When Metals Fail Due to Fatigue
SUDDENLY, without warning, death and destruction struck at the heliport atop the 59-story Pan Am Building in midtown Manhattan, New York city. As a passenger helicopter was being boarded for a flight to John F. Kennedy International Airport, the craft keeled over. In seconds, its rotors, like giant scimitars, slashed four passengers. Three died on the spot and the fourth died in a hospital. The flying blades fragmented, scattering pieces over a wide area. Fragments falling on Madison Avenue killed one woman and injured another. What caused this disaster? A report of a preliminary investigation indicated that metal fatigue was involved.
Consider another recent accident. Two women were driving on the beautiful over-the-seas highway in the Florida Keys. The car suddenly careened across the road and plunged into the sea. Fortunately, a diver and a medical doctor were on hand and the women were saved. Part of the auto’s steering gear had failed. Why? A cursory examination revealed the telltale marks of metal fatigue.
The New York heliport tragedy, the famous Silver Bridge collapse, the mysterious disappearance of the early British Comets over the Mediterranean—all these incidents are thought to have been related to metal fatigue.
The consequences of fatigue damage in metals, like cancer in humans, may be less serious if there is early detection. Also, as with cancer, the cures are often difficult and sometimes unsure. Unfortunately, the identifying marks frequently are obliterated by the accident. More often they go unrecognized because of the lack of specially trained investigators.
The Structure of Metals
To understand fatigue in metals, we must look at their structures. From the time of Tubal-cain, the first historical worker in metals, down to our day, there has been no satisfactory explanation for fatigue in these substances. (Gen. 4:22) Only in the recent past has knowledge of the basic structure of metals developed sufficiently to provide a plausible understanding. Even today, when a leaf spring or axle fails, we still hear people say that the part has crystallized. This could not have been the cause of failure, since the material already was crystallized before the failure.
When molten metals begin to cool down and solidify, tiny crystals start forming. These grow in size and number until the complete mass is crystallized. However, except for ultrapure materials, there usually are substances present that do not fit into the normal crystalline structure. Some of these are rejected and tend to end up between the crystals, or grains, in what is called “grain boundary” material. Still other materials remain in some distributed form throughout the structure. The solid particles are called “inclusions.” Even holes or voids are left. A metal that is cooled down in this manner is said to be a “cast structure.” Although the metal may be used in this form, often it will be further worked in some way. These operations may include one or more of the following: forging, rolling, swagging, machining and/or grinding. These steps may just be a beginning, since many operations often are required. Each step can, and usually does, affect the potential fatigue life of the metal.
How Metal Fatigue May Start
If a simple bar of metal is pulled in tension from its ends, usually the load can be carried at least once to near its expected full or ultimate strength. However, if it is repeatedly subjected to tension loads high enough for fatigue cracks to develop, then only a lower portion of the ultimate strength may remain, and continued loading will eventually result in failure. The reason for this reduced useful strength lies partly in the basic nature of the metallic structure. Under repeated loading, slip or shear displacement may occur in some crystals, with one atomic plane slipping past another. Certain crystalline planes offer less resistance to such slippage than others. It might be as though they were tiny decks of cards that slide more easily in one plane. The crystals usually are oriented in random fashion, and the initial slip may be triggered by some irregularity in the atomic pattern. This irregularity may result from an inclusion or a void or some other stress concentration that causes the shear limit to be exceeded. Repeated loading causes a collection of these slips or dislocations to form. They continue to accumulate until the crystal fragments. This fragmentation distorts still other crystals, and the process continues until an opening or crack forms. The crack or cracks continue to grow until the metallic part no longer can carry the load, and fatigue failure results.
There are also other means by which fatigue may begin. For example, microcracks can start in grain boundaries. The cracks may be hastened by some chemical action. Thus there are various identified causes of fatigue, although much is still to be learned. The general result, however, is a progressive weakening of the metal structure by some microcracking process during loading.
How to Identify a Fatigue Failure
Although, in some instances, a considerable amount of experience is necessary to identify fatigue as a cause of failure, there are certain general characteristics that may be helpful. It is agreed that fatigue is a progressive process. Also, the growth of cracks usually is intermittent in progress. This intermittent growth pattern is sometimes in evidence in the fractured surfaces of the broken parts. Such patterns tend to resemble irregular concentric semicircles, with the center of the semicircles being the origin of the failure. If this “oyster shell” pattern is present in the fractured surfaces, the cause is probably fatigue.
Recognition of Fatigue in Machine Design
With the start of the Industrial Revolution, powerful steam engines and locomotives began to be built. Then unexplained failures in some mechanical parts began to be noted. August Wöhler, in Germany, was one of the first persons to identify the failures as fatigue and to record his findings. He even went further and demonstrated failures, using specimens from locomotive axle material. Although the phenomenon of metal fatigue now was recognized, it remained for the time of World War I and the early automobiles to bring the problem home to the average person. In those cars it was common to have fatigue failures in crankshafts, axles and springs.
By World War II, recognition of the metal-fatigue problem increased. Widespread use of aircraft focused attention on strength, weight and fatigue reliability. Today, with the increased use of machines, including helicopters, the demands upon design and reliability are even greater. Governments and corporations are intensively investigating the problem. Sophisticated equipment has been developed and is now in use to study designs and prototypes.
One result of all this effort is that handbooks and design manuals have been improved. These manuals provide, among other things, stress loading limits for given materials so that they can be used with some safety. These are called “endurance limits.” A simple representation is shown in the accompanying graph. With this information it might seem that the problem is essentially solved. Just operate within the safe limits and fatigue worries are over.
Unfortunately, however, the data and information provided cannot cover all operating conditions. In the actual use of a metallic part, we may not be able to predict accurately the spectrum of loading. The stress conditions often are complex and involve combinations of tension, compression and shear. Also, it is important to know the order in which the low and high loads occur, if the probability of the fatigue life is to be estimated. Much of the information has been derived from work on what we might call “plain” materials. These are materials lacking stress concentrations, such as holes, notches, rivets or welds. All of these generally tend to lower the basic endurance limits. Yet, even in plain materials, there are virtually endless variations in quality. These variations in crystal size, numbers and types of inclusions, hardnesses and internal stress all complicate the problems of design and manufacture.
Solving Problems of Design and Fabrication
Many of the machines and devices that we buy are designed and fabricated with the expectation that some of their parts may fail when in use. For example, there had been a practice of designing certain automobile parts to last for 100,000 miles (160,000 kilometers). By then the upholstery could be worn out and the body rusted and damaged. On the other hand, limitation on fatigue life of aircraft parts is imposed by weight. More material may be advantageous in design. Yet, any excess weight severely limits the amount of payload and fuel that can be carried.
In machines where life and property are at stake, it is imperative that serious accidents be avoided, if possible. Because of these considerations, two general design philosophies have emerged—the fail-safe and the safe-life concepts.
With the fail-safe concept, several parallel members are used to support a given load. Thus, if one member fails, the others are capable of supporting the load until repairs are possible. Another means used is to provide “crack stoppers.” According to this method, the part is designed with a thickened portion to reduce the stress. Possibly, a strong reinforcement is employed, to which the load will be transferred. With the fail-safe concept, inspection is important.
Often it is not possible to use fail-safe design. A shaft or a gear can hardly be made with parallel load-carrying members. For this type of part, the safe-life concept must be used. By this procedure, damage-tolerant design is employed, along with rigorous testing. In the case of these parts, special care is necessary throughout production and assembly.
Sometimes both fail-safe and safe-life concepts are used. Again inspection is important, if feasible. The fitting that failed in the helicopter on the Pan Am Building reportedly was to be up for inspection at 9,900 hours. However, according to a report it had logged only 7,000 hours. Hence, even if scheduled inspections are not frequent enough, disaster may strike.
Special Protective Procedures
Some special procedures can sometimes be used to help to prevent accidents due to fatigue failures. These are not always employed because of the additional cost, the lack of knowledge or facilities, or because they do not apply. Also, certain means are available to predict failure.
One of the important procedures that often can be used is shot-peening, or blasting. This provides what may be considered a compressive skin for the part. Since fatigue failures generally originate when the part is subjected to repeated nominal tension loads, the peening helps by keeping at least the surface of the part in compression.
Another procedure sometimes is called “autofrettage.” Although this procedure has been used on guns, the principle has many applications. The idea is to overload the part excessively so that certain high-tension areas will yield. Then, upon release of the load, these yielded areas go into compression. Such localized-compression areas provide protection by reducing the tension during normal service use.
Overloading can have other beneficial effects, if performed before the part is put into service. That is so if the part employs certain types of fasteners. An example is in riveted joints. Because of the imperfect matching of holes, certain rivets may be supporting the major portion of a load. By overloading the assembly, however, the highly loaded areas yield, thus distributing the load.
There are other means employed to prevent metal-fatigue failures, and these usually are quite helpful. They include stress relieving after welding, and polishing out holes and pits to reduce localized stress concentrations.
What Can You Do?
Though the designer and manufacturer may have done much to help to prevent failures, there is much you can do. Here are some pointers:
1. Operate the equipment within the recommended loads and speeds.
2. When repairing equipment, avoid making deep scratches, nicks, or file marks, at least on critical parts.
3. Avoid overheating, as this may affect the hardness of the metal and reduce its working strength.
4. Protect the metal against rusting and pitting.
5. Protect working parts from certain chemicals, such as acids. In some metals, exposure may cause atomic hydrogen to enter, predisposing the part to embrittlement and early failure. Another effect of the chemicals may be to cause stress corrosion.
What About Metal-Fatigue Accidents?
Can accidents caused by metal-fatigue failures be prevented? Yes, eventually.
Accidents tend to be caused by forms of selfishness, ignorance and carelessness. Sometimes, inordinate desire for profit, a still deficient knowledge of design, and carelessness on the part of those who make and use equipment, leave us vulnerable to metal-fatigue failures. However, a new system of things is at our door. In that system promised by man’s Creator, all forms of selfishness will be eliminated. Knowledge will increase, including that involving design. Also, those who then make and use equipment will do so with safety for all in mind.
[Graph on page 19]
(For fully formatted text, see publication)
Maximum Stress, lbs. per Sq. Inch
Failure Expected
Endurance Limit
Safe Operating Range
500,000
1,000,000
Cycles to Failure
One method of organizing fatigue data.
[Picture on page 19]
OYSTER PATTERN