Predicting metal failure

Dr. Neil Canter, Contributing Editor | TLT Tech Beat July 2011

A new material property can predict machinery failure even when operating conditions change.

KEY CONCEPTS
• Metal fatigue over time causes a material to gradually become more disordered, increasing its level of entropy.
• A new material property called fracture fatigue entropy can predict metal failure.
• Varying the experimental conditions used to evaluate a specific material did not change the evidence that metal failure occurred when the material’s fracture fatigue entropy was reached.

Lubricants are formulated to optimize the performance of and extend the life of machinery. Research continues to determine how machinery fails, which should provide guidance for the development of better lubricants.

In a previous TLT article, a mechanism for the formation of cracks which is considered a key reason for material failure was discussed (1). Computer simulations were used to determine that cracks move through materials in a helical fashion and can generate daughter cracks over time.

Entropy is a parameter that is used in thermodynamics. The second law of thermodynamics states that the entropy of a closed system always increases. Entropy is a measure of disorder in a system. It is known that all systems tend to move in a natural progression toward disorder.

STLE fellow, professor Michael Khonsari, Dow Chemical Endowed Chair in the department of mechanical engineering at Louisiana State University in Baton Rouge, La., says, “Entropy can be used to determine how rapidly metallic components fatigue and then predict the instance of failure.”

Khonsari adds that his group developed the idea of looking at entropy from discussions with a colleague, professor Michael Bryant from the University of Texas-Austin several years ago. He says, “We discussed the results of their experimental work showing that entropy can be used to better understand friction and wear.” Khonsari then performed a different series of experiments in his laboratory and came to the same conclusion.

This work led Khonsari to think about how to use entropy to answer the fundamental questions about metal failure. He says, “As fatigue sets in when a component undergoes cyclic stress-tension-compression, torsion and repeated bending, the material gradually becomes more disordered as it degrades. We realize that this means that the level of entropy in a machine should also increase over time.” 

This factor led Khonsari and his research group to focus on how to quantify entropy associated with fatigue degradation so that a general parameter can be developed to predict metal failure even when operating conditions change. Such a parameter has now been developed.

FRACTURE FATIGUE ENTROPY
Khonsari, in collaboration with his associates Mehdi Naderi and Mehdi Amiri, has now determined a new parameter that they call fracture fatigue entropy (FFE) to predict machinery failure. This parameter happens to be a new material property. He says, “We conducted a series of fatigue tests to see if we could verify experimentally that FFE does exist. One of the key parameters we measured was the temperature Change when a piece of metal is stressed. We know that when we bend a paper clip back and forth, its junction gets hot.

The researchers used an infrared camera to measure the heat produced when a piece of metal is evaluated in a torsion test apparatus. Figure 1 shows an image of the experimental setup.


Figure 1. The experimental setup shown was used to determine that a new parameter known as fracture fatigue entropy can be used to predict metal failure. (Courtesy of Louisiana State University)

The torsional fatigue test involved clamping a round bar specimen to both ends and then rotationally oscillating one of the ends in a cyclic fashion. The entropy of the system was calculated up until the specimen fractured.

Khonsari says, “A very interesting development occurred when we did the testing. In initial work with the aluminum alloy, 6061 T6, we found that no matter how experimental conditions were changed, the FFE value we obtained was nearly constant.”

The researchers built a specimen from the same material and tested it in a bending fatigue machine. They varied the frequency of the oscillation so that the specimen moved slower and faster. They also changed the thickness of the specimen. After all the additional work, the FFE value did not change. That is, the metal fractured when the accumulated entropy reached nearly the same FFE value.

Work was repeated with a new alloy, stainless steel 304. While the FFE value for this ferrous alloy is different from aluminum 6061 T6, the result was the same. Khonsari says, “No matter what experimental condition we varied, the result was that fracture occurred when we reached the material’s FFE.”

Khonsari indicates that his group has discovered a new material property that can be used to predict failure. He says, “We have found that when cumulative entropy reaches a certain level, the material cannot function and it breaks. This factor should enable machinery users to run their equipment up to a certain point just before failure.”

As an additional experiment, the researchers decided to run tests so that materials reach 90% of their operating lives based on calculation of the FFE. Khonsari says, “We built a prototype of an automatic monitoring system that relayed a signal to the motor to shut down when the machine reached 90% of FFE. Then we turned the machine back on and found that it failed after the FFE increased by the remaining 10%.”

In all experiments conducted, the FFE was nearly constant for a specific material no matter the stress or load applied during its operating life. Khonsari says, “The operating life of a machine varies depending upon the load, but the FFE did not change. For a low load, we found that the machine showed no visible damage after operating up to a FFE of 90%. But it did fail when the remaining 10% of FFE is reached. This means that the operating life of a machine may vary, but the entropy it accumulates over its lifetime is the same.”

Khonsari believes that entropy accumulation leads a specific material to become more disordered, which is a key element of the degradation process. He adds, “This disorder means that less of the machine is available over its operating time to perform the task.” 

The finding that FFE can be used to predict machine life can be used in virtually any industrial application. Khonsari indicates that this principle can best be used in applications that are conducted in remote locations. He says, “Drilling rigs will drill a mile to 1. 5 miles down into the earth. Machinery failure is a major problem because of the time needed to replace the failed components. Use of FFE should help to predict when a drill might fail, which should improve productivity and reduce downtime.”

Additional information can be found in a recent article (2) or by contacting Khonsari at khonsari@me.lsu.edu. 

REFERENCES
1. Canter, N. (2010), “Crack Formation in 3D,” TLT, 66 (7), pp. 8–9.
2. Naderi, M., Amiri, M. and Khonsari, M. (2010), “On the Thermodynamic Entropy of Fatigue Fracture,” Proceedings of the Royal Society A, 466 (2114), pp. 423–438.


Neil Canter heads his own consulting company, Chemical Solutions, in Willow Grove, Pa. Ideas for Tech Beat items can be sent to him at neilcanter@comcast.net.