Design of orthogonal eddy-current probe for weld crack detection of carbon-steel plate

Abstract

Due to uneven surface and lift-off effect, it is difficult to detect weld crack by eddy-current testing. A new orthogonal eddy-current probe for weld crack detection of carbon-steel plate was designed in this paper. Based on COMSOL Multiphysics, the influence of scanning angle on detection sensitivity of the probe was compared firstly. Then, the effects of coil width, coil side length, detection coil height, and lift-off distance on detection sensitivity of the probe were studied, respectively. Finally, the test piece of carbon-steel plate weld with crack, and the physical probe used to verify the crack detection effect were made. The experimental results show that the weld crack of carbon-steel plate with length × width × depth of 20.0 mm × 0.3 mm × 1 mm can be effectively identified, and the lift-off noise can be effectively suppressed by the method presented in this paper. At the same time, the signal-to-noise ratio of the probe keeps constant in the lift-off distance range of 0.3 mm–3.0 mm.

Keywords

Orthogonal eddy-current probe carbon-steel weld nondestructive testing COMSOL multiphysics

1. Introduction

Steel plate welds are widely used in industrial products. Due to steel plate welds are parts with poor stability in many industrial products, the weld crack need to be early detected and evaluated to ensure stable operation and avoid major accident in production and long-term service. A large number of works for crack detection of steel plate weld have been carried out by the researchers. Ultrasonic testing (UT) technology was applied to the crack inspection of friction stir welds by Tabatabaeipour et al. [1]. Muthukumaran Malarvel et al. used X-ray (RT) to detect weld crack of steel plate [2,3]. Yunpeng Fu et al. used phased array ultrasonic inspection (PAUT) to detect vertical defect on butt-joint weld [4,5]. Xiangdong Gao et al. used magneto-optical imaging (MO) technology to identify surface and near-surface cracks of steel plate welds [6,7]. The methods of electromagnetic non-destructive testing were also widely used in weld crack inspection owing to its non-contact, high detection efficiency, and no-pollution. Such as Wei Li et al. used the method of alternating current field measurement (ACFM) to detect the complex crack on weld surface of submarine metal pipe [8–10]. Differential Ionic planar coils were used to identify crack at different angle on the weld surface by Telmo G. et al. [11,12]. However, these non-destructive testing methods also show different disadvantages in the process of weld crack detection. For example, the UT method needs couplant to fill the gap between the probe and the test piece. The RT detection exhibits great radiation damage to human body. The PAUT detection is complex and expensive. The MO detection method is not only complicated in operation, but also shows low detection accuracy. The electromagnetic non-destructive testing methods also show some shortcomings. Such as the ACFM detection method is limited by the size of the probe structure when detecting weld defect, and the detection of weld in narrow space is difficult. Due to the influences of the manufacturing process, the IOnic probe expresses low measurement accuracy and it’s flexibility is limited by the size of probe when testing small specimen.

The method based on orthogonal eddy-current probe is different from the conventional eddy current detection method. It can effectively suppress the lift-off noise [13], which shows unique advantage in detection crack of uneven surface such as weld. The application of orthogonal eddycurrent probe in detection of weld crack is studied in this paper. Firstly, the influence of different scanning angle on the detection sensitivity of the probe was compared. Then, the effects of coil width, coil side length, detection coil height, and lift-off distance on detection sensitivity of the probe were studied respectively. Finally, a physical probe was made based on optimized parameters. And the weld crack detection effects were verified by the physical probe and a weld specimen of carbon-steel plate with a crack. The experimental results show that the weld crack with length × width × depth of 20.0 mm × 0.3 mm × 1 mm can be effectively detected, and the lift-off noise can be effectively suppressed by the method presented in this paper.

2. Structural design and performance study of orthogonal eddy-current probe

Fig. 1.

The structure diagram of the probe.

The structure diagram of orthogonal eddy-current probe is shown in Fig. 1. The detection coil is placed inside the excitation coil and they are orthogonal. The bottom edges of excitation coil with side lengths of L1 × L1 and detection coil L1 × L2 are keep contact. ΔS is the relative height between the excitation and detection coils, H is the lift-off distance of the probe, which represents the distance from bottom edge of excitation coil to the surface of detected metal. The widths of excitation and detection coils are both W. When the probe is working, an eddy-current field which distributes symmetrically below the detection coil is excited in the metal.

Table 1

The initial and optimized parameters of the probe

Parameter	Symbol	Initial value	Optimized value
Excitation coil side length	L1∗ L1	6.0 mm ∗ 6.0 mm	6.0 mm ∗ 6.0 mm
Detection coil side length	L1 ∗ L2	6.0 mm ∗ 4.5 mm	6.0 mm ∗ 5.4 mm
Coil width	W	1.0 mm	1.0 mm
Coil relative height	ΔS	0.1 mm	0 mm
Lift-off distance	H	0.3 mm	0.3 mm–3 mm
Number of turns of excitation and detection coil	N	150	150
Excitation frequency	f	95 KHz	95 KHz

Due to symmetry of the eddy-current field, the output signals of probe are zero at different lift-off distance when there is no crack [13]. Otherwise, the symmetry of the eddy-current field is broken, and the output signal is disturbed. However, the output signal will quickly returns to its original state after the probe sweeps through the crack.

Based on COMSOL Multiphysics, the influences of different structural parameters on the output signal of probe were studied by control variable methods. The initial parameters of the probe are shown in Table 1. In order to ensure that the excitation and induction coils are orthogonal during the simulation, the side lengths of them increase or decrease simultaneously. In addition, a groove crack with length × width × depth of 20 mm × 0.3 mm × 1 mm was set on the surface of carbon-steel plate, as shown in Fig. 1.

2.1. The influence of scanning angle on output signal

Firstly, the influence of scanning angle on output signal of probe was investigated in this paper. The scanning angle of probe is defined as the angle between the excitation coil and the groove crack, and the scanning angle increases counterclockwise, as shown in Fig. 2. The probe scans above the groove crack at different scanning angles by simulation calculation and the output signals are shown in Fig. 3. It can be seen from Fig. 3 that the crack signal amplitude increases as the scanning angle increases from 0° to 45°. During the increase of probe scanning angle from 45° to 90°, the crack signal amplitude decreases. When the probe scanning angle is 0° or 90°, the crack signal amplitude is smallest, almost 0. Figure 3 illustrates that in the range of 0°–90°, the detection sensitivity of the probe is largest when the scanning angle is 45°. Due to the direction of the magnetic field is parallel to the detection coil and symmetrically distributes on its both sides, when scanning angle is 0° or 90°, the probe exhibits a self-zeroing characteristic, which results in the crack signal is almost zero. Therefore, the probe scanning angle of 45° was selected in this paper.

Fig. 2.

The schematic diagram of the scanning angles of the probe.

Fig. 3.

The crack signals diagram at different scanning angles.

Fig. 4.

The relationship of the crack signal amplitude depending on the coil width.

2.2. The influence of coil width on output signal

In this part, the coil width W dependence of the output signal was studied. As shown in Table 1, the other initial parameters of the probe kept constant, the scanning angle was set to 45°. The width W of the probe increases from 0.6 mm to 3.0 mm with a step size of 0.2 mm. The relationship between the crack signal amplitude and the width W of the coil was obtained, as shown in Fig. 4. As can be seen from Fig. 4, when the coil turns keep constant, the wider coil width W, the smaller crack signal amplitude. Figure 4 shows that the smaller coil width, the more concentrated eddy-current distribution in the carbon-steel test piece, and the stronger crack detection ability. However, due to the requirement of the coil turns and the wire with a certain volume, the width of the coil cannot be extremely small. In the paper, the width W of the coil was set to 1.0 mm.

2.3. The influence of coil side length L1 on output signal

The output signal is also affected by the side length L1 of probe. Under the action of excitation coil magnetic field, a narrow eddy-current field is formed on the surface of carbon-steel sheet. A change in the length of excitation coil causes a change in the length of eddy-current field. As shown in Fig. 5, the coil side length L1 increases from 4.0 mm to 5.8 mm with step size of 0.2 mm, the width of coil is 1.0 mm, the scanning angle is 45°, and the other parameters are shown in Table 1, the crack signal diagram of the groove crack at different coil side lengths L1 is obtained. It can be seen from Fig. 5 that the larger side length L1 of the coil, the larger crack signal amplitude in the range of 4.0 mm–5.8 mm. For the same crack, the larger side length L2 of the coil, the faster crack signal amplitude increases. However, in the actual eddy-current detection, as the coil geometry increases, the resolution of crack decreases. So the coil side length cannot be too large. In the paper, 5.4 mm was selected as side length L1.

Fig. 5.

The crack signal diagram of different coil side lengths L1.

Fig. 6.

The crack signal amplitude as a function of coil relative height.

2.4. The influence of relative height on output signal

When the side length L1 is 5.4 mm, the output signal also be affected by effective area of detection coil. The bottom edge of detection coil was attached to the excitation coil closely, the effective area of detection coil only depends on the relative height ΔS between the two coils. Next, the influence of the relative height ΔS on output signal of the probe was studied. Figure 6 depicts the crack signal amplitude as a function of coil relative height ΔS in the process of ΔS increases from 0.1mm to 1.0mm with a step size of 0.1 mm. As can be seen from Fig. 6, the crack signal amplitude decreases linearly with increasing ΔS in the range of 0.1 mm to 1.0 mm. The reason for this phenomenon is that as ΔS increases, the area of detection coil decreases, which results in the magnetic flux of detection coil decreases. Figure 6 shows that the sensitivity of the probe is highest when the detection coil upper side is closely attached to the excitation coil.

2.5. The influence of lift-off distance on output signal

According to the previous simulation analyses, the optimized parameters of the eddy-current probe are shown in Table 1. Then the influence of the lift-off distance on the optimized eddy-current probe was studied. A crack signal diagram was obtained in the process of the lift-off distance increases from 0.3 mm to 4.2 mm with a step size of 0.3 mm. As shown in Fig. 7, as the lift-off distance increases, the crack signal amplitude shows an decreasing trend in the range of 0.3 mm–4.2 mm. In order to further study the effect of the lift-off distance on the crack signal, the relationship between the crack signal amplitude and the lift-off distance was studied in the range of 0.3 mm–4.2 mm, as shown in Fig. 8. It can be seen from Fig. 8 that although the crack signal amplitude decreases exponentially with increasing lift-off distance in the range of 0.3 mm–4.2 mm, it is still large enough at lift-off distance of 3.0 mm. Figure 7 and Fig. 8 show that the maximum lift-off distance of the optimized probe can reach 3.0 mm.

Fig. 7.

The crack signal diagram of different lift-off distances.

Fig. 8.

The lift-off distance dependence of the crack signal amplitude.

3. Experimental results and analysis

According to simulation analysis, the optimized probe can effectively identify crack defects in the lift-off distance range of 0.3 mm–3.0 mm. In addition, due to symmetry of the probe structure, this probe shows little lift-off noise [13]. Therefore, the probe can meet the requirement of large lift-off distance and suppress the lift-off noise caused by the uneven surface during weld crack detection.

In order to verify the detection ability of the probe designed in the paper. The orthogonal eddy current probe was fabricated according to optimized parameters, and a groove crack with length × width × depth of 20 mm × 0.3 mm × 1 mm was made on the weld surface of carbon-steel plate by laser processing method, as shown in Fig. 9. Next, the weld with crack was tested at different lift-off distances of 0.3 mm, 1.0 mm, 2.0 mm and 3.0 mm, respectively. As shown in Fig. 10, a weld crack detection waveform picture was obtained. In order to better identify, zero baselines of these waveforms had been parallel moved. It can be seen from Fig. 10 that the probe can effectively detect the groove crack on the weld in range of lift-off distances from 0.3 mm to 3.0 mm, and the signal waveforms of defective part change significantly, the parts without defective show some fluctuation. It demonstrates that although there is some signal noise for uneven surface, the probe can effectively identify weld crack defect.

Fig. 9.

The photo of the test piece and the probe.

Fig. 10.

The time dependence of output signal at four different lift-off distances of 0.3 mm, 1.0 mm, 2.0 mm, and 3.0 mm.

Next, in order to analyze the influence of weld surface unevenness on the crack detection ability of the probe at different lift-off distances. The signal-to-noise ratio (SNR) of the probe was calculated, as shown in Table 2. The SNR in the paper is defined as: $\begin{eqnarray}\displaystyle SNR=V_{N}/V_{S}\bullet 100\% & & \displaystyle\end{eqnarray}$ (1)V_N and V_S are the signal peak-to-peak values of noise and crack signal, respectively.

It can be seen from Table 2 that the probe SNR changes little in the range of lift-off distance from 0.3 mm to 3.0 mm. Therefore, different lift-off distances show little influence on the crack detection capability of the probe.

Table 2

The Probe SNR at different lift-off distances

Lift-off distance	0.3 mm	1.0 mm	2.0 mm	3.0 mm
V_S	2.84E+00	1.20E+00	5.20E-01	4.80E-01
V_N	670E-01	2.60E-01	1.04E-01	9.12E-02
SNR	23.6%	21.6%	20%	19%

4. Conclusion

For the problem of signal noise caused by uneven surface during the weld crack detection of carbon-steel plate. A new orthogonal eddy-current probe was designed by the method of simulation calculation. This eddy current probe can effectively suppress the lift-off noise during the weld crack detection. Then, the orthogonal eddy current probe and the weld test piece with a groove crack were fabricated. The weld crack was detected at different lift-off distances using the eddy-current probe. The experimental results show that the weld crack with length × width × depth of 2.0 mm × 0.3 mm × 1.0 mm can be effectively detected, and the lift-off noise can be effectively suppressed by the method presented in this paper. At the same time, the SNR of the probe keeps constant in the range of lift-off distances from 0.3 mm to 3.0 mm.

Footnotes

Acknowledgements

This work was supported by National Natural Science Foundation of China (Grant nos. 51605148), Hubei Natural Science Foundation Innovation Group Project (Grant nos. 2019CFA021), and Major Technological Innovation Projects in Hubei Province (Grant nos. 2018AAA034).

References

Tabatabaeipour

Hettler

Delrue

, Non-destructive ultrasonic examination of root defects in friction stir welded butt-joints, Ndt & E International80 (2016), S0963869516000128.

Malarvel

, Anisotropic diffusion based denoising on X-radiography images to detect weld defects, Digital Signal Processing68 (2017), S1051200417301148.

Malarvel

Muthukumaran

, An improved version of Otsu’s method for segmentation of weld defects on X-radiography images, Optik - International Journal for Light and Electron Optics109 (2017), S0030402617306022.

Fortunato

, Friction stir weld-bonding defect inspection using phased array ultrasonic testing, International Journal of Advanced Manufacturing Technology (2017), 1–10.

, Phased array ultrasonic test of vertical defect on butt-joint weld of CFETR vacuum vessel port stub, Fusion Engineering and Design141 (2019), 1–8.

, Weld cracks nondestructive testing based on magneto-optical imaging under alternating magnetic field excitation, Sensors and Actuators A: Physical285 (2019), 289–299.

Gao

Nvjie

and Liangliang

, Magneto-optical imaging characteristics of weld defects under alternating magnetic field excitation, Optics Express26(8) (2018), 9972.

, High sensitivity rotating alternating current field measurement for arbitrary-angle underwater cracks, NDT & E International79 (2016), 123–131.

Papaelias

M.Ph.

, High-speed inspection of rails using ACFM techniques, NDT & E International42(4) (2009), 328–335.

10.

Raine

and Lugg

, A review of the alternating current field measurement inspection technique, Sensor Review19(3) (1999), 207–213.

11.

Santos

G. Telmo

, Application of eddy current techniques to inspect friction spot welds in aluminium alloy Aa2024 and a composite material, Welding in the World55(9–10) (2011), 12–18.

12.

Rosado

S.L.

, A new dual driver planar eddy current probe with dynamically controlled induction pattern, NDT & E International70 (2015), 29–37.

13.

Ageng

and Kasai

, The Improvement of Flaw Detection by the Configuration of Uniform Eddy Current Probes, Sensors, Basel, Switzerland, 2019, p. 19.