Aircraft turbine engines are prone to ingesting pebbles and other debris that can damage jet engine turbine blades, dramatically reducing the longevity of the components, sometimes catastrophically.

(Graphic Courtesy of Apurva Mehta)

Failures associated with such ‘foreign object damage’ cost the aerospace industry an estimated $4 billion a year. Studies at SSRL have helped show how and why turbine blades—which normally experience significant stresses during flying—will fatigue more than 100 times sooner than expected from foreign object damage.

"The results of our study are important for people on the tarmac inspecting the blades, looking for damage and deciding whether that pinhead-size ding has damaged the blade catastrophically," said SSRL scientist Apurva Mehta (ESRD).

Working with principal investigator Brad Boyce (Sandia), Mehta and collaborators simulated the damage by firing small hardened steel balls onto a titanium alloy commonly used in turbine blades. At LLNL, these ball bearings were fired at 200 meters per second (m/s), or 450 miles/hour, and at 300 m/s, the typical velocities at which runway debris encounters turbine blades when planes take off and land.

At SSRL, the team examined the resulting damage with the unique abilities of synchrotron mesodiffraction (x-ray diffraction in the sub-millimeter scale); in this case, 0.3 mm to match the size scale of the damage.

"We wanted to see how things fail, the physics behind what happens. At 300 m/s the mode of failure is qualitatively different, and we wanted to understand why," Mehta said.

The x-ray images revealed the magnitude and distribution of residual stresses in and around the craters made by the ball bearings’ impact. For example, the material was under tension in the center of the crater, under compression in a ring around the center, and under normal stress beyond the crater.

In 200 m/s craters, the actual stresses measured in the lab matched calculations from the finite-element models that engineers use to estimate residual stress. This shows that the models are dependable in predicting how much longer the dinged blades should last for impacts at 200 m/s and below. The team was surprised to find that the 300 m/s craters actually showed lower residual stresses than the 200 m/s craters. "It didn’t match our intuition or the calculations’ predictions at all," said Mehta.

Careful examination of the 300 m/s craters revealed large variations in residual stresses. The regions of lower stress often had micro-cracks created by the impact. Mehta and his colleagues believe these relieve the majority of the residual stress. It also explains the models’ off-base predictions at 300 m/s. Micro-cracks, however, are the weak point in the material. Subsequent mechanical tests have shown the material invariably fails (and fails faster) there.

The good news is that not all impacts are harmful. The team found that residual stresses around a 200 m/s crater located in the central part of the blade goes away after a few cycles. However, impacts near an edge or an angle in the blades leads to unrelieved strain on sides, edges and corners that are not relaxed on cycling and become sites of fatigue failure.

"Our study tells maintenance crews to look not only for the size of the impact crater, but for its location and for signs of micro-cracks," Mehta said.

The funders of the study, the Lufthansa Technik AG, the U.S. Air Force and the U.S. Army, were excited about the results. Turbine blades are expensive, and the new data give more clues about which craters are dangerous and which might ‘self-heal.’

This new understanding has been incorporated into a mathematical model of failure to help design new blades to prevent failures resulting from foreign object damage.

Boyce shared the results in a workshop last month called "Probing Mechanical Deformation and Failure via Synchrotron X-ray," during SSRL’s 30th Annual Users Meeting.

"This example illustrates the utility of a synchrotron x-ray source to solve real-world engineering problems," said SSRL scientist Mike Toney.