Defects Identification and Crack Depth Determination in Porous Media on the Brick Masonry Example Using Ultrasonic Methods: Numerical Analysis and Machine Learning DOI Open Access
Alexey N. Beskopylny, Sergey A. Stel’makh, Evgenii M. Shcherban’

и другие.

Journal of Composites Science, Год журнала: 2025, Номер 9(6), С. 267 - 267

Опубликована: Май 28, 2025

Automation of the structural health monitoring process involves use successful methods for detecting defects and determining their critical characteristics. An efficient means crack detection in composite materials is ultrasonic method, but its application to determine parameters, such as depth construction practice, difficult or leads large errors. This article focuses on machine learning usage detect cracks like brickwork. Ceramic bricks with various mechanical properties pre-grown from 2 60 mm are considered. To understand processes occurring during pulse transmission, modeling was performed ANSYS environment. The brick considered a porous medium weakened by crack. Numerical allows identification main features signal response determination amplitude-time range different porosity values. Using made it possible solve two related problems. first, binary classification, i.e., presence absence crack, solved 100% accuracy. second depth. A neural network built using an ensemble decision trees. accuracy prediction R2 = 0.983, error predicted values within 8%.

Язык: Английский

Unidirectional Frequency-Steerable Acoustic Transducer for guided ultrasonic wave damage imaging DOI Creative Commons
Masoud Mohammadgholiha, Jochen Moll,

Kilian Tschöke

и другие.

Mechanical Systems and Signal Processing, Год журнала: 2025, Номер 229, С. 112505 - 112505

Опубликована: Март 7, 2025

Язык: Английский

Процитировано

0
Defects Identification and Crack Depth Determination in Porous Media on the Brick Masonry Example Using Ultrasonic Methods: Numerical Analysis and Machine Learning DOI Open Access
Alexey N. Beskopylny, Sergey A. Stel’makh, Evgenii M. Shcherban’

и другие.

Journal of Composites Science, Год журнала: 2025, Номер 9(6), С. 267 - 267

Опубликована: Май 28, 2025

Automation of the structural health monitoring process involves use successful methods for detecting defects and determining their critical characteristics. An efficient means crack detection in composite materials is ultrasonic method, but its application to determine parameters, such as depth construction practice, difficult or leads large errors. This article focuses on machine learning usage detect cracks like brickwork. Ceramic bricks with various mechanical properties pre-grown from 2 60 mm are considered. To understand processes occurring during pulse transmission, modeling was performed ANSYS environment. The brick considered a porous medium weakened by crack. Numerical allows identification main features signal response determination amplitude-time range different porosity values. Using made it possible solve two related problems. first, binary classification, i.e., presence absence crack, solved 100% accuracy. second depth. A neural network built using an ensemble decision trees. accuracy prediction R2 = 0.983, error predicted values within 8%.

Язык: Английский

Процитировано

0