·Home ·Table of Contents ·Materials Characterization and testing

Cracks Detection by Electromagnetic and Acoustic Emission J. Sikula, B. Koktavy, J. Pavelka, P. Koktavy , V. Sedlakova Department of Physics, Technical University of Brno, Czech Republic T. Lokajicek Geophysical Institute, Academy of Science, Prague, Czech Republic M. Tacano Meisei University, Hino, Tokyo S. Hashiguchi Yamanashi University, Kofu, Japan Contact

Abstract

Electromagnetic (EME) and acoustic (AE) emission, which are generated in stressed solids, can be used to indicate cracks and defects position, dimension and their orientation. When a crack is created in a solid, the faces of the crack make up plates of an elementary capacitor and the charges located on them constitute an electric dipole or quadrupole. The magnitude of the electric charge configuration and the dipole moment depend on the time. The electromagnetic phenomena are related to a redistribution of the electric charges. Electrostatic and magnetic fields can be detected in the vicinity of the crack, while in greater distance electromagnetic field is assumed. As a EME sensor two conducting plates are fitted to the specimen making the parallel plate capacitor. This capacitor is used to pick-up the electrostatic field given by electric charge distribution. New crack creation is accompanied by the phonons. They are measured by acoustic emission signal. The electromagnetic signal precedes the acoustic emission response and time delay corresponds to the difference of propagation velocities of sound and electromagnetic signals.

The granite samples with dimensions 4x4x1 and 8x4x1cm³ were examined. They were loaded with a ramp increasing stress up to their failure. Electric voltage measured on parallel plate capacitor is time dependent due to vibration of crack walls. From the high frequency component, the length crack was estimated. We found crack length 0.3 - 1.5 mm. Crack position was determined in experiment with four capacitor and four acoustic emission sensors. The simultaneous investigation of EME and AE signals enable us to localise the crack position in solids and estimate crack dimension. This method can be used for samples with low electrical conductivity and was successfully applied for granite and composite materials.

Introduction

It is shown that electromagnetic phenomena which are created in stressed solids can be used to indicate cracks and defects. Electromagnetic and acoustic emission signals generated by crack creation in mechanically stressed solids are phenomena, which have been known for many years already [1], [2]. Similar electromagnetic signals were observed, when birth of cracks in hard dielectrics occur. There are many papers related to electromagnetic and acoustic emission associated with rock fracture [3] - [8].

The experimental results (see [4] to [7]) can be summarised as follows: EME signals occur simultaneously with acoustic emission events. EME signals with large amplitude occurs together with large AE events. EME and AE signals increases just before the main rupture. EME signals are created by the electric charges redistribution, the piezoelectric effect, the electrokinetic effect and electrification due to friction between the newly created surfaces. In previous studies reported measurements of electric potentials related to rock fracture have been carried out under the monotonously increasing compressive stress in laboratory. Mori et al. [7] studied electric potential signals from a rock sample under the cyclic loading and found, that electric potential and AE signals mainly occur in the stage, in which the stress is increasing and reaches the highest level in each cycle. They confirmed, that the electromagnetic signals occur primarily due to a generation of a new crack and its extension. The result, that electromagnetic signals rapidly increase just before the main rupture suggest, that the EME signals may be useful as a precursory phenomena.

Experimental set-up

To give evidence of the mentioned phenomena following experiment was designed and carried out. In the experiment one or four parallel plate capacitors were used to pick up the electric field component (see Fig.1. and 2.).

Fig 1: Experimental set-up with one capacitor antenna

Fig 2: Experimental set-up with four capacitor antennas

This experimental set-up indicates also the electrostatic charge redistribution. The wavelength of the emitted radiation was in all cases much larger than the sample size so that near-field effects could be observed only. The conducting plates were fitted to the specimen making the parallel plate capacitor. The electric voltage measured on this plate capacitor C is time dependent due to the vibration of crack walls, its damping and sample electrical conductivity. The time dependence of electrical voltage is determined by the charge relaxation and phonon field interaction.

Experimental results

The granite samples dimensions of 4x4x1 and 4x8x1 cm3 were examined. They were loaded with a constant stress rate up to their failure. The capacitor antennas were fixed on DUT (device under test) according to Fig.1.- 2. Acoustic emission signal was detected on the sample edges (see Fig. 2.), where position of each AE sensor is denoted as 1 to 4. Electromagnetic emission signals (EME) are arising across the capacitors antennas (EME sensors are denoted as 5 to 8) and afterward amplified by low-noise preamplifiers (see Fig. 3. records 5-8).

Fig 3: The time dependence of AE and EME signals for the event 7123

Time dependencies of recorded signals are depicted in Fig. 3. In this event EME signals start approximately at the same time, while AE signals are delayed. The time difference between AE and EME signals depends on the crack position and the wave velocity both of EME and AE wave propagation.

Crack position

The crack responsible for these signals is located close to the antenna No. 5. and than AE signal No.1. has a minimum delay, while a maximum delay was detected by AE sensor No. 4., which is in opposite direction with respect to sensor No.1. This method can improve a crack positioning and give useful information about crack orientation, charge on crack walls and active crack area.

In Fig.4. the time dependence of AE and EME signals for the event 7130 is depicted. In this case we found that two cracks are created approximately at the same time near sample edges. The position of this cracks is schematically shown in Fig.5., points A and B.

Fig 4: Time dependence of AE and EME signals for event 7130

Fig 5: Crack position

The time dependence of EME signal in Fig.4. shows that when the crack is created near the top or the bottom edge of the sample, then electric signal in capacitor antenna is inversely proportional to the cube of the distance. In this case signal in antenna No.7 and 8. is on the background noise level for the crack B.

Fig 6: Electromagnetic emission signal

Fig 7: Acoustic emission signal

Stimulated creation of cracks was observed for event 8118 (Fig.6,7), where first crack appears at the time t = 10 ms probably on surface near AE sensor No.3. After this follow volume crack. In next step the corner near AE sensor No.3. was split off.

Crack orientation

The time dependence of EME signal is influenced by crack walls orientation. Theoretical analysis [8] show, that the time dependence of voltage on capacitor antenna is given by


This describe DC signal with t = RC constant and AC signal with damping constant , g =d + b ,d due to the mechanical damping and b the sample electrical conductivity. Here V₀ is integration constant, w_o=g- t /t is cut off angular frequency,w crack wall vibration angular frequency and factor g is given by



where v₀ is velocity amplitude, d is parallel plate capacitor thickness and a is the angle between electric field intensity and electric charge velocity .

Fig 8: Time dependence of EME and AE signals

The crack walls are mainly oriented according to a rock crystalline structure. Samples with crystalline layers parallel to area of capacitor antennas generate EME signal given by superposition of damped DC and AC signal, as in Fig.8. This was measured on experimental set-up according to Fig.2.

Crack length

We found, that the frequencies of the EME signals for granite samples were f_C = 1 to 6 MHz. The crack length l_C can be estimated from this frequency and for granite sample with velocity of mechanical wave propagation 3 km/s we find crack length l_C = 1.5 - 0.25 mm.

Crack's eigen vibrations are attenuated due to mechanical damping and electrical sample conductivity. In our experiment this high frequency crack vibration was observed for a time interval of the order 10µs. After this initial time interval, the main source of crack wall vibration is given by the all sample vibration due to acoustic emission energy. This frequencies are given by sample geometry and frequency of EME signal was about one order lower than for initial crack walls vibration.

Conclusion

Mechanical stress creates electrical potentials on the walls of the sample. Electrical voltage is a measurable quantity for such effects and generally this is created by redistribution of electric charges, piezoelectric phenomena and by friction during crack walls sliding. There are two sources of acoustic emission events: tensile and shearing processes. Acoustic emission events appear either when new cracks are generated or by shearing process in existing cracks.

We suppose, that electric charge redistribution is a main effect, which is manifested in our experiments. Electric charge redistribution is related to time dependent electric dipole moment and then electromagnetic field must be generated. Measured voltage has two components, DC and AC given by crack walls vibration and mechanical parameters of measured sample. High frequency component is of the order form 1 to 6 MHz and give information on crack length. We estimate the length of the crack of the order of 1 mm. Sensitivity of the value of electrical potential on crack orientation with respect to parallel plate capacitor can be used to study orientation of sample's layers.

Low frequency component below 500 kHz give information on mechanical vibration spectrum and this component is correlated to acoustic emission signals. For our samples, this frequency was of the order of 100 kHz. The second time constant, which is given by parameters of mechanical vibrations and electrical conductivity is dependent on humidity.

Acknowledgement

This research has been partially supported by the MSMT of the Czech Republic under grant ME285 and also as a part of a research project CEZ J22/98:261 100007 .

References

1. T. Yoshino: Low-Frequency Seismogenic Electromagnetic Emission as Precursors to Earthquakes and Volcanic Eruptions in Japan, Journal of Scientific Exploration, Vol. 5, No. 1, pp 121-144, Pergamon Press, 1991.
2. N. G. Chatiašvili :Vozmožnye mechanizmy elektromagnitnovo izlucenija pri razrušenii kristallov i gornych porod. Geofiziceskij žurnal, 1988, t. 10, s. 45-53.
3. T. Ogawa, K. Oike and T. Miura: Electromagnetic radiation from rocks, J. Geophys. Res., 90, 6245-6249, 1985.
4. I. Zamada, K. Masuda, H. Mizutani: Electromagnetic and acoustic emission associated with rock fracture, Phys. Earth. Planet Int., 57, 157-168, 1989.
5. T.Lokajícek and J. Šikula, Acoustic emission and electromagnetic effects in rocks, Progress in Acoustic Emission VIII, 1996, pp. 311-314
6. Y. Mori, K. Saruhashi and K. Mogi, Acoustic emission from rock specimen under cyclic loading, Progress in Acoustic Emission VII, 1994, pp. 173-178
7. Y. Mori, K. Sato, Y. Obata and K. Mogi, Acoustic emission and electric potential changes of rock samples under cyclic loading, Progress in Acoustic Emission IX, 1998
8. J. Šikula, B. Koktavý, P. Vašina and T. Lokajícek, Cracks and quality testing by acoustic and electromagnetic emission, Proc. of the 23rd European Conference on Acoustic Emission Testing, TUV Austria, Wien 6-8, May, 1998, p. 130-133.

© AIPnD , created by NDT.net

|Home|    |Top|