What Is Fatigue in Materials and Why Do Strong Parts Still Break?

What Is Fatigue in Materials?

Fatigue in materials is the damage that adds up when a part is loaded again and again. It can come from loading, unloading, vibration, bending, pressure cycling, or a mix of these. When you compare material properties for a new design, tensile strength is useful, but it is not enough by itself. A bracket, shaft, spring, welded frame, aircraft panel, battery tab, or pump impeller may pass a static test and still crack after thousands or millions of cycles in service.

The U.S. Federal Aviation Administration explains fatigue and damage tolerance as the study of how aircraft materials and structures respond to repeated loading and environmental factors over time. Its page, last updated on April 9, 2026, points to repeated or fluctuating stresses, strains, and stress intensities. That wording is worth noting because fatigue is not only about the highest load on the drawing. It also depends on load range, cycle count, local shape, surface condition, environment, and the small flaws already inside the material.

Cyclic Loading Creates Small Damage

A cycle can be one engine start, one wheel rotation, one pressure pulse, or one bend in a hinge. The load may look safe on a gauge, but the material is still being worked each time. In metals, small areas near notches or surface marks can slip back and forth at the microscopic level. In polymers, heat and creep can add to the damage, and in composites, the matrix, fibers, and interface may not fail in the same way.

Cracks Start at Stress Raisers

Fatigue cracks often begin where stress gathers, such as a sharp corner, weld toe, thread root, drilled hole, inclusion, pore, machining groove, or corrosion pit. These spots are easy to miss when the part still looks clean from the outside. For additive manufactured metals, internal lack-of-fusion defects and gas pores can matter a lot. NIST’s Additive Manufacturing Fatigue and Fracture project, updated on September 25, 2025, shows high-cycle fatigue fracture surfaces in AM Ti-6Al-4V where cracks initiated at lack-of-fusion defects, with entrapped gas pores also visible.

Final Fracture Comes Fast

Most of the working life can pass while a crack is too small for a normal visual check. After the crack reaches a critical size, the remaining section may fail with little warning. A broken shaft often shows a smooth fatigue region with beach marks, then a rough final overload region. The last break may look like the main event, but the problem usually started much earlier.

Why Can Low Stress Still Cause a Fatigue Failure?

Fatigue can happen even when the nominal stress is below yield strength. This catches many buyers off guard because material data sheets usually focus on tensile strength, yield strength, elongation, and hardness. Those values come from short-term tests. Fatigue asks a different question: how does the material behave after many service events, with flaws, surface marks, humidity, heat, or vibration in the same picture?

Yield Strength Is Not the Whole Story

A part can stay elastic in a basic calculation but still have local plastic strain near a notch. That small local area may be enough for a crack to start. A polished laboratory specimen and a welded field part also do not behave the same. The weld toe, residual tensile stress, and local hardness change can cut fatigue life even when the base metal looks strong on paper.

Life Depends on Stress Range and Cycles

Fatigue design often uses an S-N curve, which plots stress against cycles to failure. Higher stress usually means fewer cycles, and lower stress usually means more cycles. Some steels show a fatigue limit in many traditional tests, while many aluminum alloys are rated at a chosen life, such as a stress level at a specified number of cycles. NASA-HDBK-5026, dated August 12, 2024, defines fatigue limit as a cyclic stress or strain range below which fatigue initiation failures are unlikely. The word “unlikely” is practical wording; it is not a guarantee.

Mean Stress Changes the Result

The same alternating stress can give different fatigue lives when the mean stress changes. A bolt that cycles between high tension and higher tension is not the same as a spring wire that cycles around zero load. Engineers use the stress ratio, often called R, to describe that condition. When you ask for fatigue data, do not accept a curve without the stress ratio, specimen type, surface finish, and test environment.

Which Materials and Processes Are Most Sensitive?

No single material is best in every fatigue comparison. A clean steel, a heat-treated aluminum alloy, a titanium alloy, a glass-fiber composite, and a high-performance polymer can all do the job when the application matches the material. The risk comes when the material, process, and service load are not aligned. New materials can make this harder to judge because small process changes may move fatigue life by a large margin.

Metals Show Clear Crack Growth Patterns

Metals are widely studied because many machines and structures rely on them. Crack initiation, stable crack growth, and final fracture can often be linked to fracture mechanics. That is why aircraft, bridges, pressure vessels, rotating parts, and medical devices use fatigue testing instead of only static strength checks. For a machined metal component, surface finish and edge radius can be just as important as the alloy grade.

Polymers and Composites Need Separate Tests

Polymers may heat up during fast cycling, so test frequency matters. Composites may lose stiffness through matrix cracking, fiber breakage, or delamination before a clear final fracture appears. Moisture and temperature can also shift the result. A composite tube that works indoors may not behave the same near salt spray, fuel vapor, or daily outdoor temperature swings. For this reason, a supplier should show fatigue data close to your use case, not only a general brochure number.

Additive Manufacturing Adds Defect Risk

Additive manufacturing can build shapes that machining cannot, but fatigue-critical use needs careful data. NASA-HDBK-5026 gives guidance for strength, fatigue, and fracture control of additively manufactured spaceflight hardware. NIST also states that AM fatigue and fracture work targets process control, non-destructive evaluation, material variability, and qualification. In shop terms, the printed part is not only a material; it is a material plus powder, build direction, heat treatment, porosity, surface roughness, and inspection history.

How Do Engineers Test Fatigue in Materials?

Good fatigue testing tries to connect lab control with real service. A lab test is cleaner than the field, which helps with comparison, but it can miss problems if the setup is too simple. Public NASA facility pages show how serious this work can be. NASA Johnson Space Center lists static and fatigue load testing, cyclic testing up to 100 Hz depending on load and stroke, and load frames with capacities from 10 to 220 kip. NASA Langley’s Fatigue and Fracture Laboratory describes servo-hydraulic machines from 5 to 400 kips.

S-N Curves for Early Screening

S-N testing is common for comparing materials, surface conditions, heat treatments, and manufacturing routes. Several specimens are tested at different stress levels until they fail or reach a runout cycle count. The result helps you screen choices early in a project. Still, S-N data is not universal. A smooth rotating-bending specimen cannot fully represent a welded bracket with bolt holes and paint damage.

Crack Growth Tests for Damage Tolerance

Damage tolerance starts with the idea that a crack or flaw may already exist. The question is how fast it can grow and how large it can become before repair or failure. The U.S. rule 14 CFR § 25.571 requires transport airplane structure evaluations to address fatigue, corrosion, manufacturing defects, accidental damage, typical loading spectra, temperatures, humidities, and test evidence for critical structural elements. That aviation rule is not your pump housing rule, but the logic is useful in many industries.

Service Spectra Make Tests More Realistic

Real loading is rarely a clean sine wave. A delivery drone arm sees takeoff vibration, gusts, landing shock, and handling damage. A rail part sees millions of similar cycles plus the odd overload. A pressure vessel sees startup, hold, shutdown, and temperature drift. Engineers often reduce those histories into load spectra, then use cycle counting methods. It is not exciting work, but it keeps the test tied to the actual duty. See also: Application.

What Design Choices Cut Fatigue Risk?

Better fatigue life usually comes from many small decisions, not one big material change. A stronger alloy may help, but a sharp corner can waste that gain very quickly. A polished surface may help too, but corrosion can take away the benefit. In practice, fatigue design is a chain of geometry, material, process, surface, environment, inspection, and service behavior.

Smooth Geometry and Better Surface Finish

Large fillet radii, clean transitions, good hole quality, and proper weld profiles reduce local stress concentration. Removing machining marks in the main stress direction can also help. For castings and printed parts, near-surface pores are often worse than deep pores because surface cracks can open more easily. A small burr at a hole edge looks cheap to remove, until it becomes the first crack site.

Residual Compression Helps the Surface

Processes such as shot peening, laser peening, cold expansion of holes, and surface rolling can add compressive residual stress at the surface. Since many fatigue cracks start at the surface, compression can slow initiation and early growth. The process still has to be controlled. Too much cold work, poor coverage, or later grinding can reduce the benefit. Ask for process records, not just the phrase “peened surface.”

Inspection Intervals Match Crack Growth

The NTSB report on Aloha Airlines Flight 243 is a hard reminder. On April 28, 1988, a Boeing 737-200 suffered explosive decompression at 24,000 feet, and about 18 feet of cabin skin and structure separated. The NTSB found that the maintenance program failed to detect significant disbonding and fatigue damage at a lap joint; corrosion and premature fatigue cracking were also cited. The lesson for any industry is direct: inspection must find the damage before the remaining life is gone.

What Data Should You Ask Suppliers to Provide?

When you buy new materials for cyclic service, a clean data sheet is only the starting point. Fatigue numbers depend on test setup, material lot, processing route, and environment. If a supplier gives only tensile strength and says the material is “high performance,” slow the discussion down. A strong static material can still be a poor fatigue choice in a vibrating bracket or rotating shaft.

Test Method and Specimen Details

Ask which standard or internal method was used, how many specimens were tested, and whether the specimens were smooth, notched, welded, printed, machined, polished, coated, or as-built. Also ask how failure was defined. Did the test stop at complete fracture, a stiffness drop, a visible crack, or a set runout cycle count? These details can change the meaning of the number.

Environment and Loading Ratio

Request stress ratio, frequency, temperature, humidity, medium, and corrosion exposure. If your part will see saltwater, fuel, hydrogen, sterilization, or hot oil, room-temperature air data is only a rough guide. If reliable public data cannot be found for your exact material and environment, write that in the engineering file and plan a targeted test. Guessing is not a design method.

Inspection and Traceability Records

For fatigue-critical parts, ask for heat treatment records, surface treatment records, non-destructive inspection results, lot traceability, and any crack growth or fracture toughness data. For additive manufactured parts, add powder history, build orientation, scan parameters, hot isostatic pressing status, surface finishing, and CT or other NDE records where suitable. You do not need a thick file for every washer. For cyclic load paths, though, the paper trail can save real money later.

FAQ

Q1: What Causes Fatigue in Materials? A: Repeated or fluctuating load causes small damage to build over time. Cracks often start at stress raisers such as holes, weld toes, pores, scratches, inclusions, or corrosion pits.

Q2: Can Fatigue Happen Below Yield Strength? A: Yes. The nominal stress may be below yield strength while local stress near a notch is high enough to start microscopic damage. That is why static strength alone is not enough.

Q3: Which Fatigue Property Should You Compare First? A: Start with S-N data at the same stress ratio, surface condition, and environment. For safety-critical parts, also compare crack growth rate, threshold behavior, and fracture toughness.

Q4: Is Additive Manufacturing Safe for Fatigue-Critical Parts? A: It can be, but only with controlled processing, defect inspection, surface treatment, heat treatment, and test data. Lack-of-fusion defects and pores can strongly affect fatigue life.

Q5: How Can You Reduce Fatigue Risk in a New Design? A: Use smoother geometry, better surface finish, suitable material data, realistic load spectra, controlled manufacturing, and inspection intervals based on crack growth. Small details matter a lot.