Electro Magnetic Acoustic Transducers (EMA) is a well known type of ultrasonic probes used for non-destructive testing (NDT)

of electrically conductive materials. Conventional Ultrasonic Testing (UT) is performed by longitudinal waves and mode converted shear-waves generated by piezoelectric probes. EMAT's are broadening the range of usable wave modes by the direct conversion of polarized shear waves with normal and angle beams and the selective generation and detection of nearly all types of guided waves. Despite their limitations (low efficiency, lift-off sensitivity, limited frequency range etc.) they have the big advantage to perform UT without couplants. The dry coupling allows UT at elevated temperatures, in media which do not tolerate liquids (e.g. natural-gas pipelines) or on sensitive and coated surfaces of blanks used for car bodies in the automotive industry etc. NDT is not only related to defect detection but more and more to material state and condition analysis and monitoring. To this end, polarization, dispersion, propagation velocity, and mode conversion are physical parameters usable for improved and high sophisticated non-destructive measurements. For inspection of a storage tank and pipeline in service, the application of an automatic inspection system (non-destructive inspection robot) is desirable, because manual inspection is difficult to perfectly and exactly perform due to the enormous amount of inspection needed. Development of an ultrasonic inspection robot with an electromagnetic acoustic transducer (EMAT) which did not require a coupling medium to inspect the circumferential pipe parts was conducted. They developed a special EMAT that could transmit and receive alternately a Lamb wave with high sensitivity. Generally, an angle beam method is used as the non-destructive inspection method of welded parts, but precise depth and lateral scanning is absolutely required to cover the entire welded part. In addition, we must control any ultrasonic transducers so that they touch with constant pressure on the surface of the steel pipe. Development of an ultrasonic non-destructive inspection robot which carried an EMAT that could alternately drive a Lamb, because an EMAT can theoretically transmit and receive ultrasonic waves without any coupling medium. Development of an EMAT for Lamb wave The basic oscillation pattern of a fundamental plate has a perpendicular oscillating direction in the advancing direction, and is parallel to the material surface. Therefore, it is not influenced by the surface condition of the sample plate. Next, the oscillating components of the S0-mode Lamb wave mainly consists of the parallel direction to the advance direction. Therefore, it is easily influenced by the surface condition of the sample plate and that the distance attenuation is large. The EMAT consists of a magnet that produces a bias magnetic field, and a sensor coil that produces a dynamic magnetic field. The driving force uses a high frequency vibration of magnetostriction generated in the direction of the compounds magnetic field by combining the dynamic magnetic field generated by a high frequency electric current in the sensor coil, and the static field by the electromagnet.

Although the direction of the magnetostrictive change occurs in the direction of the slant to the direction of a sensor coil, the power ingredient, which causes the magnetostrictive change occurs in the space frequency to the direction of a sensor coil and in a perpendicular direction, and is considered to be changed into the SH0-mode plate wave. Next, in the case of S0-mode Lamb wave, the change of magnetostriction occurred due to the compounds generated magnetic field is in a parallel direction to the traveling direction. This magnetostrictive change was converted into the S0 mode Lamb wave. The experimental method The sensor coil was incorporated into a structure, so that the interval between the lead lines is 3 mm, this sensor coil produced a plate wave of 6 mm wavelength. There are two sensor coils for the receiver and transmitter. These two sensor coils were placed parallel to the travelling direction of the plate wave by having a 30 mm distance between both magnetic poles. These sensor coils and two electromagnets are connected to the experimental system. As a result, the magnetic field distribution in part of the sensor coil induced by one electromagnet might be influenced by the other electromagnet. When two electromagnets were on the plate, the T-direction magnetic field density decreased by 2% and the Vdirection magnetic field density increased by about 20% compared to the magnetic field density when only one electromagnet was used. We considered that this result does not significantly influence the measured received signal by the SH0 and S0-mode plate waves. Basic performance of the EMAT In order to check whether an alternating generation of two plate wave modes could be realized by the same transducer, the following evaluation tests were then carried out. The relation between the static magnetic field and the received signal amplitude was measured for 0.6, 1.0, 2.0 and 6.0 mm thick plates. It is observed that the optimum magnetic current of the SH-plate wave is different from that of the Lamb wave. This phenomenon could not be simply expected because the EMAT for the S0 and SH0-mode plate waves use the same magnetostrictive effect. It might be that the drive source of the Lamb wave was the slope of the magnetostrictive curve (MagnetostrictionMagnetic field) and that of the SH0-mode plate wave was the strength of the magnetostriction curve. The basic characteristics of Lamb wave From the experiment, the signal amplitude of both modes decreased as the plate thickness increased. This feature is effective for inspection on a manufacturing line where the plate thickness is always variable. The S0-mode Lamb wave remained almost constant on the reflected signal amplitude from the drilled hole and the diameter. The signal amplitude increased for the Lamb wave as the surface condition became rougher.

Evaluation of Lamb wave modes Four selected Lamb wave modes was evaluated for detection of flaws in a steel pipe specimen. They were first tested at the distance of 304.8mm from one end of the pipe specimen to determine energy loss. The A0 Lamb wave mode was evaluated first. the signal response from one end of the pipe specimen is strong and sharp. This was determined by placing an ultrasonic couplant in front of the hole and damping the signal with a fingertip. If the transducer was aligned with the hole circumferentially, signals from other flaws were detected. As the transducer was moved axially along the pipe specimen, the circumferential location of the transducer in relation to the flaw signal would change. When the transducer was positioned axially within 121.9 cm from the flaws, the circumferential position for the signal reflection was aligned with the flaws. When the transducer was positioned axially farther than 121.9 cm, the circumferential alignment began to change. Further investigations revealed that the transducer beam divergence of generated Lamb waves completes one wrap 360 around the pipe specimen. This explains why all of the flaws were seen from a distance of 121.9 cm. When the transducer was positioned at distances farther than 121.9cm, the beam diverged Lamb waves then began to interact with axial waves. This condition causes the Lamb waves to have high and low energy points circumferentially. The second Lamb wave mode to be evaluated was the So mode. The 0.5MHz transducer was mounted on the variable angle lucite wedge. The incident angle was set to 50 to generate the So mode. The transducer was positioned on the pipe specimen at a distance of 304.8 cm away from one end of the pipe specimen containing artificial flaws. The end reflection was very weak and could very easily be damped on the surface. Some of the larger holes could be detected but were easily lost in the noise. The smaller signals around screen division 4 were water drops on the surface of the pipe specimen. When the transducer was located at a distance 91.4cm from one end of the pipe specimen, that is, a distance 68.5 cm from the centre of flaws, only three were detected. These reflections were not detected from six holes but from three holes because the beam divergence does not wrap around the entire pipe once. The final Lamb wave mode to be evaluated was the St mode. The variable angle lucite wedge was adjusted to the incident angle of 34. The transducer was positioned at the distance of 304.8cm away from the end containing the artificial flaws. Similar to the At mode, two signal reflections from one end of the pipe specimen were seen.

Conclusion An ultrasonic inspection mechanism equipped with an EMAT that alternately excites the SH-plate wave and the Lamb wave alternately for steel pipes was developed. Mutual excitation was found and fundamentally possible. A Lucite wedge containing a water column that generates the Ao Lamb wave mode has been developed and used for inspection of the steel pipe. However, the experimental results indicated that different drive conditions are required to drive the Lamb waves. We also confirmed that it is effective to combine the information of the received signal from ultrasonic modes with respect to the detection of a defect. However, further improvement of the detecting ability of the system is required, because that ability decreases, as the sheet thickness increases.