Sprayed-on foam insulation covers the space shuttle external fuel tank and is necessary to prevent ice buildups as the shuttle sits awaiting liftoff with cryogenic propellants in the external tank. Flaws present in the foam can result in initiation sites for foam loss. Manually sprayed foam areas are especially prone to flaws. A major effort is underway at NASA to find and improve inspection methods to locate cracks, voids, delaminations (between the foam and the underlying tank, or between the foam and knitlines), and crushed areas of the foam in order to minimize foam debris release upon launch.
The latest external tank for the July 2006 shuttle launch (STS-121) had the protuberance air load ramp windshield removed to minimize large foam debris release upon launch. As a result, the highest priority areas to inspect are now the ice frost ramps, which prevent ice from forming on underlying aluminum brackets used to fasten fuel-pressurization lines and on a tray of electrical cables to the tank’s exterior. It is impossible to completely prevent foam from coming off of the ramps during launch, but inspection methods can help locate damaged ramps prior to launch and also aid in the understanding of the conditions and flaws that cause foam release. (Future design changes also include possible removal of the ice frost ramp.)
Foam inspection methods in use and being researched for improvements at NASA include terahertz, microwave, shearography, and x-ray backscatter. Terahertz inspection in the reflection mode requiring access to only one side of a material has shown significant promise for the detection of voids and crushed foam. The schematic shows the overall operation of this method. Briefly, terahertz waves are electromagnetic waves with wavelengths on the order of 200 to 1000 μm (just shorter than those in the microwave domain). Electrically conducting materials such as metals reflect terahertz waves, whereas dielectric(nonconducting) nonpolar liquids, nonmetallic solids, and gases are transparent to terahertz energy. Reflections also occur off of interfaces, such as between a solid and air, where a dielectric discontinuity (difference in the indices of refraction) occurs. Thus, small reflections will occur off of voids in foam and the signal transmitted through the void will have reduced signal amplitude. Significant time- and frequency-domain information is available in these reflection signals.
Reflection-mode terahertz methodology. Reflections will be received off of the various interfaces. Reflection from metal will be the strongest. The horizontal dotted line over the large, initial portion of the wave shows the time gate of the echo typically used during signal processing. A sample power spectral density is shown with the centroid, fC, denoted. FSH, full scale height.
Currently applied signal-processing techniques for terahertz data are generally based on simple parameter extractions such as peak detection of the time-domain signal or the Fourier-transformed signal. The focus of this research effort is to develop signal analysis techniques that capitalize on all the information that is available in the terahertz signal to improve the resolution of flaws and differentiation of flaws from normal foam variations such as knit lines.
This past year, the multicenter effort focused particularly on the computation and image display of the centroid of the power spectral density for the reflections off of a metal substrate located beneath foam (the configuration of such a sample simulates the external tank configuration) for foam samples containing seeded voids. The centroid images were compared with those formed from traditional waveform parameters such as peak-to-peak amplitude (time-domain) and peak magnitude (frequency-domain). The centroid is reasonably simple and rapid computationally, and proved to significantly enhance flaw resolvability in foam samples containing seeded voids of various sizes and depths as shown in the following figure.
Comparison of different terahertz images including line profiles (right side of figure) at the location of the bottom (smallest) row of voids. The voids are relatively deeper as one moves to the right in the image.(a) Time-domain peak-to-peak amplitude. (b) Power spectral density peak magnitude. (c) Power spectral density at 0.16 MHz. (d) Centroid, fC, of the power spectral density.
The product of this study is a commercial-grade software package that is being used both by Michoud Assembly Facility and NASA Glenn Research Center personnel to analyze data and visualize terahertz (and ultrasonic) images. The software was developed by Glenn in cooperation with the NASA Langley Research Center, Michoud Assembly Facility, NASA Marshall Space Flight Center, and Cleveland State University.
Glenn contact: Don Roth, 216-433-6017, Donald.J.Roth@nasa.govLast updated: December 14, 2007
Responsible NASA Official:
Gynelle.C.Steele@nasa.gov
216-433-8258
Point of contact for NASA Glenn's Research & Technology reports:
Cynthia.L.Dreibelbis@nasa.gov
216-433-2912
SGT, Inc.
Web page curator:
Nancy.L.Obryan@nasa.gov
216-433-5793
Wyle Information Systems, LLC
