Nondestructive Testing of Aging Aircraft

New technology will allow better detection of fatigue cracks that can lead to deadly crashes

By Gary L. Burkhardt and Jay L. Fisher, ScD

A team of Southwest Research Institute engineers is developing more effective methods of detecting cracks in airplanes that can lead to deadly crashes -- an increasingly urgent concern as the service life of aging aircraft is further extended.

Several types of U.S. Air Force aircraft already have exceeded their designed 20- to 30-year lifetimes; the Air Force plans to extend some airframe service lifetimes up to 80 years. For example, the 1950s-vintage B-52 bomber and KC-135 tanker will operate until 2030 and 2040, respectively. The nation's commercial aircraft fleet also is aging as airlines curb capital equipment costs by extending the service lives of their aircraft. Aging can result in widespread fatigue cracking of the aircraft's structure. Cracks emanating from fastener holes in the outer skin of the wing and fuselage constitute a major safety issue. The in-air skin failure 13 years ago of an Aloha Airlines jet was the result of skin failure produced by cracks that formed from many small rivet holes. Aloha Airlines Flight 243 was en route to Honolulu on April 28, 1988, at 24,000 feet when a large section of the fuselage was ripped from the Boeing 737 airliner. Miraculously, the airplane landed safely, and the incident resulted in only one fatality. The National Transportation Safety Board accident report noted that a passenger boarding the airplane noticed a long crack in a row of rivets above the cabin door but did not tell the flight crew before takeoff.

The eddy-current testing (ECT) method of nondestructive evaluation (NDE) can detect tiny cracks, which must be repaired before they become large enough to threaten the safety of an aircraft. The most common technique involves systematically moving a probe around each fastener. This process is time consuming because of the large number of fasteners involved. For example, the fuselage of a C-5 cargo plane has more than 500,000 fasteners.

Although some new ECT technologies are in use, some are slow and difficult to set up and calibrate. One ECT technique can locate small cracks but cannot reliably detect cracks that connect two fastener holes. This shortcoming is of major concern because these connecting cracks can lead to the kind of catastrophic damage that caused the Aloha Airlines incident.

SwRI's internal research and development program has funded work by engineers in the Institute's Applied Physics Division who proposed development of a more effective ECT method for detecting these dangerous cracks.

The current flow (I) direction shown by the arrows alternates in opposite directions for each adjacent conductor line in the meander coil. The meander excitation coil generates eddy currents in the part under inspection, and flaws are detected using sensors arranged so as to be sensitive to magnetic fields (B) parallel to the inspection surface.

Background

ECT systems use alternating magnetic fields to induce electrical current flow into a test specimen. This is usually accomplished using single or multiple electromagnetic coils, with the coil axes perpendicular to the surface under inspection, or a sheet conductor that induces current flow over a large area. SwRI researchers believed that a meander coil system could detect cracks of all sizes consistently over a relatively large area at one time.

A meander coil consists of a strip of conducting metal that repeatedly doubles back on itself, resulting in a series of parallel metallic pathways. An electrically activated meander coil generates eddy currents in the inspected part, and the magnetic fields from these currents are detected by giant magnetoresistive (GMR) sensors, solid-state devices that change resistance in response to a magnetic field. These GMR sensors are arranged to sense the fields parallel to the inspection surface. Where no crack exists, there is no field component in this direction, and no direct coupling occurs between the main field and the sensor. If a crack does exist, a magnetic field is produced in this parallel direction. This parallel field is generated in the inspected component by eddy currents, whose paths are disturbed by the crack.

The bottom of the small-scale array probe shows the sensors positioned in holes cut between the meander coil traces.

Modeling

As a first step, the research team developed a three-dimensional electromagnetic model to optimize the meander coil design. The software model calculated the eddy-current distribution in the inspected material and the associated magnetic field distribution above its surface. The team used the resulting information to determine the orientation and location of the receiver sensors and to design the meander coil. The modeling showed that a meander coil device would produce detectable flaw responses for short and long cracks. This device would thus be superior to certain conventional sensors that have difficulty detecting long cracks.

The team built a prototype probe with a single GMR sensor based on promising modeling results. The meander coil was fabricated with conductors spaced close together on printed circuit board material. A GMR sensor was placed against the meander coil with the sensor's active area positioned between two conductors. Engineers bonded these components to form a probe package and placed a bar magnet on the package to provide the magnetic bias field required by the GMR sensor.

Engineers used the probe to systematically scan several aluminum test panels. The panels contained fastener holes with no cracks as well as flawed holes with induced notches to simulate cracks. The flaws that connected fastener holes showed up well, as did those that were not connecting. The probe was also scanned over a plastic shim fastened to the panel's surface to simulate paint that frequently separates a sensor from the metal part under inspection. These separations can result in high electromagnetic noise and reduced crack detection capability in conventional ECT devices, but the meander probe sustained almost no negative effect from this "lift-off" condition.

Small-scale array probe

The research team then prepared a probe with several GMR sensors to demonstrate the feasibility of an array probe and to gain experience with fabricating an array. Engineers employed the same coil configuration as with the single-probe sensor, but they incorporated four GMR sensors in die form (a bare device with no package).

The team evaluated the array probe for sensitivity and spatial resolution. The probe was scanned over the same panel used in the single-probe tests. The probe readily detected connecting and nonconnecting flaws, even though the resolution was reduced because the 1.2-mm spacing between the array sensors is greater than the 0.5-mm spacing between scans used for the single sensor. In a test with an unflawed hole and holes with 3.2- and 1.6-mm flaws, the probe detected the 3.2-mm flaw. To detect the 1.6-mm flaw clearly, the sensor spacing would need to be reduced to approximately 0.75 mm.

Conclusions

SwRI staff demonstrated the feasibility of meander probe technology. The Institute is seeking external funding to develop the sensor probe further. Future development could involve an array with more elements and containing on-board electronics. Externally funded development may lead to a chip containing sensors and integrated electronics.

Comments about this article? Contact Burkhardt at (210) 522-2705 or gburkhardt@swri.org.

Published in the Fall 2001 issue of Technology Today®, published by Southwest Research Institute. For more information, contact Maria Martinez.

Technics Fall 2001 Technology Today
SwRI Publications SwRI Home