Cracking the Fatigue Mystery
New Machine Simulates Repetitive Stress to Study High-Cycle Fatigue
Understanding metal fatigue, a mysterious and potentially catastrophic phenomenon that can destroy an aircraft engine or shatter an industrial turbine, is made all the more difficult by the lengthy study periods and complex equipment needed to duplicate, observe, and measure it in a laboratory.
Yet, scientists and engineers at Southwest Research Institute (SwRI) have developed a machine* that accomplishes just that by combining high-frequency applications of precisely measured stress with the ability to view the process and results in a scanning electron microscope (SEM). Commercial laboratory equipment capable of such studies presently does not exist.
"High-cycle fatigue" (HCF), so-called because it appears after millions of repetitive cycles of use, can affect even high-strength metals such as the titanium alloys used in the gas turbines that power aircraft or large electric generators. The engines must endure repeated, long-term exposure to high pressures and temperatures, with rotation speeds of 7,000 to 10,000 revolutions per minute. Stress levels alter with changes in engine speed, and because the blades act as miniature airfoils they are also subject to aerodynamic vibrations as they pull air through the engine.
Faced with these operating environments, blade and disk materials must be able to withstand these very high cyclic stresses. HCF cracks can be initiated and grow at stress levels that are low in relation to the material's yield stress, which is the design stress used for static components.
Because of the very large number of cycles involved (between 107 to 109), laboratory study of HCF failure phenomena must be carried out at high frequencies to characterize the behavior of the cracks within reasonable time limits.
To address this important industrial problem, scientists at SwRI, using internal research funds, designed and built a unique high-frequency machine to study HCF failure characteristics. One of the machine's special features is that it is small and light enough to fit within the sample chamber of an SEM.
The machine applies stress to a specimen under study, such as a piece of turbine blade material, by two means: a steady stress is applied by hydraulic pressure that simulates the centrifugal forces in the engine, and at the same time high frequency stresses are applied by piezoceramic plates that cause the machine as well as the specimen to resonate between 1,000 to 1,700 cycles per second (1-1.7 kHz).
The resonance conditions amplify the loads generated by the piezoceramic plates and are a second important design feature of the machine. Static loads of up to 26,700 Newtons (N), or 6,000 pounds, and dynamic loads of 5,400 N or 1,200 pounds, can be applied to a specimen. These very large loads mean that specimens with cross-sectional areas similar to an actual turbine blade can be studied.
A resonance frequency of approximately 2 kHz was chosen for the machine because it allows experiments using cycles of up to 108; to be conducted in a 40-hour work week, with time included for research and analysis of the changes in specimens under study.
The machine was designed to fit within an SEM so that the depth of field and high resolution of that instrument could be exploited to investigate the growth of fatigue cracks that can occur under HCF conditions. Examining and measuring the growth of these cracks is a slow and difficult process. For example, average crack growths of one-tenth of an atomic spacing are common. Fatigue cracks growing at these rates have never been fully characterized before and it is expected that knowledge of how these cracks develop and interact with the microstructure of the material will provide useful information for the manufacture, use, and inspection of a variety of advanced engine materials.
The HCF machine is the third in a series of innovative and successful instruments designed and built by SwRI engineers and scientists for fundamental research on material fatigue and fracture. It complements a cyclic loading machine, built in 1978, that operates at tensile loads of up to 4,000 N (1,000 pounds) and up to 4 Hz, and a high-temperature machine, built in 1985, that can operate at the same loads and at temperatures of up to 850 degrees C. Both of these machines also are operated in conjunction with the SEM so that the crack tip can be imaged at magnifications of up to several thousand times.
As one result of the unique ability to photograph cracks under various loading conditions, SwRI scientists have developed a technique, termed stereoimaging, to measure the material response, or the displacements, within the material near the crack tip.
The stereoimaging technique was automated in 1992 by incorporating image processing. The automated stereoimaging technique is called DISMAP (displacement mapping). Using DISMAP, the micromechanics of crack tips that are actually growing can be determined at high resolution. As a result, the loads required to open the crack to the tip can be accurately measured, as can the crack opening displacement which is often less than one micrometer. This measures the elastic and plastic response of the material to the loads being remotely applied to the specimen. All these micromechanical parameters can then be used to define the fracture mechanics of the material under fatigue loading conditions.
The HCF machine, used in conjunction with the SEM, extends previous technology and facilitates study of the micromechanics of crack formation and growth into the very difficult but important area of low crack growth rates. Already, observations using the machine have shown that fatigue cracks grow only intermittently. Indeed, some cracks, after an initial appearance, remain dormant for millions of cycles before growing again. In addition, it has been found that fatigue cracks that were thought to be open at minimum load under conditions of high static stress are, in fact, usually closed at the minimum cyclic load, and don't open until about half the maximum load is applied. It has also been learned that the region of plastic material response at the crack tip under the very low growth conditions of HCF is very small, not exceeding approximately 2 micrometers. These findings may appear modest but they are critical to understanding more about the character of HCF.
Although the HCF machine was designed to operate under conditions similar to the compressor section of a jet engine, it has many other potential uses, such as in power turbines for electric generators which use similar technology.
While the materials used in manufacturing are generally steels rather than titanium alloys, they suffer from many related design and operational material problems. Automobile engines, for example, operate under HCF conditions, and the machine could be used to study fatigue in valve spring and connecting rod materials. There are also potential biomedical applications as some parts of the body, such as heart valves, cycle billions of times during a lifetime. (See Smart Skeletons.)
The successful design and operation of the high cycle fatigue machine has led to its use as a prototype model for a laboratory machine now under construction. This laboratory instrument will be used for continuous and heavy-duty operation, will be capable of tackling a variety of problems, and will, therefore, have a more sophisticated control system.
*Patent pending for high-cycle fatigue test machine
Staff Engineer Andrew Nagy, SwRI Consultant and former employee Dr. Thomas Owen, and Engineering Technologist John B. Campbell designed, built, and developed the operational protocols for the high-cycle fatigue machine. The contributions of P A. Cox, Forrest S. Campbell, and Art Nicholls are also recognized.
D.L. Davidson, A. Nagy, T.S. Owen, "A 2 kHz SEM Loading Stage, High Cycle Fatigue of Structural Materials," Symposium Proceedings in Honor of Professor Paul C. Paris, p. 263-269, 1997.
Published in the Fall 1998 issue of Technology Today®, published by Southwest Research Institute. For more information, contact Joe Fohn.