An Efficient Algorithm for Finding the Level of Useful Signals on Interpretation of Magnetic and Eddy Current Defectograms

Gorbunov²,

et al. 2021

To ensure traffic safety of railway transport, non-destructive tests of rails are regularly carried out by using various approaches and methods, including eddy-current flaw detection methods. An automatic analysis of large data sets (defectograms) that come from the corresponding equipment is an actual problem. The analysis means a process of determining the presence of defective sections along with identifying structural elements of railway tracks in defectograms. This article continues the cycle of works devoted to the problem of automatic recognizing images of defects and structural elements of rails in eddy-current defectograms. In the process of forming these images, only useful signals are taken into account, the threshold levels of amplitudes of which are determined automatically from eddy-current data. The previously used algorithm for finding threshold levels was focused on situations in which the vast majority of signals coming from the flaw detector is a rail noise. A signal is considered useful and is subject to further analysis if its amplitude is twice the corresponding noise threshold. The article is devoted to the problem of correcting threshold levels, taking into account the need to identify extensive surface defects of rails. An algorithm is proposed for finding the values of threshold levels of rail noise amplitudes with their subsequent correction in the case of a large number of useful signals from extensive defects. Examples of the algorithm’s operation on real eddy-current data are given.

Section: нахождение пороговых уровней полезных сигналовunclassified

Section: Introductionunclassified

See 2 more Smart Citations

An Algorithm for Correcting Levels of Useful Signals on Interpretation of Eddy-Current Defectograms

Gorbunov²,

et al. 2021

“…Данная статья продолжает цикл работ [4][5][6][7][8], посвященных задаче автоматического распознавания образов дефектов и конструктивных элементов железнодорожных рельсов по дефектограммам многоканальных вихретоковых дефектоскопов. Дефектограммы разбиваются на фрагменты (блоки анализа), каждый из которых обрабатывается отдельно.…”

Section: Introductionunclassified

Severity Estimation of Defects on Interpretation of Eddy-Current Defectograms

Gorbunov²,

et al. 2021

To ensure traffic safety of railway transport, non-destructive tests of rails are regularly carried out by using various approaches and methods, including eddy-current flaw detection methods. An automatic analysis of large data sets (defectograms) that come from the corresponding equipment is an actual problem. The analysis means a process of determining the presence of defective sections along with identifying structural elements of railway tracks in defectograms. At the same time, severity estimation of defined defects is also of great interest. This article continues the cycle of works devoted to the problem of automatic recognition of images of defects and rail structural elements in eddy-current defectograms. In the process of forming these images, only useful signals are taken into account, the threshold levels of amplitudes of which are determined automatically from eddy-current data. The article is devoted to the issue of constructing severity estimation of found defects with various lengths. The construction of the severity estimation is based on a concept of the generalized relative amplitude of useful signals. A relative amplitude is a ratio of an actual signal amplitude to a corresponding threshold level of useful signals. The generalized relative amplitude is calculated by using the entropy of the half-normal distribution, which is assumed to be a model for a probability distribution of an appearance of certain relative amplitudes in an evaluated defect. Tuning up the formula for calculating severity estimation of a defect is carried out on the basis of eddy-current records of structural elements. As a reference of the most dangerous defect, the bolted rail joint is considered. It models a fracture of a rail. A reference weak defect is a flash butt weld, a defectogram of which contains signals with low amplitude values. The proposed approach to severity estimation of defects is shown by examples.

Application of Convolutional Neural Networks for Recognizing Long Structural Elements of Rails in Eddy-Current Defectograms

Gorbunov²,

et al. 2020

To ensure tra c safety of railway transport, non-destructive test of rails is regularly carried out by using various approaches and methods, including eddy-current aw detection methods. An automatic analysis of large data sets (defectograms) that come from the corresponding equipment is an actual problem. e analysis means a process of determining the presence of defective sections along with identifying structural elements of railway tracks in defectograms. is article is devoted to the problem of recognizing images of long structural elements of rails in eddy-current defectograms. Two classes of rail track structural elements are considered: 1) rolling stock axle counters, 2) rail crossings. Long marks that cannot be assigned to these two classes are conditionally considered as defects and are placed in a separate third class. For image recognition of structural elements in defectograms a convolutional neural network is applied. e neural network is implemented by using the open library TensorFlow. To this purpose each selected (picked out) area of a defectogram is converted into a graphic image in a grayscale with size of 30 x 140 points.