1
|
Defect Shape Classification Using Transfer Learning in Deep Convolutional Neural Network on Magneto-Optical Nondestructive Inspection. APPLIED SCIENCES-BASEL 2022. [DOI: 10.3390/app12157613] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 02/01/2023]
Abstract
To implement a magneto-optic (MO) nondestructive inspection (MONDI) system for robot-based nondestructive inspections, quantitative evaluations of the presence, locations, shapes, and sizes of defects are required. This capability is essential for training autonomous nondestructive testing (NDT) devices to track material defects and evaluate their severity. This study aimed to support robotic assessment using the MONDI system by providing a deep learning algorithm to classify defect shapes from MO images. A dataset from 11 specimens with 72 magnetizer directions and 6 current variations was examined. A total of 4752 phenomena were captured using an MO sensor with a 0.6 mT magnetic field saturation and a 2 MP CMOS camera as the imager. A transfer learning method for a deep convolutional neural network (CNN) was adapted to classify defect shapes using five pretrained architectures. A multiclassifier technique using an ensemble and majority voting model was also trained to provide predictions for comparison. The ensemble model achieves the highest testing accuracy of 98.21% with an area under the curve (AUC) of 99.08% and a weighted F1 score of 0.982. The defect extraction dataset also indicates auspicious results by increasing the training time by up to 21%, which is beneficial for actual industrial inspections when considering fast and complex engineering systems.
Collapse
|
2
|
Distribution of Magnetic Flux Density under Stress and Its Application in Nondestructive Testing. APPLIED SCIENCES-BASEL 2022. [DOI: 10.3390/app12157612] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 02/01/2023]
Abstract
Carbon steels are commonly used in railroad, shipment, building, and bridge construction. They provide excellent ductility and toughness when exposed to external stresses. They are able to resist stresses and strains effectively, and guarantee safe operation of the devices through nondestructive testing (NDT). The magnetic metal memory (MMM) can be used as an NDT method to measure the residual stress. The ability of carbon steel to produce a magnetic memory effect under stress is explored here, and enables the magnetic flux density to be analyzed. The relationship between stress and magnetic flux density has not been fully presented until now. The purpose of this paper is to assess the relationship between stress distribution and the magnetic flux density measured by the experiment. For this, an experimental method for examining a carbon steel plate (SA 106), based on the four-point loading test, was used. The effect of stresses resulting from the applied loads on the response of the experimented SA 106 specimen was examined. A three directional tunnel magnetoresistance (TMR) measurement system was used to collect the triaxial magnetic flux density distribution in the SA 106 specimen. In addition, finite element method (FEM) analyses were performed, and provided information on the direction and distribution of the stress over the studied SA 106 specimen. Indeed, a correlation was derived by comparing the stress analysis by FEM and the measured triaxial magnetic flux density.
Collapse
|
3
|
Special Issue: Novel Approaches for Nondestructive Testing and Evaluation. APPLIED SCIENCES-BASEL 2022. [DOI: 10.3390/app12020565] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 11/16/2022]
Abstract
Nondestructive testing and evaluation (NDT&E) is one of the most important techniques for determining the quality and safety of materials, components, devices, and structures [...]
Collapse
|