Fatigue Failure in Mechanical Engineering
(Level 3 - Cyclic Loading and Time-Dependent Damage)
Fatigue is one of the most common causes of mechanical failure. Parts fail under repeated loading at stress levels far below what they could survive in a single application. Understanding fatigue is essential because it is often invisible until catastrophic failure occurs.
At this stage, the focus is on recognizing how cyclic stress causes progressive damage and why time and load cycles matter as much as stress magnitude.
What Fatigue Failure Is
Here's what catches people off guard: a part passes all your static load tests with huge safety margins, goes into service, and six months later it's cracked clean through. Fatigue is the silent killer—progressive weakening under repeated loading even when stress levels are completely safe for a single application. That shaft could handle 10,000 pounds once. Cycle it with 2,000 pounds a million times and it fails anyway.
No visible warning. Cracks grow internally, hidden from inspection until they're catastrophic. And it's time-dependent—failure isn't about peak stress alone, it's about how many times you apply that stress. Even low stress will eventually kill a part if you cycle it enough. People design for strength and forget about cycles. That's how rotating equipment fails in the field while the calculations said it was fine.
Fatigue accounts for over 80% of mechanical failures in rotating machinery, vehicles, aircraft, and anything that vibrates or cycles loads. Not overload. Not material defects. Fatigue. You ignore cyclic loading, you're not designing—you're guessing.
How Fatigue Cracks Form and Grow
Fatigue kills in three stages. First: crack initiation. Micro-cracks form at stress concentrations—drilled holes, sharp corners, notches, even machining marks on the surface. Surface roughness accelerates this. Manufacturing flaws you can't see with your eyes become crack nucleation sites. This stage can take thousands to millions of cycles depending on stress levels and how clean your geometry is.
Second: crack propagation. Now the crack grows slowly with every load cycle. Growth rate depends on the stress range you're cycling through and the material's fracture toughness. This is the stage where inspection might catch it—if you know where to look. Most people don't. They inspect the wrong spots or use methods that can't detect subsurface cracks. Meanwhile the crack is growing.
Third: final fracture. The crack reaches critical size. The remaining cross-section can't support the load anymore. Sudden catastrophic failure. This happens fast—one cycle the part's holding, the next it's in pieces. Here's the thing: most of the component's life is spent in stages one and two. Crack initiation and propagation take time. Final fracture? That's over before you notice.
The S-N Curve (Stress-Life Approach)
If you want to predict fatigue life, you need the S-N curve. Also called the Wöhler curve, it plots stress amplitude versus number of cycles to failure. This is your primary tool for answering the question: how long will this part last under cyclic loading?
High stress? The part fails in relatively few cycles—low-cycle fatigue territory. Lower the stress and it survives more cycles—high-cycle fatigue. Some materials, especially steels, have an endurance limit: a stress level below which the part theoretically has infinite life. Stay below that threshold and the part should never fail from fatigue no matter how many cycles you put on it.
Here's the catch: aluminum and most non-ferrous metals don't have an endurance limit. They will eventually fail at any stress level if you cycle them long enough. That changes how you design. A steel shaft might be good forever at low stress. An aluminum component? You're always on borrowed time. Cycle count matters, always.
Factors That Affect Fatigue Life
Fatigue life is brutally sensitive to details most people ignore. Stress concentrations—sharp corners, drilled holes, keyways, thread roots—these spots are where cracks start. A smooth radius versus a sharp corner can be the difference between 100,000 cycles and 10 million cycles. Surface finish matters just as much. Rough machining marks? Crack initiation sites. Polished surface? Dramatically longer life.
Residual stress changes everything. Compressive residual stress from shot peening or cold rolling can double or triple fatigue life. Tensile residual stress from welding or machining? Cuts life short. Mean stress—the average stress level you're cycling around—matters too. Cycling in tension reduces life. Cycling in compression increases it. A part loaded in pure alternating tension-compression will fail faster than the same part with a compressive mean stress.
Environment kills you. Corrosion accelerates crack growth—corrosion fatigue is worse than either corrosion or fatigue alone. High temperatures reduce fatigue resistance across the board. Here's the reality: fatigue is not just a material property. It's a system property. Design details, manufacturing quality, operating environment—all of it affects whether your part lasts a year or a decade.
Low-Cycle vs. High-Cycle Fatigue
Fatigue splits into two regimes based on how many cycles cause failure. High-cycle fatigue (HCF) is what you get in rotating machinery—engines, turbines, bearings—where parts see millions of cycles. Stress levels stay below yield strength. The part never plastically deforms on any single cycle, but eventually micro-damage accumulates and cracks form anyway. HCF is usually defined as more than 10,000 to 100,000 cycles to failure.
Low-cycle fatigue (LCF) is different. Fewer than 10,000 cycles, stress levels high enough to cause plastic deformation every cycle. You see this in pressure vessels that cycle between pressurized and depressurized states, or thermal cycling equipment where heating and cooling create huge stress swings. The part yields a little bit each cycle, and that accumulated plastic strain eventually tears it apart.
Different models and testing methods apply to each regime. Use the wrong approach and your predictions will be garbage. Know which regime your application lives in—HCF or LCF—because the physics, the failure mechanisms, and the design strategies are not the same.
Designing Against Fatigue Failure
Preventing fatigue isn't about picking a stronger material and calling it done. It's about obsessive attention to geometric details and manufacturing quality. Avoid stress concentrations—use generous fillet radii, eliminate sharp corners, chamfer edges. Every sharp transition is a crack initiation site waiting to happen. Improve surface finish on critical areas. Polish bearing surfaces, remove machining marks, eliminate surface defects. A rough surface can cut fatigue life in half.
Induce compressive residual stress with shot peening or cold rolling. This physically prevents cracks from opening—the compressive stress has to be overcome before a crack can grow. Design for damage tolerance: assume cracks will form eventually, then design so you can detect them before catastrophic failure. Inspection access, detectable crack paths, redundant load paths—these keep failures from being fatal.
Pick materials with high fatigue resistance, considering endurance limits and fracture toughness. But here's what separates good engineers from careless ones: inspection intervals based on predicted crack growth rates, not arbitrary calendar time. You need to know how fast cracks propagate in your conditions, then inspect before they reach critical size. Time in service means nothing if you don't understand the crack growth physics.
Why This Level Matters
Fatigue is the silent killer of mechanical systems. Parts that pass every static strength check can still fail catastrophically from cyclic loading. Your FEA shows green stress plots under peak load? Doesn't matter if you're cycling that load a million times. Most engineers learn stress analysis and think they're done. They're not even close.
This level teaches you to recognize when fatigue is a real risk—rotating equipment, vibrating structures, anything with repeated loading. You learn how cyclic stress causes progressive damage even at "safe" stress levels, how to apply S-N curves for life prediction, and which design details actually extend fatigue life versus which ones are just hopeful thinking. Surface finish, fillet radii, residual stresses—these aren't minor details, they're the difference between reliable operation and field failures.
This knowledge is essential for designing rotating machinery, vehicles, aircraft, and any structure that vibrates or cycles loads. Miss this and you're shipping parts with invisible expiration dates. Once you understand fatigue, you're ready for the next level: creep, buckling, and thermal failure—time-temperature effects and structural instability that kill parts in completely different ways.
Ready for the Next Level?
Once you understand fatigue failure, you're ready to learn about creep, buckling, and thermal failure—time-temperature effects and structural instability.
Continue to Level 4: Creep, Buckling & Thermal Failure →