Metal Fatigue Analysis: A Complete Guide to Fatigue Failure and Prevention

Metal fatigue can cause a component to fail even when its normal operating stress remains below the material’s yield strength. Shafts, gears, brackets, couplings, fasteners, and other mechanical parts may experience millions of bending, torsional, vibration, pressure, or thermal cycles during service. Each individual cycle may appear harmless, but microscopic damage can gradually develop into a crack and eventually cause metal fatigue failure. Understanding what is metal fatigue therefore requires more than checking tensile strength. Engineers must evaluate cyclic loads, local stress concentrations, material behavior, surface condition, environment, and manufacturing quality. This guide explains the main methods used in fatigue failure analysis and how better design and CNC machining decisions can reduce fatigue risk.

What Is Metal Fatigue?

Metal fatigue is the progressive damage and eventual cracking of a metal component caused by repeated or cyclic loading. This is the simplest definition of metal fatigue, but an important detail is often missed: fatigue damage can accumulate even when the nominal stress is below the material’s yield strength.

Therefore, if an engineer is asked to define metal fatigue or explain what is fatigue in metals, the answer should focus on repeated loading rather than a single overload. A component may remain apparently elastic during normal operation while microscopic regions around the surface, inclusions, notches, holes, or geometric transitions experience localized damage. Repeated cycling allows this damage to accumulate until a fatigue crack forms and grows.

Static Failure vs. Fatigue Failure

This distinction explains why checking only a material’s yield strength is insufficient for components subjected to repeated service loads.

How Does a Fatigue Crack Develop?

A typical fatigue failure develops in three stages.

1. Crack initiation. Repeated loading causes localized microscopic deformation. Surface irregularities, scratches, machining marks, inclusions, corrosion pits, thread roots, holes, and sharp transitions can increase local stress and provide favorable locations for a crack to begin.

2. Crack propagation. Once a crack exists, each additional load cycle can extend it slightly. The remaining load-bearing section becomes progressively smaller.

3. Final fracture. Eventually, the remaining cross-section is no longer sufficient to carry the applied load. The component then fractures rapidly, producing a final overload region that may look very different from the earlier fatigue-growth area.

For this reason, a broken part may appear to have failed suddenly even though the underlying damage developed over thousands or millions of cycles.

What Is a Fatigue S-N Curve?

An S-N curve, also called a stress-life curve, relates cyclic stress amplitude to the number of cycles a material can withstand before fatigue failure. “S” represents stress and “N” represents cycles to failure.

In general, increasing cyclic stress reduces fatigue life. Lower stress levels allow a greater number of cycles before failure. This relationship is widely used when the component operates primarily within the elastic range.

Усталостная прочность

Fatigue strength describes the cyclic stress a material can withstand for a specified number of cycles under defined test conditions. It should always be interpreted together with the material grade, heat treatment, loading ratio, surface condition, temperature, and test method.

Fatigue Limit or Endurance Limit

Some steels show a region on the S-N curve where further reductions in stress correspond to extremely long fatigue lives, commonly described as an endurance or fatigue limit. Many aluminum alloys, however, are normally designed using fatigue strength at a specified number of cycles rather than assuming a distinct endurance threshold. The source article similarly distinguishes the fatigue behavior of ferrous and non-ferrous alloys when discussing long-life design.

Engineers should therefore avoid statements such as “steel has infinite fatigue life” or “aluminum always fails after a certain number of cycles.” Real fatigue behavior depends on the specific alloy, condition, component geometry, loading, manufacturing state, and environment.

What Causes Stress Concentration in Fatigue Failure?

Fatigue cracks frequently begin where component geometry raises local stress above the nominal stress calculated for the overall section. The theoretical stress concentration factor, Kt, describes this geometric amplification.

Common stress-concentrating features in CNC machined parts include:

• sharp internal corners;
• small fillet radii;
• drilled and cross-drilled holes;
• keyways;
• grooves and undercuts;
• thread roots;
• shaft shoulders;
• abrupt changes in section thickness;
• diameter transitions.

Kt vs. Kf

Kt describes the theoretical increase in stress created by geometry, whereas the fatigue notch factor, Kf, represents how strongly that notch actually reduces fatigue performance for a particular material. The difference matters because materials do not respond equally to the same geometric notch.

How Can CNC Part Geometry Reduce Fatigue Stress?

Designers can often reduce fatigue risk by increasing fillet radii where practical, introducing gradual section transitions, removing unnecessary sharp corners, optimizing the position of holes and grooves, and keeping thread roots or keyways away from peak-stress zones.

However, simply making every radius as large as possible is not a realistic design rule. Bearing seats, mating parts, tool access, assembly clearances, sealing features, and packaging constraints may limit the available geometry. The objective is to reduce unnecessary stress concentration while maintaining the component’s functional requirements.

How Does Surface Finish Affect Metal Fatigue?

Surface condition is particularly important because many fatigue cracks initiate at or near the surface. CNC machining directly determines surface topography and can also influence near-surface residual stress.

CNC Machining Marks

Tool marks, feed marks, scratches, burrs, chatter, torn material, and other machining irregularities can behave as microscopic stress raisers. Their importance increases when they occur in a highly loaded area such as a shaft shoulder, fillet, thread root, thin section, or bending surface.

The orientation of machining marks can also matter. A groove running across the primary tensile stress direction may be more damaging than a shallow feature aligned differently. Therefore, specifying a surface finish without considering load direction and critical geometry may not fully control fatigue performance.

Surface Roughness and Fatigue Life

A smoother critical surface generally reduces the number and severity of surface irregularities from which fatigue cracks can initiate. This does not mean that lower Ra automatically guarantees longer life. Geometry, loading, material condition, residual stress, corrosion, heat treatment, and local defects must also be considered.

Residual Stress

Manufacturing operations can leave residual stress in the surface layer after the cutting or finishing force has been removed. Tensile residual stress can make crack initiation and propagation more favorable, while compressive residual stress is generally beneficial because it opposes tensile crack opening.

The original engineering guide therefore treats surface roughness and residual stress together rather than evaluating surface appearance alone.

Which Surface Treatments Can Improve Fatigue Resistance?

Shot Peening

Shot peening intentionally plastically deforms the surface with small impacts, creating a compressive residual stress layer. This can delay surface crack initiation and is commonly considered for components exposed to repeated loading.

Grinding and Polishing

Precision grinding or polishing can remove significant machining marks and improve critical surfaces. However, grinding must be controlled. Excessive heat, inappropriate parameters, or grinding burn can damage the surface and potentially introduce unfavorable residual stress.

Case Hardening

Processes such as carburizing and nitriding can modify surface hardness, wear behavior, microstructure, and residual stress. Their effect on fatigue depends on the material and process parameters, so a surface treatment should be selected for the actual component rather than assumed to universally improve fatigue life.

What Is the Difference Between High-Cycle and Low-Cycle Fatigue?

High-cycle fatigue and low-cycle fatigue are sometimes separated only by cycle count, but the more useful engineering distinction is the dominant deformation mechanism.

A value around 10⁵ cycles is sometimes used as a rough practical boundary, but it should not be treated as a universal material law. The key question is whether local deformation remains predominantly elastic or whether significant cyclic plastic strain occurs. This deformation-based distinction is also emphasized in the reference guide.

Stress-Life vs. Strain-Life: Which Method Should You Use?

Stress-Life Method

The stress-life or S-N method is commonly appropriate when stresses remain mostly elastic and the expected life is relatively long. Examples include many rotating shafts, gears, springs, and structural components exposed to repeated vibration.

Strain-Life Method

The strain-life or E-N method becomes more relevant when local plastic deformation is significant. This can occur during severe cyclic loading, thermal cycling, or around strong stress concentrations even if the nominal stress of the whole component does not appear particularly high.

When Should Fracture Mechanics Be Used?

Traditional stress-life and strain-life approaches are often used to estimate the initiation and total fatigue life of an initially uncracked component. Fracture mechanics is particularly useful when a crack, flaw, or defect is already known or assumed to exist.

Linear Elastic Fracture Mechanics can be used to examine crack-driving force and estimate how a crack may grow from an initial size toward a critical size.

Crack Growth Rate

Crack growth is often expressed as da/dN, representing the increase in crack length per load cycle. Combining crack-growth behavior with the component’s stress state allows engineers to estimate how quickly an existing flaw may propagate.

Remaining Useful Life

This approach supports damage-tolerant design. Instead of assuming a critical component is completely defect-free forever, engineers can evaluate whether an existing or detectable flaw can remain below a dangerous size during a defined inspection interval. The source article uses this crack-growth approach to explain remaining useful life and damage tolerance.

Which Material Properties Affect Fatigue Life?

Tensile strength is important, but it does not fully describe fatigue resistance. More specialized material properties may include fatigue strength coefficient, fatigue ductility coefficient, cyclic strain hardening behavior, toughness, ductility, crack-growth resistance, and cyclic stress-strain behavior.

Microstructure also matters. Inclusions, pores, segregation, improper heat treatment, decarburization, and other defects can provide locations for fatigue cracks to initiate.

Consequently, changing from one metal to a nominally stronger alloy is not automatically the best solution to a fatigue problem. A stronger material used with a poor surface, sharp notch, unfavorable heat treatment, or severe residual tensile stress may still fail prematurely.

How Do Different Metals Behave Under Fatigue?

Сталь

Many steels provide good fatigue strength and can be modified through alloy selection and heat treatment. Some steels exhibit an identifiable endurance region under particular laboratory conditions. Their actual component performance still depends strongly on surface condition, section size, stress concentration, heat treatment, and environment.

Алюминий

Aluminum alloys are widely used where weight reduction is important. Because many aluminum alloys do not exhibit a clearly defined endurance limit, designers generally specify an allowable fatigue strength or service life at a defined number of cycles.

Титан

Titanium alloys combine relatively low density, high specific strength, and strong corrosion resistance, making them useful for aerospace and other demanding components. Their fatigue behavior remains sensitive to surface condition, stress concentration, machining damage, environment, and alloy condition.

Нержавеющая сталь

Stainless steels can be attractive when cyclic loading occurs in corrosive environments. However, corrosion resistance alone does not guarantee high fatigue life. Grade, heat treatment, surface finish, geometry, residual stress, and actual environmental exposure still need to be considered.

How Does the Environment Affect Fatigue Life?

Corrosion Fatigue

Corrosive environments can accelerate fatigue by creating pits or surface damage that act as local stress concentrators. A component that performs adequately in dry laboratory air may therefore have a shorter fatigue life in seawater, humid environments, chemical processing equipment, or aggressive industrial fluids.

Thermal Fatigue

Repeated heating and cooling causes materials to expand and contract. If this movement is restrained or temperature gradients exist across the part, cyclic thermal stresses can develop even without a conventional external force.

High-Temperature Fatigue

At elevated temperatures, material strength, creep resistance, cyclic behavior, oxidation, and crack-growth behavior may all change. Room-temperature S-N or E-N data should not simply be transferred to a high-temperature component. The reference guide likewise treats temperature-dependent material data and creep-fatigue interaction as important extensions of the normal fatigue workflow.

How Do Welded Joints Affect Fatigue Strength?

Welded structures require special consideration because several fatigue risks may occur simultaneously.

Weld Geometry

The weld toe and root create abrupt geometric changes and therefore natural stress concentration zones.

Зона термического воздействия

Welding alters the local thermal history and microstructure surrounding the weld. The resulting heat-affected zone may respond differently to cyclic loading than the unaffected base material.

Welding Residual Stress

Cooling and contraction can introduce substantial tensile residual stress. For this reason, the fatigue resistance of a welded joint should not be estimated simply by applying the polished base-metal fatigue strength to the entire assembly.

How Is Fatigue Life Evaluated Under Variable Loads?

Real machinery rarely operates under one perfectly constant stress amplitude. A vehicle component may experience acceleration, braking, cornering, vibration, impact, and road inputs. A rotating machine may encounter start-stop cycles, fluctuating torque, imbalance, and transient loads.

Load Spectrum

The first step is therefore to define the real operating load spectrum rather than relying only on the maximum design load.

Rainflow Counting

Rainflow counting can convert a complex time history into identifiable stress or strain cycles that are easier to use in fatigue calculations. The original guide includes this method in its CAE workflow for processing variable real-world loads.

Cumulative Damage

Methods such as Miner’s rule estimate how damage from cycles at different stress levels accumulates toward failure. They are useful engineering approximations, but their predictions should not be interpreted as exact physical descriptions of every loading sequence.

How Is Fatigue Analysis Performed With FEA?

Finite element analysis is commonly used to identify critical stress or strain regions before fatigue life is calculated.

1. Build or import the component geometry.
2. Define the relevant material properties.
3. Apply realistic constraints and boundary conditions.
4. Apply operating forces, pressures, torques, acceleration, or thermal loads.
5. Calculate stresses and strains.
6. Identify critical local hotspots.
7. Assign appropriate S-N or E-N fatigue data.
8. Define the loading history.
9. Calculate damage or predicted fatigue life.
10. Review the critical regions and modify the design if required.

FEA is powerful, but a visually detailed life contour is not automatically an accurate prediction. Incorrect material curves, oversimplified supports, inaccurate loads, unrealistic contact conditions, ignored residual stress, or an incorrect surface assumption can all produce misleading results. The reference article similarly emphasizes that fatigue solver accuracy is highly dependent on material data and realistic loading histories.

How Is a Metal Fatigue Test Performed?

Physical testing remains important when fatigue performance needs to be established or validated.

Specimen Preparation

Fatigue specimens need tightly controlled geometry, dimensions, surface condition, and material consistency. Poor machining can introduce unintended notches, scratches, burrs, dimensional variation, or residual stress and thereby distort test results.

Stress Level Selection

Specimens are normally tested at multiple stress levels so that the relationship between applied stress and cycles to failure can be established.

Specimen Count

Fatigue data naturally shows statistical scatter. Testing multiple specimens at each condition improves confidence in the resulting design curve.

Runout

A test may be stopped after a predefined number of cycles if the specimen has not failed. Such a specimen is normally reported as a runout rather than assumed to have literally unlimited life.

S-N Curve Generation

The resulting stress-life data can then be statistically evaluated to establish a fatigue curve suitable for the intended engineering purpose. The consistency of fatigue specimens is therefore not merely a laboratory concern; precision machining quality can directly influence the reliability of the data.

How Can You Identify Fatigue Failure From a Fracture Surface?

Fractography is one of the most valuable tools in fatigue failure analysis because the fracture surface may reveal where a crack started, how it propagated, and where final overload occurred.

Crack Initiation Site

The investigation normally begins by tracing the fracture back to its origin. Potential initiation locations include a sharp radius, hole, thread root, inclusion, machining mark, corrosion pit, surface scratch, or manufacturing defect.

Crack Propagation Zone

Macroscopic beach marks or similar patterns may indicate progressive crack growth under changing service conditions. Microscopic examination can reveal additional crack-growth features.

Final Overload Zone

Once the fatigue crack has consumed enough of the load-bearing section, the remaining material fractures rapidly. This final overload area should not be mistaken for the original cause of the failure.

Combining fracture evidence with load history, material records, dimensional inspection, surface inspection, and stress analysis helps distinguish a design problem from a material defect, machining problem, corrosion-assisted crack, or unexpected overload.

What Are Common Metal Fatigue Examples?

Useful metal fatigue examples are not limited to famous aircraft or bridge failures. Fatigue occurs in ordinary machined components whenever cyclic load combines with a vulnerable local feature.

The common theme is that fatigue failure is usually controlled by the local condition at the most highly stressed feature rather than by the average stress across the complete part.

How Can CNC Machining Improve Fatigue Resistance?

CNC machining cannot eliminate fatigue as a physical mechanism, but manufacturing decisions can reduce several conditions that encourage fatigue crack initiation.

Optimize Fillets and Transitions

Precision machining can accurately produce specified fillets, tapers, reliefs, and gradual transitions. Maintaining the designed radius is particularly important around shaft shoulders, pockets, mounting features, and load-carrying transitions.

Control Critical Dimensions

Bearing journals, press-fit regions, locating diameters, bores, mating faces, and threaded features influence how load is distributed through an assembly. Dimensional errors or misalignment can introduce bending, uneven contact, or unintended preload that increases cyclic stress.

Improve Critical Surface Finish

Finish milling, precision turning, grinding, and polishing can be applied selectively to highly stressed surfaces rather than unnecessarily tightening roughness requirements across the entire component.

Remove Burrs and Unintended Sharp Edges

Burrs and uncontrolled edge conditions can create local stress concentration or interfere with correct assembly. Deburring should therefore be considered part of fatigue-critical manufacturing rather than only a cosmetic process.

Control Machining-Induced Surface Damage

Excessive cutting heat, severe tool wear, chatter, aggressive grinding, and unstable cutting conditions can degrade the surface layer. Process planning should prioritize stable cutting and appropriate finishing where fatigue loading is significant.

Maintain Material and Heat-Treatment Traceability

For high-duty components, verifying alloy condition and heat treatment helps ensure that the material used in production corresponds to the fatigue properties assumed during engineering analysis.

How Can Engineers Design CNC Parts for Better Fatigue Life?

• Avoid unnecessarily sharp internal corners.
• Increase fillet radii where function and assembly permit.
• Use gradual section and diameter transitions.
• Keep unnecessary holes away from peak-stress regions.
• Evaluate keyways, grooves, cross-holes, and undercuts carefully.
• Consider thread-root stress under cyclic tensile loading.
• Specify surface finish selectively on fatigue-critical regions.
• Select material and heat treatment according to cyclic rather than static requirements alone.
• Use realistic operating load spectra.
• Account for corrosion, temperature, and other service environments.
• Use FEA to identify local stress hotspots where necessary.
• Validate safety-critical designs using appropriate fatigue tests.

There is no universal “best” fillet radius, surface roughness, or material for fatigue resistance. These requirements must be derived from component geometry, loading, environment, service life, and manufacturing constraints.

Example: Fatigue Failure in a CNC Machined Shaft

Consider a CNC machined steel shaft operating in rotating equipment. The shaft repeatedly experiences bending as well as torsional load during startup, normal operation, and shutdown.

After an unexpectedly short service life, a crack is discovered close to a shoulder where the shaft diameter changes.

A systematic analysis may proceed as follows:

1. Fractography confirms progressive fatigue crack growth rather than a single overload.
2. The crack origin is located close to the shaft shoulder.
3. Service data confirms repeated bending at this location.
4. FEA identifies a local stress hotspot around the diameter transition.
5. Dimensional inspection shows that the shoulder radius is relatively small for the applied loading.
6. Surface inspection finds circumferential machining marks close to the crack origin.
7. The shoulder geometry is redesigned with a larger practical transition radius.
8. The critical surface receives an improved finishing process.
9. A compressive-stress treatment such as shot peening is considered if compatible with the component requirements.
10. The revised design is reanalyzed and validated under representative cyclic loading.

The important lesson is that simply replacing the shaft with a higher-strength steel might not address the root cause. Effective fatigue improvement usually requires understanding the interaction between load, geometry, material, surface condition, and manufacturing process.

How Should You Perform Fatigue Failure Analysis?

When a component fails prematurely, fatigue failure analysis should follow a structured process rather than immediately blaming the material or increasing the safety factor.

Step 1: Confirm the Failure Mode

Use visual examination and fractography to determine whether progressive fatigue crack growth actually occurred.

Step 2: Identify the Crack Origin

Trace the fracture back to the initiation site and document nearby geometry, surface condition, defects, corrosion, inclusions, and machining features.

Step 3: Determine Actual Operating Loads

Compare design assumptions with real service conditions. Measurement using strain gauges, vibration data, torque measurements, pressure records, or operating histories may reveal loads that were not included in the original design.

Step 4: Review Component Geometry

Evaluate fillets, holes, grooves, threads, shoulders, wall transitions, and other locations where local stress could be substantially higher than nominal stress.

Step 5: Verify Material Condition

Check alloy grade, heat treatment, hardness, microstructure, defects, and material certification where relevant.

Step 6: Review Manufacturing Quality

Inspect the component for tool marks, burrs, scratches, chatter, grinding burns, dimensional errors, residual stress concerns, and deviations from the specified surface condition.

Step 7: Reproduce the Failure With Analysis

A useful FEA or fatigue model should be capable of identifying the same critical region observed on the failed component. If the model predicts failure somewhere completely different, the assumptions may still be incomplete.

Step 8: Develop Corrective Actions

Potential solutions may involve geometry, material, heat treatment, machining process, surface treatment, operating load, or several of these factors together.

Step 9: Validate the Improvement

Reanalysis alone may not be sufficient for a critical part. Accelerated fatigue testing, component testing, or controlled field validation can demonstrate whether the change actually improves service life. This diagnose-analyze-improve-validate sequence follows the same structured philosophy emphasized by the reference engineering guide.

Common Metal Fatigue Analysis Mistakes

Using Tensile Strength as the Only Material Criterion

Static tensile strength does not describe the entire cyclic response of a material.

Assuming Stress Below Yield Means the Part Is Safe

Localized cyclic damage can occur even when nominal stress remains below yield strength.

Ignoring Stress Concentration

Nominal stress may be moderate while a hole, thread, groove, or small radius experiences much higher local stress.

Ignoring Surface Condition

A fatigue-critical design can lose much of its intended advantage if the manufactured surface contains deep tool marks, scratches, burrs, or grinding damage.

Using Nominal Stress Instead of Local Stress

Fatigue cracks begin locally, so average section stress may not represent the condition that actually controls life.

Using S-N Data When Plastic Strain Is Significant

A strain-life method may provide a more appropriate model when repeated local yielding occurs.

Ignoring Residual Stress

Two components with the same geometry and roughness can behave differently if their manufacturing processes create different residual stress states.

Ignoring Corrosion and Temperature

Laboratory room-temperature data may not represent a component exposed to moisture, chemicals, seawater, heat, or thermal cycling.

Using Unrealistic Loading Assumptions

Fatigue predictions based only on a simplified maximum load can miss vibration, impacts, startup cycles, variable torque, or other important service events.

Treating FEA Fatigue Life as an Absolute Number

Simulation is a prediction based on assumptions. The output should be interpreted together with uncertainty in loads, materials, manufacturing conditions, and boundary conditions.

Changing Material Before Finding the Root Cause

A more expensive or stronger alloy will not necessarily solve a fatigue problem created primarily by poor geometry, surface damage, misalignment, or unexpected loading.

Frequently Asked Questions About Metal Fatigue

What is metal fatigue?

Metal fatigue is progressive material damage caused by repeated or cyclic loading. Microscopic cracks can initiate and grow over time until the remaining section can no longer support the applied load. Fatigue can occur even when nominal stress is below the material’s yield strength.

What causes metal fatigue failure?

Metal fatigue failure is normally caused by the combination of cyclic loading and a vulnerable local condition. Common contributors include stress concentration, poor surface condition, unfavorable residual stress, material defects, corrosion, temperature, and unexpected operating loads.

What are common metal fatigue examples?

Common examples include rotating shafts failing at shoulders, gear teeth cracking at the root, vibrating brackets cracking around mounting holes, threaded components failing at thread roots, and couplings developing cracks around keyways or section transitions.

Can metal fatigue occur below yield strength?

Yes. This is one of the defining characteristics of fatigue. Although the overall component may remain macroscopically elastic, repeated loading can cause localized microscopic damage and eventually initiate a fatigue crack.

What is the difference between fatigue strength and fatigue limit?

Fatigue strength normally refers to the stress a material can withstand for a specified number of cycles. Fatigue or endurance limit refers to a stress region below which some materials exhibit extremely long fatigue lives under defined conditions. Not every metal has a clearly defined endurance limit.

Does a smoother surface increase fatigue life?

A smoother fatigue-critical surface generally reduces severe surface irregularities that can initiate cracks, but surface roughness is only one factor. Geometry, residual stress, material condition, loading, heat treatment, and environment also influence fatigue performance.

Can CNC machining cause fatigue failure?

CNC machining does not inherently cause fatigue failure, but inappropriate geometry or machining conditions can reduce fatigue life. Sharp transitions, deep tool marks, burrs, chatter, surface damage, excessive cutting heat, and unfavorable residual stress can all make crack initiation easier in a cyclically loaded component.

Is “metal fatique” the same as metal fatigue?

Yes. Metal fatique is a common misspelling of metal fatigue. The correct engineering term is metal fatigue, meaning progressive damage caused by repeated or cyclic loading.

Заключение

Reliable metal fatigue design requires more than choosing a high-strength alloy. Engineers need to evaluate cyclic loads, local stress concentration, material behavior, surface condition, residual stress, temperature, corrosion, and manufacturing quality together. When failure occurs, systematic fatigue failure analysis should identify where the crack started and why before corrective action is selected. For CNC machined shafts, gears, brackets, couplings, housings, suspension components, and other cyclically loaded parts, accurate fillets, smooth transitions, controlled dimensions, suitable surface finishes, proper materials, and stable machining processes can all help reduce fatigue crack initiation risk and improve long-term reliability.