Electromagnetic testing of moisture separator reheater tubes using a bobbin-type integrated Hall sensor array

Abstract

This paper presents an electromagnetic testing method for moisture separator reheater (MSR) tubes, which is based on partial saturation eddy current testing (PSECT). The sensor probe has a permanent magnet that is used to partially saturate the MSR tube material, a bobbin coil is used to supply an alternating electromagnetic field that penetrates into the tube, and a bobbin-type integrated Hall sensor (BIHaS) array is used to measure the radial component of the electromagnetic field distribution in the tube. The BIHaS has 15 Hall sensor elements arrayed at 24 ${}^{\circ}$ (2.72 mm) intervals on the circumference to make a 13.5 mm diameter sensor probe. Thus, the spatial resolution of measurement can be improved rather than using a single/differential-type sensing coil as used in conventional PSECT systems. The efficiency of PSECT over magnetic flux leakage testing and eddy current testing is presented. Furthermore, the quantitative evaluation of damage volume can be performed.

Keywords

Hall sensor array partial saturation eddy current testing electromagnetic testing moisture separator reheater

1. Introduction

The moisture separator reheater (MSR) system is set up between the high pressure (HP) and low pressure (LP) steam turbines in a pressurized-water nuclear power plant. The MSR is usually composed of finned tubes that have about three times the surface area of un-finned tubes. The MSR removes moisture and reheats the cycle steam to protect the LP turbine blades from erosion and corrosion and improve the thermodynamic efficiency of the LP turbines [1]. In the past, the MSR tube was made of 90-10 CuNi, but this has been replaced by ferritic stainless steel (SS439), a ferromagnetic material that is stronger and has excellent corrosion resistance [2]. However, after long-term operation, the erosion and corrosion can appear on the outer surface of the MSR tubes [2, 3]. The failure of the MSR tubes could stop the operation of the nuclear power plant. Thus, nondestructive testing is required to inspect the damage at an early stage.

Ultrasonic testing (UT) systems using guided waves have been developed for inspection of MSR tubes [4]. However, the UT system requires a coupling material between the guide wave and tube surface that is difficult to maintain during inspection. Electromagnetic testing is a preferred technique, and both magnetic flux leakage testing (MFLT) and eddy current testing (ECT) have been successfully used for inspection of tubes with no requirement for coupling material [5, 6, 7, 8, 9]. MFLT is used for inspection of ferromagnetic material by measuring the magnetic flux leakage around the damage [5, 6]. However, it is limited in a finned tube because the fins produce a high level of noise. ECT can be applied for both ferromagnetic and paramagnetic materials [7, 9], but it is difficult for an electromagnetic field to penetrate into ferromagnetic material targets because the high permeability of the material prevents deep penetration of the eddy current. Remote field eddy current testing (RFECT), in which the sensing coil is placed far from the exciting coil (2–3 times of the inner diameter of the tube), is an efficient method for inspection of ferromagnetic material [10]. RFECT is equally sensitive to inner and outer diameter damages, so they could not be distinguished. Therefore, full/partial magnetic saturation of the tube material is performed to reduce the permeability, and then the ECT technique can be applied; this is called full/partial saturation eddy current testing (FSECT/PSECT). A sample partial saturation is shown in Fig. 1. This is finite simulation result by ANSYS software for a static magnetic field distribution in a ferromagnetic tube produced by a permanent magnet. The material under the defect was highly magnetized due to the loss of the material. To fully saturate the material, a strong permanent magnet or a coil with a high DC current is necessary, but it’s difficult in real applications. Conventional FSECT/PSECT systems use a sinusoidal wave current to supply the exciting coil [11, 12], but a pulse wave form has been recently used with more advantages by the mean of superposition of infinite frequencies [13]. However, the current FSECT/PSECT systems still have limited spatial resolution because of the use of a single/differential-type sensing coil.

Figure 1.

Finite simulation results of magnetic field distribution in a ferromagnetic tube by a permanent magnet. The simulation was performed by ANSYS software.

Therefore, this paper proposes a PSECT system using a bobbin-type integrated Hall sensor (BIHaS) array instead of a single/differential-type sensing coil as used in conventional PSECT systems. The BIHaS is composed of many Hall sensor elements arrayed in the circumferential direction of the probe so that the location of damage in the circumference of the MSR tube can be identified and a quantitative evaluation of damage size can be performed [14, 15]. The BIHaS is used in three inspection systems (MFLT, ECT and PSECT), and the advantages of the PSECT system over the MFLT and ECT systems are presented.

2. Principles

Figure 2 shows the three inspection techniques, which are MFLT (Fig. 2a), ECT (Fig. 2b), and PSECT (Fig. 2c), using a BIHaS. In MFLT, a permanent magnet is used to produce magnetic leakage around damage in the tube, the radial component of which is measured by the BIHaS [14, 15, 16]. In ECT, an exciting bobbin coil is used to produce eddy currents in the circumferential direction of the tube. If there is a damage in the tube, the eddy current will be distorted and concentrated around the damage. The radial component of the secondary magnetic field, which is induced from the eddy current, will be measured using the BIHaS. PSECT is a combination of the two techniques, in which a permanent magnet is used to partially saturate the tube material and a bobbin coil is used to produce the eddy current in the tube.

Figure 2.

Three types of inspection technique using a bobbin-type integrated Hall sensor array: a) magnetic flux leakage testing (MFLT), b) eddy current testing (ECT), and c) partial saturation eddy current testing (PSECT).

At certain scan speed of MFLT or ECT system; an extra induced current in conductive specimen is produced according to Lorentz’s Law and Lentz’ Law. This current makes a reduction of electromagnetic field of damage. Thus, the detectability of the system could be reduced at a high scan speed. Tian el at. observed in theirs research [17] that the magnetic field distribution around a crack was changed in pattern as well as intensity when the scan speed is about 36–108 km/h. However, at a low speed as using in this study, the effects could be ignored [18].

The output voltage of the BIHaS is processed using high-pass filters (in ECT and PSECT) or low-pass filters (in MFLT), differential-type amplifiers, root-mean-square-circuits for converting to DC signals (in ECT and PSECT) and through an AD converter to a computer.

3. Experimental setup

Figure 3 shows the experimental setup of the three inspection systems. The BIHaS has 15 Hall sensor elements arrayed at 24 ${}^{\circ}$ (2.72 mm) intervals. The bobbin coil has a length of 10 mm and 130 turns of 0.2-mm-diameter copper wire. The permanent magnet has a diameter of 6 mm and a length of 20 mm with 120-mT strength in a pole. The diameter of the sensor probes is approximately 13.5 mm. The numbers of HPFs, LPFs, amplifiers, and RMSs are same as the number of sensor elements [15], which allows simultaneous processing. The cut-off frequency of the HPFs and LPFs is 300 Hz, and the gain of the amplifiers is 60 dB for ECT and PSECT and 10 dB for MFLT. The NI-PCI 6255 used as the ADC and I/O controller has a total of 80 analog input channels, 16 bit resolution, 1.25 MHz sampling rate, and a $\pm$ 10 V input range.

A ferritic stainless steel (SS439) MSR finned tube was used in the experiment. The inner diameter, outer diameter, and wall thickness (W.T) are 14.1, 19.1, and 1.2 mm (excluding fin height). The fin has a height of 1.3 mm and a step of 1.1 fins/mm. Artificial defects (flat-bottom hole) were produced on the outer surface of the pipe with depths from 10 to 100% of the W.T and diameters of 1.3, and 2.5 mm. The detailed defects sizes are shown in Table 1. The MSR tube was scanned at a speed of 28 mm/s and 1 mm/data interval. In addition, the attraction force between the permanent magnet and the MSR tube could make noise and damage to the sensor during scan. Thus, two centering parts have been attached at the front and end sides of the sensor probe, as shown in a picture sensor prototype in Fig. 3.

Table 1
Sizes of defects on the MSR tube

#	Depth	Depth	Diameter	Volume	#	Depth	Depth	Diameter	Volume
	(% W.T)	(mm)	(mm)	(mm ${}^{3}$ )		(% W.T)	(mm)	(mm)	(mm ${}^{3}$ )
1	100	1.20	2.5	5.91	6	60	0.72	1.3	0.95
2	87	1.05	2.5	5.04	7	40	0.48	1.3	0.63
3	58	0.70	2.5	3.33	8	20	0.24	1.3	0.31
4	100	1.20	1.3	1.59	9	10	0.12	1.3	0.19
5	80	0.96	1.3	1.27

Figure 3.

Experimental setup of the three inspection systems.

4. Experimental results

In experimental results, the differential ( $\Delta V_{\textit{RMS}}$ ) of the RMS voltage ( $V_{\textit{RMS}}$ ) of the BIHaS in the scan direction is used, as expressed in Eq. (1). Because, the differences of offset and sensitivity of each Hall sensor is eliminated and defects could be observed clearly [14, 15, 16].

$\Delta V_{\textit{RMS}}(i,j)=V_{\textit{RMS}}(i+1,j)-V_{\textit{RMS}}(i,j)$ (1)

where $i$ and $j$ are the scan data index and sensor element index, respectively.

Figure 4 shows the experimental results of MFLT (Fig. 4a), ECT (Fig. 4b), and PSECT (Fig. 4c). A current of 300 mA at 900 Hz was supplied to the bobbin coil in ECT and PSECT. In middle of defects #2 and #6, a high signal appeared due to the deformation of the specimen during the milling operation of defect #2. It can be observed that the MFLT result has the highest noise, which is mainly from the fins of the MSR tube. Thus, it is difficult for MFLT to detect small defects. It can detect a minimum size of 60% W.T depth – Ø1.3 (#6). The noise in the PSECT is slightly higher than in the ECT because the eddy current in the PSECT penetrates deeper into the wall thickness and approaches to the fins, and that produces higher noise. However, the intensity of the defect signal at each frequency in PSECT was significantly higher than that in the ECT. ECT can detect defects with 40% W.T depth – Ø1.3 (#7), and 20% W.T depth (#8) can be observed, but with low intensity. The PSECT provides the best detection probability, i.e., 20% W.T depth – Ø1.3 (#8) can be detected. It can be observed that the intensity of the defect signal increases as the defect size increases in all three inspection techniques. Thus, the evaluation of defect size can be performed. Furthermore, the locations of defects on the circumferential direction of the MSR tube can be observed.

Figure 4.

Experimental results of defects on the MSR tube by a) MFLT, b) ECT with exciting current of 300 mA–900 Hz, and c) PSECT with exciting current of 300 mA–900 Hz.

For quantitative evaluation of defect volume (Vol) for ECT and PSECT, the sum of the absolute defect area signal ( $S$ , expressed in Eq. (2)) was used. The Vol and S are approximated as having linear relationship, as shown in Fig. 5a. The relation is mathematically expressed in Eq. (3).

$\displaystyle S=\sum\limits_{i=1}^{N}{\sum\limits_{j=1}^{M}{\left|{\Delta V_{% \textit{RMS}}(i,j)}\right|}}$ (2) $\displaystyle S=a\times\textit{Vol}+b$ (3)

where $M$ and $N$ are the number of data points in the scan length and the number of sensor elements, respectively. Here, $M=$ 16 and $N=$ 10. The slope ( $a$ ) and intercept ( $b$ ) of the fitting line are called “sensitivity” and “bias” of defect detection. They’re depending on the inspection method (e.g. ECT, PSECT), exciting frequency, exciting current, etc. The “sensitivity” means that how much sensitivity of the system with a certain condition responses for a defect, and the “bias” means that the basic level of the response signal for a defect free. The higher “sensitivity” $a$ shows a better condition of a system for defect detection and evaluation.

Figure 5.

Evaluation of defect volume: a) linear relationship between sum of absolute defect area signal with defect volume at 900 Hz, b) sensitivity of defect area signal against exciting frequencies, and c) estimation of defect volume results at 900 Hz.

The “sensitivity” $a$ according to differences of exciting frequencies in ECT and PSECT is described in Fig. 5b, in which PSECT provides higher sensitivity than ECT. The sensitivity increases as frequency increases in PSECT. However, there is a slight decrease at 2000 Hz in ECT because the high permeability of the MSR tube reduces the penetration depth of the eddy current at high frequency, and thus, the intensity of the defect signal is reduced. Besides, by using the linear relationship between the $S$ and Vol, an estimation of Vol can be performed by the inverse of Eq. (3). Figure 5c shows estimation result of defect volume at 900 Hz for ECT and PSECT. The factors ( $a$ , $b$ ) are (0.0278, 0.0454) and (0.048, 0.0728) for ECT and PSECT, respectively. Normalized root-mean-square error, which is RMS error of the estimation over measured defect volume range, was obtained at 14.58% and 11.58% for ECT and PSECT, respectively.

5. Conclusions

This paper presents partial saturation eddy current testing (PSECT), a method of electromagnetic testing using a bobbin-type integrated Hall sensor (BIHaS) array. The BIHaS has 15 Hall sensor elements arrayed at 24 ${}^{\circ}$ (2.72 mm) intervals on a 13.5 mm diameter probe that improves the spatial resolution compared to conventional PSECT systems. To verify the advantages of PSECT compared to eddy current testing (ECT) and magnetic flux leakage testing (MFLT), the BIHaS was used in all the three systems for inspection of artificial defects on a finned MSR tube. The experimental results show that PSECT provides the best detection results and can inspect Ø1.3 mm – 20% of tube wall thickness depth (0.24 mm). In addition, the defect volume can be estimated using sum of absolute defect area data with a normalized root-mean-square error of 11.58%.

Footnotes

Acknowledgments

This research was funded by KEPCO Plant Engineering, and the Korea Institute of Energy Technology Evaluation and Planning (KETEP 20171520101610). We are grateful for the support.

References

EPRI: Moisture separator reheater source book. Technical Report, TR-106345, Nov. 1997.

Gibson

J.N.

, Use of type 439L stainless steel tubing in moiture separator/reheaters, Journal of Materials for Energy Systems4(3) (1982), 136–141.

EPRI: Electromagnetic NDE guide for balance-of-plant heat exchangers. Rev. 2, TR-101772-R2, 3.2-21, 4.0-29

\sim

4.0-33, 1997.

EPRI: Evaluation of the Guided Wave Inspection Technique Using 439 Stainless Steel Finned Tubing. Report GC-109349, Oct 1997.

Kikuchi

Kurisawa

Ara

Kamada

and Kobayashi

, Feasibility study of magnetic flux leakage method for condition monitoring of wall thinning on tube, International Journal of Applied Electromagnetics and Mechanics33 (2010), 1087–1094.

Kim

Choi

and Lee

, Employing magnetic sensor array for inspecting defects in a pipeline, IEEE Sensors Applications Symposium (SAS), San Antonio, TX, 22-24 Feb. 2011.

Yusa

Chen

and Miya

, Quantitative profile evaluation of natural defects in a steam generator tube from eddy current signals, International Journal of Applied Electromagnetics and Mechanics12 (2000), 139–150.

EPRI: The Development of Eddy Current Technique for WWER Steam Generators Inspection, Technical Results, ID.1011252, 2005.

Uchanin

and Najda

, NDE for Carbon Steel Heat Exchanger Tube Support Plates and Tube Fretting, InTech (2011), 145–164.

10.

Chen

Rebican

Miya

and Takagi

, 3D simulation of remote field ECT by using Ar method and a new formula for signal calculation, Research in Nondestructive Evaluation16 (2005), 35–53.

11.

Lee

H.J.

Cho

C.H.

Nam

M.W.

Yoo

H.J.

and Hong

S.Y.

, Eddy current testing of type-439 stainless steel tubing using magnetic saturation technique, Journal of Mechanical Science and Technology26(7) (2012), 2081–2085.

12.

Sadek

H.M.

, NDE technologies for the examination of heat exchangers and boiler tubes – principles, advantages and limitations, Insight48(3) (2006), 181–183.

13.

Xie

Chen

Takagi

and Uchimoto

, Evaluation of Wall Thinning in Carbon Steel Piping Based on Magnetic Saturation Pulsed Eddy Current Testing Method, Stud. Appl. Electromagn. Mech.39 (2013), 296–303.

14.

Lee

Jun

Kim

Choi

and Le

, Bobbin-Type Solid-State Hall Sensor Array With High Spatial Resolution for Defects Inspection in Small-Bore Piping Systems, IEEE Trans. Mag.48(11) (2012), 3704–3707.

15.

Lee

Jun

and Kim

, Estimation of sizes of defects on pipes in nuclear power plants using dipole moment and finite element methods, NDT & E Int.58 (2013), 56–63.

16.

Kim

H.S.

and Lee

, Localization and evaluation of corrosion in a small-bore piping system using a bobbin-type magnetic camera, International Journal of Applied Electromagnetics and Mechanics45(1–4) (2014), 739–745.

17.

Tian

G.Y.

and Ward

, Numerical simulation on magnetic flux leakage evaluation at high speed, NDT & E Int.39(5) (2006), 367–373.

18.

Jun

Kim

and Lee

, Nondestructive testing of train wheels using differential-type integrated Hall sensor matrixes embedded in train rails, NDT&E Int.55 (2013), 28–35.

Electromagnetic testing of moisture separator reheater tubes using a bobbin-type integrated Hall sensor array

Abstract

Keywords

1. Introduction

Table 1 Sizes of defects on the MSR tube

Footnotes

Acknowledgments

References

Table 1
Sizes of defects on the MSR tube