EMATs for on orbit wall remaining thickness measurement after an impact - Feasibility studyG. Ithurralde, D. Simonet
Aérospatiale Corporate Research Center Louis Blériot
DCR/MI, 12 rue Pasteur, 92152 Suresness Cedex, France; 33 1 4697 3216; fax 3730
Aérospatiale, Space and Defense Division
66 route de Verneuil, BP 3002, 78130 Les Mureaux, France; 33 1 3492 2021
Corresponding Author Contact:
|TABLE OF CONTENTS|
Space nondestructive testing (NDT) looks likely to grow if important projects such as the International Space Station (see Figure 1) are achieved. These kind of orbital facilities may collide with fragments coming from the 11000 previous satellites launched since Sputnik in 1957. An estimated three thousand tons of fragments are orbiting the Earth with an average satellite speed of 11 km/s (6.85 mi/s) and a density of about 2.8 g/cm3 (175 lb/ft3).
The European Space Agency (ESA) wants to have practical NDT tools for diagnose and repair of collision damage. This is the only way to ensure a lifetime of 15 years for the European ISS module, the Columbus Orbital Facility (COF). Hurling projectiles onto aluminum plates demonstrated that a large variety of defects are created after a high-speed impact. In particular local thinning down was observed at the center of craters (see Figure 2).
Conventional UT with piezoelectric transducers is not applicable for thickness measurement in space environment because the vacuum and wide variations in temperature do not permit the use of any liquid couplant. For this reason electromagnetic acoustic transducers (EMATs) are a good alternative. Our objective was to prove their efficiency under space conditions.
|Fig 1: Drawing of the Columbus Orbital Facility|
|Fig 2: Top view of a crater after an impact|
A probe was specially designed and manufactured to generate and receive 0° shear waves vibrating between 3 MHz and 4 MHz. It was tested in a homemade temperature regulated vessel, between -150°C (-238°F) and +170°C (+338°F).
The results were encouraging since thickness was measured down to 1 mm (0.04 in.) with satisfactory accuracy on a calibration block with steps made of aluminum alloy (2219 T6). On the one hand, the influence of temperature was not as important as expected. On the other hand signal amplitudes were very sensitive to the lift-off. Miniaturized flexible circuits instead of wound coils should solve the lift-off problem.
Space standards and other requirements such as maximum power consumption, procedure, etc. have to be considered and integrated into the design of the probe and the driver before planning to inspect a space structure on orbit. However it has been demonstrated that EMATs can operate in space environment.
|Fig 3: Cross view of the probe showing design and principle of shear wave generation|
Figure 3 shows the four superposed elements of the probe:
Samarium cobalt was selected for the magnet because this material provides strong magnetic fields and it is not demagnetized until it reaches about 350°C (662°F). Type 28 was the highest grade found on the market for samarium cobalt magnets.
A brass mask was inserted between the magnet and the transmitting coil in order to avoid the generation of ultrasound in the magnet. Its thickness was 0.1 mm (0.004 in.).
The transmitting and receiving coils were hand wound separately. The copper wire is heat insulated by a polyester amide imide coat. Selected wire diameters were 35 AWG for the transmitter and 37 AWG for the receiver. The external coil diameter is approximately 17 mm (0.66 in.). Coils were soldered to two insulated coaxial cables RG 178B/U.
These four elements were bound together with epoxy glue withstanding temperature above 180°C (356°F), in a polyetheretherketone case.
Inducing a circular alternative eddy current density in a vertical magnetic field produces radial oscillating Lorentz forces, in a skin depth inferior to 0.1 mm (0.004 in.). These forces act like an ultrasound source inside the material itself. Receiving follows the opposite mechanism.
|Fig 4: Set-up for thickness measurement under simulated space conditions|
The installation was composed of:
The EMAT driver generates high voltage square bipolar tone bursts. The frequency and the number of cycles in tone bursts are adjustable (from 0.1 to 4 MHz and from 1 to 10 cycles, respectively). The optimal values ranged between 3 and 4 MHz, and 1 and 3 cycles. Alternative current pulsed in the transmitting coil was about 17.5 A peak-to-peak. No cooling was needed.
To maximize signal-to-noise ratio, received signals were electronically filtered, rectified and smoothed. Furthermore, the software averaged 16 acquisitions.
The calibration block is made of the same aluminum alloy as the COF external structure, that is, alloy 2219 T6. Several steps were machined with thickness ranging between 1 mm (0.04 in.) and 5 mm (0.2 in.).
Fig 5: Time of flight measurement on a rectified A-SCAN at 20°C
Fig 6: Relative differences between measured and expansion corrected values between -150 and +170°C
Tests were performed at nine temperatures between -150°C (-238°F) and +170°C (+338°F). Thickness was deduced from the known 20°C (68°F) velocity and the measured time of flight difference between two successive back wall echoes, as shown in figure 5.
Lift-off influence was also studied at 20°C (68°F), by inserting thickness gauges between the receiving coil and the calibration block.
The ultrasound wavelength was too large for the step of 1 mm (0.04 in.) to be evaluated. Larger thickness was measured with satisfactory accuracy (lower than six percent) and signal-to-noise ratio, even at +170°C (+338°F) when electrical conductivity, thus eddy current density, decrease.
Relative errors are given in Figure 6. Their calculation took into account theoretical temperature expansion. Thickness is overevaluated at high temperature and is underevaluated at low temperature. These differences may be explained by changes in shear waves velocity due to temperature variations.
Lift-off has a dramatic influence on signal amplitude. With a gap between the receiving coil and the calibration block greater than 1 mm (0.04 in.), echoes are not strong enough to exceed noise. This will cause problem for thickness measurement at the center of deep concave craters unless the probe diameter is minimized. A solution may consist in replacing winded coils by flexible circuit coils (obtained by photolithography or similar process) for the active surface to fit with the crater surface.
Temperature variations are not a problem if EMATs are used for wall thickness measurement purpose on the COF, in space environment. Results demonstrate that a probe generating 0° shear wave provides values with relative errors lower than ± 5% between -150 (-238°F) and +170°C (+338°F), for a thickness 1-5 mm (0.04-0.2 in.).
Development should be continued to optimize the probe, particularly regarding the lift-off effect, and to adapt the whole system to space standards and procedures.
EMATs might also be used for crack detection and sizing, for example with probes generating surface or guided waves. But other NDT tools such as eddy currents could prove to be more efficient.
Appreciation is gratefully acknowledged to R/D Tech personnel for their constant assistance.