Fatigue life is the total number of stress cycles a material or structure can withstand before it breaks. Every time a bridge flexes under traffic, an airplane wing bends in turbulence, or a hip implant bears a step, that counts as one cycle. Accumulate enough cycles and even strong materials will crack and eventually fail, often at stress levels far below what it would take to break them in a single pull.
Understanding fatigue life matters because most mechanical failures in the real world aren’t caused by a single catastrophic overload. They’re caused by repetitive, seemingly harmless loads applied thousands or millions of times. Engineers use fatigue life predictions to decide how long a part can safely operate before it needs inspection or replacement.
How Fatigue Failure Develops
A material doesn’t just snap after a set number of cycles. Failure builds through a sequence of stages that starts at the microscopic level and ends with a visible break. First, repeated loading causes tiny, permanent changes in the material’s internal structure. These changes concentrate into microscopic cracks, usually at the surface or at internal flaws like voids and inclusions.
Those microscopic cracks then grow and merge into a single dominant crack. This dominant crack propagates in a stable, predictable way for a while, slowly extending with each load cycle. Eventually the remaining cross-section of material is too small to carry the load, and the part fractures suddenly. The final break often looks brittle and abrupt, even in materials that are normally ductile. That deceptive appearance is one reason fatigue failures historically caught engineers off guard.
The S-N Curve
The primary tool for characterizing fatigue life is the S-N curve (sometimes called a Wöhler curve, after the German engineer who pioneered fatigue testing in the 1860s). The “S” stands for stress amplitude and the “N” stands for the number of cycles to failure. You plot stress on the vertical axis and cycles on the horizontal axis, usually on a logarithmic scale because fatigue lives can range from hundreds of cycles to billions.
The relationship is intuitive: higher stress means fewer cycles to failure, and lower stress means more. For many steels, the curve flattens out at a certain stress level called the fatigue limit or endurance limit. Below that threshold, the material can theoretically endure an infinite number of cycles without failing. Steel’s fatigue limit is typically around 40 to 50 percent of its ultimate tensile strength.
Aluminum and many other non-ferrous metals behave differently. Their S-N curves never fully flatten out, meaning there’s no true endurance limit. Given enough cycles, even very low stresses will eventually cause failure. That’s why aluminum aircraft structures rely on scheduled inspections and defined service lives rather than assuming any stress level is permanently safe.
Predicting Fatigue Under Varying Loads
Real components rarely experience a single, constant stress level. A car suspension encounters potholes, highway cruising, and speed bumps all in the same trip. Engineers need a way to add up the damage from different stress levels, and the most widely used method is the Palmgren-Miner rule (often just called Miner’s rule).
The idea is simple. For each stress level, you divide the number of cycles actually experienced by the total number of cycles that stress level would take to cause failure on its own. You then add up all those fractions. When the sum reaches 1, the material is predicted to fail. If a component experiences 10,000 cycles at a stress level where failure occurs at 100,000 cycles, that contributes 0.1 (or 10 percent) of its fatigue life. Stack enough partial damage from different load levels and you reach 1.0.
Miner’s rule is popular because it’s straightforward, but it has a known weakness: it ignores the order in which loads are applied. In reality, a period of high stress followed by low stress can produce different damage than the reverse sequence. Under alternating multi-level loads, Miner’s rule can overpredict or underpredict the actual fatigue life. More sophisticated nonlinear damage models exist to handle those cases, but Miner’s rule remains the starting point for most engineering calculations.
How Fatigue Life Is Measured
Fatigue testing involves subjecting a specimen to repeated loading until it breaks, then recording the number of cycles. The most traditional setup is the rotating beam test, where a cylindrical specimen spins while a bending load is applied. Each rotation counts as one stress cycle. This method is popular because the specimen design is simple, the loading is well defined, and the machine can run at high frequency with relatively low energy, making it practical to accumulate millions of cycles.
Other common methods include axial loading (pulling and pushing the specimen along its length) and torsion testing (twisting it). Hydraulic machines can apply complex, variable-amplitude load patterns that better simulate real service conditions. Large-scale setups exist for testing full-size components like wind turbine blades or bridge sections, though these are far more expensive and time-consuming than small specimen tests.
Building an S-N curve for a single material typically requires testing dozens of specimens at different stress levels, because fatigue life is inherently variable. Two identical specimens tested at the same stress can fail at noticeably different cycle counts due to microscopic differences in their internal structure. Statistical methods are used to account for this scatter and establish reliable design values.
Factors That Shorten Fatigue Life
Fatigue life is highly sensitive to conditions that might seem minor in a static strength calculation. Surface finish is one of the most important. Machining operations leave small scratches and grooves that act as stress concentrators, giving cracks a head start. A polished surface can have a fatigue life several times longer than a rough-machined one, even though both surfaces are made of the same material.
Geometric features like sharp corners, holes, notches, and abrupt changes in cross-section create localized stress spikes that dramatically reduce fatigue life. This is why engineers use generous fillet radii and smooth transitions in parts subject to cyclic loading.
Mean stress also plays a significant role. If a component is preloaded in tension (imagine a bolt that’s already stretched tight), even a small cyclic stress on top of that static load will shorten fatigue life compared to the same cyclic stress applied with no preload. The higher the average tension, the fewer cycles the part can survive.
Corrosion is particularly damaging. When cyclic stress and a corrosive environment act together, the combination is worse than either one alone. Chemical reactions create small surface pits that serve as crack initiation sites, and the corrosive environment accelerates crack growth once a crack has started. This phenomenon, called corrosion fatigue, can reduce fatigue life by orders of magnitude and effectively eliminates the fatigue limit that steel would otherwise exhibit in dry air.
Fatigue Life in the Human Body
The concept of fatigue life isn’t limited to metal and plastic. Human bone behaves in a remarkably similar way. Stress fractures are essentially fatigue failures of bone: repetitive loading, typically in the legs and feet, accumulates microscopic damage faster than the body can repair it.
Bone is a living material that constantly remodels itself. Cells called osteoclasts break down old bone while osteoblasts build new bone, a cycle that normally takes three to four months. Under repetitive loading, bone cells that sense mechanical stress can become compromised, disrupting the signaling that coordinates repair. Resorption outpaces formation, weakening the bone until a stress fracture develops.
This is why stress fractures are common in military recruits and distance runners who ramp up activity too quickly. The bone’s “fatigue life” is being consumed faster than remodeling can restore it. Rest and gradual training progression are the biological equivalent of keeping stress below the endurance limit.