Back to Basics

[ click here for the Back to Basics Archive ]

Explaining Differing Magnetic Particle Testing
Results in the Same Specimens

by Wei-Chang Zhong*

 

It is pleasing to see our colleagues in China submitting articles to Materials Evaluation. The Nanjing Gas Turbine Research Institute has contributed several papers on magnetic particle testing during the past few years. This one is interesting because it sheds light on work performed in 1979, when little was known here of Chinese NDT. Enjoy! Roderick K. Stanley Associate Technical Editor

Figure 1-3

INTRODUCTION
several years ago, the author read a problem in the Nondestructive Testing Handbook, second edition: Volume 6, Magnetic Particle Testing (ASNT, 1989) to which there was yet no answer. The details are as follows: "The absence of verification has led to widely varying effectiveness. Two often cited reliability studies are excellent examples of the problems associated with the lack of a sensitivity check. The first of the studies was the result of a round-robin among four major aerospace contractors, two jet engine manufacturers, three landing gear manufacturers, one major forging supplier and a commercial test lab. Each company processed and tested twenty-four pedigreed specimens with known discontinuities. All of the discontinuities were considered detectable at that time in the method's development." Figure 1 shows the results of that study. The NDT Handbook continues, "the companies varied from less than 20 percent detection to over 90 percent. Only one of the eleven companies scored better than 60 percent of the discontinuities. At the time of its completion, the study drew few conclusions about the cause for this variability."

As soon as the author read these paragraphs, he knew the cause of this problem, because he encountered a similar phenomenon in 1979 and solved that difficult problem at that time.

---

In cases like this, the magnetic flux leakage on hairline cracks approaches zero.

---

A STRANGE PHENOMENON 

The Problem Encountered
In 1979, the author evaluated the magnetic particle testing skills of six NDT technicians in the third professional grade. (There were eight professional grades in China at that time, the eighth being the highest.) The workpieces of a steam turbine were selected as the test specimens. Their descriptions and materials, as well as the test results, are shown in Table 1.

Table 1 The tested workplaces and the results
Number	Workpiece	Material	Test Results/Discontinuities (Number of Detectors)
1	long blade	S40300	1 transversal crack (4), 3 transversal cracks (1), no discontinuity (1)
2	long blade	S42000	7 longitudinal cracks (6)
3	long blade	S40300	1 transversal crack (6)
4	small blade	S40300	1 transversal crack (6)
5	small blade	S42000	1 transversal crack (6)
6	small blade	S42000	2 transversal crack (4), 1 transversal crack (2)
7	large bolt	G41350	1 transversal crack (5), 1 transversal crack (2)
8	small bolt	G41350	many cracks (4), a lot of hairline cracks (1), no discontinuity (1)
9	spring	G10650	scratch (3), not discontinuity (3)
10	flat iron	G10450	small cracks (1), hair line cracks (1) no discontinuity (4)

 

Table 1 shows that the test results were not identical. Of course, this may be considered a reflection of the differences in the technicians' levels of proficiency. However, it is worth noting that the test results of workpieces 1 and 8 are especially scattered. There seemed to be an unknown reason for the discrepancies in the results, because the technicians all had nearly 10 years of NDT experience and they were earnest and conscientious in performing this testing.

The author selected some small bolts, which had been rejected by magnetic particle testing as containing hairline cracks on their surface, and retested them. He was surprised to find no magnetic particle indications on the surfaces of bolts where such testing had discovered hairline cracks not long before. There was no problem with the equipment or suspension.

The author then increased the magnetizing electric current step by step and carried out the experiment repeatedly. However, the result was still the same - the magnetic particle pattern of the hairline cracks, which previously appeared normal, had disappeared.

 

The Treatment Method
The specimens were all workpieces that were rejected by magnetic particle testing in normal production. The key was to determine whether there would be an error between the two instances of magnetic particle testing. The workpieces, materials, discontinuities, equipment, suspension, magnetizing current and magnetizing direction remained the same. So why did the magnetic particle pattern appear only once and nothing appeared after that?

Through repeated experiments, a treatment method designed to ensure the reappearance of the original magnetic particle pattern was finally found: remove the specimen from the magnetic particle detector, demagnetize it, wash it and clear it. After remagnetizing it and applying the suspension to it, the magnetic particle indication on the hairline cracks reappeared (Zhong and Nian, 1984; Zhong and Nian, 1993).

 

EXPLANATION

Observation
The hairline cracks are caused by nonmetallic inclusions or blowholes elongated in rolling or press forging, so the cross sectional width of the crack is extremely small in comparison to its length and the magnetic leakage flux across the gap is very small. Thus, the magnetic lines of force in the leakage field are not markedly curved (Figure 2), so that during magnetic particle testing they can attract only a small amount of magnetic particles from the ink and the visible indication of the crack is not very clear.

Although the amount of magnetic particles attracted by the crack gap is small, they can force the magnetic flux leakage released from the workpiece surface to return back to the body of the specimen through themselves (Figure 3). In cases like this, the magnetic flux leakage on hairline cracks approaches zero.

As a result, when the specimen is remagnetized, even if the magnetizing field is increased, the hairline cracks will not attract magnetic particles from the ink. Only when the specimen is demagnetized, washed, cleaned and the few magnetic particles attracted to the crack are removed, can the magnetic leakage field attract new particles and reveal the crack again.

 

Verification
In the summer of 1995, a magnetic particle testing experiment was again carried out. It was discovered that there were four long, straight magnetic particle patterns of more than 10 mm (0.4 in.) around a square hole (0.6 by 0.6 mm [0.02 in.]) on the workpiece surface (Zhong, 2000). In the summer of 2000, when this experiment was repeated, the length of the magnetic particle pattern was found to be only about 5 mm (0.2 in.). In May 2002, the same sample was tested and there was no magnetic particle indication whatsoever around the square hole on the specimen surface. The sample was demagnetized and washed every time before remagnetization. Because the hole was very small, however, the cleaning and removal of the magnetic particles in the hole was especially difficult. As a result, at the second magnetization of the sample the magnetic leakage field over the hole was less than half that during the first time; at the third magnetization, there was no magnetic leakage field over the hole at all.

 

CONCLUSION
Why, then, were the magnetic particle testing results different for the same 24 specimens with known discontinuities when tested by 11 companies?

The author believes that the new phenomenon introduced above must have been occurring when the 11 companies successively tested the 24 specimens, because the ability of each company to detect the discontinuities varied so widely, from less than 20% to over 90%. On the basis of the author's explanation of this phenomenon, it is believed that the earlier the specimens were tested by a company, the higher success rate was obtained. No two companies shared the same success rate, as can be seen in Figure 1.

This theoretically solves the problem in the Nondestructive Testing Handbook mentioned above; likewise, the reliability study discussed in the Nondestructive Testing Handbook verifies again the author's explanation of the phenomenon which he discovered in 1979.

 

ACKNOWLEDGEMENTS
The author gratefully acknowledges Wen-Xue Nian, who supplied the same information regarding the phenomenon and the treatment technique as the author, and Lian-Hua Hua, who carried out the magnetic particle testing experiment on the square hole and supplied the information regarding another experimental verification of the phenomenon. The author also acknowledges the financial support of this work by the National Natural Science Foundation of China (Grant No. 59571064).

This paper was first presented at the 8th Conference of the Chinese Society for NDT/International Research Symposium on NDT held in Su-Zhou, China, in September 2003.

 

REFERENCES
American Society for Nondestructive Testing, Nondestructive Testing Handbook, second edition: Volume 6, Magnetic Particle Testing, Columbus, Ohio, ASNT, 1989.

Zhong, Wei-Chang, "Strange Magnetic Particle Pattern around a Square Hole on the Workpiece Surface," Materials Evaluation, Vol. 59, 2001, pp. 1085-1086.

Zhong, Wei-Chang and Wen-Xue Nian, "Disappearing and Reproducing of Magnetic Indications of Remagnetized Hairline Seams," Chinese Journal of NDT, Vol. 6, No. 5, 1984, p. 34.

Zhong, Wei-Chang and Wen-Xue Nian, "Disappearance and Reappearance of Magnetic Particle Indications on Hairline Cracks, When Remagnetized," British Journal of Non-Destructive Testing, Vol. 35, 1993, p. 717.

 

* Nanjing Gas Turbine Research Institute, Nanjing 210037, China; e-mail njgtt@jlonline.com.

 

Copyright © 2005 by the American Society for Nondestructive Testing, Inc. All rights reserved.