Introduction to Fatigue Analysis

Understand Crack Initiation and Propagation to Prevent Failures and Enhance Structural Integrity

Summary

Fatigue Analysis can be used to evaluate the service life of components and reduce the probability of accidents. Most equipment vibration problems are due to thermal, mechanical, acoustic, or flow related excitation – and if left unchecked, will cause equipment fatigue and subsequent failure, which may lead to greater repair costs.

Fatigue failures are characterized as a progressive structural failure phenomenon that occurs by the initiation and propagation of cracks to an unstable size – until failure. Fatigue failures continue to be a major concern in engineering design. 

History

In the early 1800’s there were reports of various engineers studying fatigue and failure of railroad equipment. One of the first documented fatigue investigations was reported by a German mining engineer, W. A. J. Albert, who in 1829 performed repeated loading tests on hoist chains. Later in the 1850s, August Wöhler’s research on railway axle failures led to the development of strategies for averting fatigue failure caused by bending, torsion, and axial loads. Wöhler also demonstrated the influence of both cyclic and accompanying steady stresses on fatigue. For a more comprehensive understanding of the historical development, read more about the History of Fatigue Analysis.

Discussion

Fatigue failure investigations over the years have led to the observation that the fatigue process embraces two domains of cyclic stressing or straining that are significantly different in character, and with each – failure is produced by different physical mechanisms. One domain of cyclic loading is when significant plastic straining occurs during each cycle, which is associated with loads exceeding the material elastic limit, resulting in low cycles to failure – commonly referred to as low-cycle fatigue.

The other domain of cyclic loading is when strain cycles are largely confined to the elastic range, and is associated with lower loads and long lives, or high numbers of cycles to failure – commonly referred to as high-cycle fatigue. Low-cycle fatigue is typically associated with cyclic lives from one up to about 104 or 105 cycles, and high-cycle fatigue for lives greater than about 104 or 105 cycles.

The final failure occurs when the remaining section—the part of the whole section not yet affected by the progressing crack—is no longer able to withstand normal operating forces. See Introduction to Fracture Mechanics.

Low-Cycle vs. High-Cycle Fatigue — Cyclic Regime Comparison
LOW-CYCLE FATIGUE ≤ 10⁴ – 10⁵ cycles HIGH-CYCLE FATIGUE > 10⁴ – 10⁵ cycles Exceeds yield — plastic strain each cycle High Δσ • Loads exceed elastic limit • Significant plastic deformation • Strain-based analysis required Elastic range — low strain amplitude Low Δσ • Stresses within elastic range • Very long life (>10⁵ cycles) • Stress-based (S-N) analysis used
Figure 1. Schematic comparison of low-cycle fatigue (large plastic strain, few cycles) and high-cycle fatigue (elastic strain, large number of cycles). Original diagram — O'Donnell Consulting Engineers.
S-N Curve (Wöhler Curve) — Stress Amplitude vs. Cycles to Failure
S-N curves: stress amplitude versus cycles to failure. Steel levels off at an endurance limit; aluminum keeps declining. Stress Amplitude (σ) Number of Cycles to Failure, N (log scale) 10¹ 10² 10³ 10⁴ 10⁵ 10⁶ 10⁷ High Low LOW-CYCLE HIGH-CYCLE ~10⁴–10⁵ σe(endurance limit) Non-ferrous (aluminum) No endurance limit; continues to decline Ferrous (steel) Endurance limit plateau Transition point Ferrous / steel (endurance limit) Non-ferrous / aluminum (no limit)
Figure 2. Typical S-N (Wöhler) curves comparing ferrous and non-ferrous metals. Ferrous materials exhibit a defined endurance limit; non-ferrous materials do not. Original diagram — O'Donnell Consulting Engineers.

In some instances, the fracture’s appearance exhibits recognizable signs that it was caused by fatigue and may give other clues that point to the failure’s origin. Characteristic features, commonly called “beach marks,” represent progressive stages of crack development. In certain metals (such as aluminum alloys) beach marks may not be clearly visible by optical means. In this case, striations usually can be observed with a scanning electron microscope (SEM). 

Components that see a very large number of load cycles can be designed for infinite life by keeping the alternating stress amplitude below the material’s endurance limit, after adjusting for mean stress, surface finish, size, and stress concentrations. This approach applies only to materials that show an endurance limit, such as carbon and low-alloy steels in non-corrosive environments. Aluminum alloys and most other non-ferrous metals do not have one, and a corrosive environment can remove it in steels as well. Components with stress amplitudes above the endurance limit, or made from materials without one, are designed for finite life: the number of load cycles expected in service must stay below the allowable number of cycles from the applicable fatigue design curve.

Fatigue life calculations involve more than reading a single stress value off an S-N curve. Mean stress matters: a tensile mean stress shortens fatigue life compared with fully reversed loading at the same alternating stress. In general practice it is accounted for with corrections such as the Goodman, Gerber, or Soderberg relationships, while the ASME Section III design curves already include an adjustment for the maximum possible mean stress. Real components rarely see one constant load cycle, so the stress history is broken into individual cycles using a counting method such as rainflow counting (ASTM E1049) or the peak-and-valley pairing used in ASME Section III. The damage from each group of cycles is then combined using Miner’s rule (the Palmgren-Miner linear damage rule), which sums the ratio of applied cycles to allowable cycles, n/N, for each stress range. In Section III this sum is the cumulative usage factor, and it may not exceed 1.0.

The allowable cycles themselves come from design curves that build in deliberate margin. The group that developed the Section III method took the best-fit curve of strain-controlled test data on polished laboratory specimens and reduced it by a factor of 2 on stress or a factor of 20 on cycles, whichever was more conservative at each point. The factor of 20 was intended to cover the differences between test specimens and real components: roughly 2 for data scatter, 2.5 for size effects, and 4 for surface finish and atmospheric environment.

In equipment design, preventing fatigue failure during the designated lifespan is imperative. Components that see a very large number of load cycles can be designed for infinite life by keeping the alternating stress amplitude below the material’s endurance limit, after adjusting for mean stress, surface finish, size, and stress concentrations.

Components that see a very large number of load cycles can be designed for infinite life by keeping the alternating stress amplitude below the material’s endurance limit, after adjusting for mean stress, surface finish, size, and stress concentrations. This approach applies only to materials that show an endurance limit, such as carbon and low-alloy steels in non-corrosive environments. Aluminum alloys and most other non-ferrous metals may not have one, and a corrosive environment can remove it in steels as well. Components with stress amplitudes above the endurance limit, or made from materials without one, are designed for finite life: the number of load cycles expected in service must stay below the allowable number of cycles from the applicable fatigue design curve.

Crack Initiation, Propagation, and Final Fracture
Stage I Crack Initiation Stage II Stable Propagation Stage III Final Fracture Surface notch (stress riser) Micro-crack Cyclic load "Beach marks" — concentric crack front progression Remaining cross-section Unstable fast fracture Section too small to carry load Crack progresses from surface stress riser → stable growth (beach marks visible) → catastrophic fracture when remaining section is insufficient.
Figure 3. Three stages of fatigue crack development: initiation at a surface stress riser, stable propagation with characteristic beach marks, and final unstable fracture. Original diagram — O'Donnell Consulting Engineers.

Presently, three major approaches exist for design against fatigue failures:

  • Stress-Based Approach: This traditional method relies on nominal (average) stresses within the component area. It determines the cyclic load’s resistible nominal stress by accounting for mean stresses and adjusting for stress-raising features like grooves, holes, and fillets.
  • Strain-Based Approach: This involves a more intricate analysis of localized yielding at stress-raising points during cyclic loading.
  • Fracture Mechanics Approach: Primarily considering material toughness, crack size, and stress levels, this approach requires an understanding of crack growth and fracture.
Three Fatigue Design Approaches — Conceptual Overview
① Stress-Based Nominal (average) stress in the component S-N curve approach High-cycle applications more detail Hysteresis ② Strain-Based Localized yielding at stress-raising features ε-N curve approach Low-cycle applications crack present ③ Fracture Mechanics Crack growth rate, material toughness & stress intensity da/dN — Paris Law approach Remaining life prediction
Figure 4. The three primary fatigue design approaches — stress-based (S-N), strain-based (ε-N), and fracture mechanics — each suited to different loading regimes and levels of crack characterization. Original diagram — O'Donnell Consulting Engineers.

Fatigue Failures

Fatigue failure is the most common cause of failure in mechanical and thermal processing equipment. The basic reason relates to the myriad of the causes of failures which include deficiencies in design, materials, processing complexities and owner/operator error. Failures have been traced to corrosion, material defects, stress risers, structural hot spots, dynamic loads, variable weather conditions and other factors not adequately covered by Design, Safety and Fabrication Codes & Standards.

A formal metallurgical investigation attempts to pinpoint the exact cause of failure – faulty design, defective material, improper treatment, surface damage, or pure high cycle fatigue. Interdisciplinary teams are essential for performing most failure analyses, since the root cause is rarely the result of a single variable. In performing failure analysis we investigate operational issues, environmental factors, load paths and stresses to a system or component.

ASME Code Origin: The fatigue design life evaluation procedures in Section III of the ASME Boiler and Pressure Vessel Code were originally developed in the U.S. Naval Nuclear Program. Those involved included Dr. William J. O’Donnell Sr., Bernard F. (Bernie) Langer, W.E. (Bill) Cooper and James (Jim) Farr – who, in the late 1950’s and early 1960’s developed the initial formulation of this technology in the Tentative Structural Design Basis for Reactor Pressure Vessels and Directly Associated Components, which became known as “SDB-63.”

Section III of the ASME Code “Vessels in Nuclear Service” was the first to include specific Code rules to prevent low cycle fatigue failure. Its first edition was published in 1963. Section VIII, Division 2, “Alternate Rules for Pressure Vessels” followed in 1968.

 


References

  • “Mechanical Behavior of Materials – Engineering Methods for Deformation, Fracture and Fatigue” Norman E. Dowling, College of Engineering, Virginia Polytechnic Institute and State University, Blacksburg, VA, 1993, Prentice Hall, Englewood Cliffs, NJ. Current edition: Mechanical Behavior of Materials, 5th ed., Norman E. Dowling, Stephen L. Kampe, and Milo V. Kral, 2019, Pearson, Hoboken, NJ.
  • “Failure of Materials in Mechanical Design – Analysis, Prediction, Prevention” J.A. Collins, Ohio State University, 1981, John Wiley & Sons
  • “Metal Failures – Mechanisms, Analysis, Prevention” Arthur J. McEvily, Professor Emeritus, Department of Metallurgy and Materials Engineering, University of Connecticut, 2002, John Wiley & Sons
  • “Engineering Against Fatigue” J.H. Beynon, M.W. Brown, T.C. Lindley, R.A. Smith, & B. Tomkins, Eds., 1999, A.A. Balkema, Rotterdam, Netherlands
  • “Code Design and Evaluation for Cyclic Loading – Sections III and VIII” W.J. O’Donnell, Chapter 39 Companion Guide to the Boiler & Pressure Vessel Code – Criteria and Commentary on Select Aspects of the Boiler Pressure Vessel and Piping Codes, Second Edition, Vol. 1, Ed. K.R. Rao, 2006. ASME
  • “Design Application of Section III Fatigue Analysis”
    W.J. O’Donnell, PE, Presentation at the Workshop on Cyclic Life and Environmental Effects in Nuclear Applications, Clearwater Beach, FL, January 20, 1992. In Technical Information from Workshop on Cyclic Life and Environmental Effects in Nuclear Applications, Vol. 1, Pressure Vessel Research Council.
  • “The Initial Scope and Intent of the Section III Fatigue Design Procedures”
    W.E. Cooper, PE, Technical Information from Workshop on Cyclic Life and Environmental Effects in Nuclear Applications, Vol. 1, Clearwater Beach, FL, January 20–21, 1992, Pressure Vessel Research Council

Fatigue/Failure Analysis Resources

O’Donnell Consulting’s approach to fatigue analysis draws on decades of direct involvement in ASME code development and hands-on failure investigation across pressure vessels, piping, rotating equipment, and structural components. Depending on the nature of the problem, our engineers apply stress-based, strain-based, or fracture mechanics methods to assess remaining service life, identify root causes, and develop practical remediation strategies. Whether the engagement involves a proactive design review, an in-service crack assessment, or a post-failure forensic investigation, we bring the same depth of technical rigor to each evaluation.

(412) 835-5007

Scroll to Top