Natural Rail Surface Defect Inspection and Analysis Using 16-Channel Eddy Current System
Round 1
Reviewer 1 Report
applsci-1297886 Resubmitted from applsci-1222460 is acceptable for Publishing in the presented form.
Author Response
Comment 1:
Added references.
Comment 2:
1) There is no special equation, and the calibration method was written in the paper.
2) y-axis alignment complete
3) Both channels 13 & 15 have a signal. The signal level is less than other sensors.
4) Other channels are not noise. channel 4 & 8 have a lot of noise.
Comment 3:
1) Edit abstract and introduction
2) fig. 3 & 4 Delete
3) fig. 7 uses the specimen in fig. 5
4) fig. 9, 10, 13 modified
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The authors have addressed concerns raised by the reviewer in the previous review (applsci-1297886 Resubmitted from applsci-1222460.
Author Response
Comment 1:
Added references.
Comment 2:
1) There is no special equation, and the calibration method was written in the paper.
2) y-axis alignment complete
3) Both channels 13 & 15 have a signal. The signal level is less than other sensors.
4) Other channels are not noise. channel 4 & 8 have a lot of noise.
Comment 3:
1) Edit abstract and introduction
2) fig. 3 & 4 Delete
3) fig. 7 uses the specimen in fig. 5
4) fig. 9, 10, 13 modified
Author Response File: Author Response.pdf
Reviewer 3 Report
applsci-1297886 Resubmitted from applsci-1222460 is acceptable for Publishing in the presented form
Author Response
1) Edit abstract and introduction
2) fig. 3 & 4 Delete
3) fig. 7 uses the specimen in fig. 5
4) fig. 9, 10, 13 modified
Author Response File: Author Response.pdf
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
The reviewed article deals with Natural Rail Surface Defect Inspection and Analysis. The structure of this article is suitably arranged. I have several comments concerning content of the article. Firstly, the chapter “Introduction” is insufficiently worked out. This chapter is very brief, it contains only a minimal literature review and concept of this chapter is inappropriate for the given article because it does not provide sufficient amount of information. The chapter “Conclusion” is elaborated insufficiently and with a low-level of understandability. I recommend to rework significantly the first and last chapter of the given article.
Author Response
- Introduction
The steadily developing railway industry provides the representative mode of transportation for both public and industrial resources. Particularly, rails are crucial in tracks that distribute load through train supports, sleepers, and roadbeds. However, damage and defects occur owing to external environmental factors such as rail aging and fatigue load caused by contact between the wheel and rail, thus resulting in human injury and economic loss [1]. Particularly, various surface defects in rails occur most frequently during rolling stock operation, and internal defects may develop which leads to accidents if maintenance such as grinding is insufficient or not performed in advance.
Therefore, researchers are investigating various non-destructive techniques to ensure rail stability. Non-destructive techniques for rail inspection include various techniques such as contact ultrasonic inspection, non-contact eddy current inspection, and magnetic leakage inspection [2].
A commonly used technique is inspection using ultrasonic waves, which is widely used to detect defects occurring inside rails. However, in the case of inspection using ultrasonic waves, it is difficult to detect defects occurring on the pole surface due to the initial signal in the case of the pole surface, and in the case of squat defects, there is a problem in the propagation of the ultrasonic beam. Therefore, it is advantageous to apply a different inspection technique to inspect the surface of the rail.
The eddy current inspection technique is a non-contact method in which a high-frequency alternating current is applied to generate a magnetic field to detect changes in impedance between a coil and specimen, or changes in voltage induced in a coil. This provides an excellent detection sensitivity of the surface defects, and electrical signal can provide information such as the shape and depth of the defect [3]. The received signal output are X (resistance) and Y (reactance) components, and can be expressed through the amplitude, phase, and Lissajous plane to derive the defect information [4].
Existing eddy current uses single sensor used to determine the defect generated in the center of the rail. However, defects do not occur only in the center of the rail, and mainly occur outside the rail, so the entire head of the rail must be inspected. In this study, multiple eddy current sensors were placed to inspect the entire head of the rail. The sensors are arranged to minimize the interference between the sensors so that they are optimized for detection of defects occurring on the rail surface.
Furthermore, a software was developed specifically catered to the new set up. Since multi-channel eddy current flaw detection is performed, two-dimensional (2D) and three-dimensional (3D) shapes of the rail surface can be obtained. The inspector can conduct the experiment only by looking at the flaw screen, and since the result data can be confirmed as an image later, defects can be monitored.
In this study, an eddy current flaw detection optimized for rail surface defect inspection was used, and 16 eddy current sensors were used to simultaneously inspect the entire rail head. In addition, two-dimensional and three-dimensional rail surface shapes was obtainable through separate software, and a relational expression for analyzing the depth of the defect was derived by manufacturing artificial defects on the rail surface.
- Conclusions
In this study, an artificial defect was installed on the UIC60 rail surface, and using a 16-channel eddy current flaw detection system, a relational equation to analyze the depth was derived. In addition, two-dimensional and three-dimensional defect images could be acquired using the newly developed program.
Sensor calibration and signal processing techniques were used to improve the inspection accuracy and signals. The comparison of signal characteristics indicated that the peak width of the head check defect was narrower than shelling defect. Two defects types (head check and shelling) were reproduced in 2D images and the images were compared to the actual defects. The location and distribution of the defect signals in the images overlapped with those of the actual defects. Furthermore, the position of the defect and the part in which the defect propagated in the depth direction were confirmed, which could not be confirmed visually in the previous stages. The 3D images of the defects allowed us to confirm the signal size of the section in which the depth had substantially increased, as well as the surface roughness according to the degree of noise. The comparison of the 2D and 3D images using the 16-channel eddy current system with actual defects confirmed the detection accuracy.
Furthermore, to analyze the depth of defects, artificial defects were produced to derive the correlation between defects and phase. To confirm the reliability of the derived correlation, it can be confirmed that the error range does not exceed ±1mm through a number of experiments. It was confirmed that the depth can be analyzed using the correlation derived from the depth measurement of natural defects.
The 16-channel eddy current flaw detection equipment can be used to measure the defects on the rail surface, and it is expected to contribute to the stable management of the rail by applying it to 50K and 60K specimens.
Reviewer 2 Report
The paper discusses about development of the 16-channel eddy current based sensing system, including evaluation, for inspecting defects on natural rail surfaces. It seems interesting and highly practical as railway plays an important economic role in transportation. The reviewer has two minor comments.
- The references are not adequate. The authors should review more related works to demonstrate superiority of their contributions.
- The manuscript must be proofread by a professional proofreader.
Author Response
Revised the references.
Reviewer 3 Report
- How about the cross-talk between channels?
- For the Y signal of Fig.13, the edge signal decreased with the increasing of the depth, normally the edge signal should increase with the depth, how to explain you result.
- For Fig.14, the phase signal has relation with both the size and the depth of the defect. When you use the phase signal to judge the depth, how to remove the influence of the size of defect?
- What the scanning speed of your measurement? How about the influence of the scanning speed to the NDE result?
Author Response
1. Cross-talk refused to disclose as a technology of sensor manufacturers.
2. From the beginning, it is 8mm, 6mm, 4mm, 2mm.
3. To measure the size of the defect, we measure the size of two feet, so it does not affect the depth.
4. Measurement speed 2km/h
The signal is distorted because the data processing speed of the device cannot keep up with the measurement speed.
Round 2
Reviewer 3 Report
The authors have addressed concerns raised by the reviewer in the previous review (applsci-1297886 Resubmitted from applsci-1222460.
