - 108710
Published: March 23, 2024
Language: Английский
Citations
6