Failure may be an opportunity for growth, but I don’t want to be anywhere near the collapsing bridge or malfunctioning airplane that everyone else learns from. When it comes to structural failure, the best place to learn about it is in the lab and the best time to detect it is well before it happens.
|
|
| Cars rest on the collapsed portion of I-35W Mississippi River bridge, after the August 1st, 2007 collapse that killed 13 and injured 145. One of the steel components was likely too thin and failed under stress. Image Credit: Kevin Rofidal, United States Coast Guard. |
Last week, at the 253rd National Meeting & Exposition of the American Chemical Society (ACS), a team of scientists from Brigham Young University presented a new method for detecting weak spots in metal structures. Initial studies show that it can detect earlier warning signs than other nondestructive testing methods, in addition to being portable and less expensive.
The standard way to predict when a metal part will fail is to make it fail—for example, to bend it repeatedly until it breaks or expose it to colder and colder temperatures until it cracks. Do this enough times in the environment that the part will be in, and you can calculate its average performance. The average performance can then inform guidelines on how often that piece should be replaced. The average performance doesn’t tell you when a particular component will fail though—some will fail much sooner than the average and others much later.
There’s a big downside to this method: It requires breaking something. If you want to test the integrity of the metal lining a ship hull or airplane wing, it’s best to avoid cutting out a sample for testing. This is where nondestructive testing comes in, a billion-dollar industry that looks for evidence of weakness without stressing or destroying a material. Existing nondestructive techniques, such as X-ray imaging, can be expensive, cumbersome, require highly trained technicians, and have limited sensitivity—by the time they detect microscopic cracks in a piece of metal, the breaking point is imminent.
In James E. Patterson’s lab at Brigham Young University, he and students Shawn Averett, Alex Farnsworth, Kaylee Rellaford, and Scott Smith are exploring another option: predicting failure with green laser light.
Metal contains defects called dislocations, places where the planes of the metal’s atomic lattice become irregular, rather than repeating perfectly. The dislocations are initially spread throughout a metal structure, but when exposed to stress they migrate toward the surface of the metal. According to the researchers, the density of these dislocations on a metal surface appears to be correlated to the microscopic fractures that precede failure.
When you shine a green laser on a metal surface, some of the light is reflected back and some of it is converted into ultraviolet light through a process called second harmonic generation, or frequency doubling. The amount of green light that is converted into ultraviolet light depends on the conditions on the surface, including the dislocation density. If you know your surface well, you can determine the dislocation density from measurements of the ultraviolet light. Because dislocations migrate under stress, this means that you can identify areas under high stress and take action before it’s too late.
Although this research is in its early stages, results look promising. Experiments show that this method can differentiate between damaged and intact metal parts. Preliminary results suggest that it and similar methods can also be used on composites and plastics, and can even detect some plastic deformation that current nondestructive testing techniques can’t see. This technique could potentially verify the purity of metals and uniformity of alloys, and detect invisible corrosion in naval ships. It can be run through fiber optics and incorporated into a hand-held tool, or even an automated probe that can climb into hard-to-reach places, enabling easier, cheaper, and more effective inspections.
The extreme sensitivity of this technique is one of its strong points, but it’s also one of the challenges. Measuring the ultraviolet light won’t tell you anything useful unless you can separate the noise from meaningful data, i.e. the surface changes caused by other influences from the changes caused by dislocations. As this work moves forward, the team is mapping signal changes to their physical causes and making sure that this method is valid and reliable. In light of the aging infrastructure in much of the world, hopefully their quest to predict failure will be a resounding success.
—Kendra Redmond