Fatigue Crack Growth & Critical Flaw Evolution
How an observed or assumed crack is propagated through the future load spectrum to establish critical flaw size, remaining cycles and inspection or operating limits.
From Current Flaw to Future Integrity
A current-condition fracture assessment answers whether a flaw is acceptable now. Crack-growth analysis asks how long that conclusion remains valid. The starting flaw, future load spectrum, crack-growth model and terminal condition are combined to predict how the defect evolves and when it reaches a defined critical state. This is the core link between inspection, damage tolerance and remaining-life decisions.
Crack-Growth Rate
For many metallic applications, growth per cycle is related to the stress-intensity-factor range. The Paris relation is the best-known representation of the mid-growth regime, but real assessments may require threshold, near-instability, mean-stress and environment corrections.
da/dN = C(ΔK)^m
Do Not Treat Paris Law as Universal
C and m are material- and environment-dependent fitted parameters, and the Paris region is only part of the full crack-growth curve. At low ΔK, threshold behaviour may suppress growth; near fracture, growth accelerates. Load ratio, closure, corrosion environment, temperature and frequency can alter the rate materially. The chosen law should be supported by data appropriate to the service regime.
Variable-Amplitude Loading
Real structures rarely experience one constant stress range. A future spectrum may include starts, shutdowns, pressure cycles, manoeuvres, vibration, thermal transients and occasional overloads. Crack growth should therefore be integrated using a representative sequence or a demonstrably conservative envelope. Sequence effects can matter because overloads can cause retardation or acceleration depending on the mechanism and modelling approach.
Crack Shape Evolution
Surface flaws generally change both depth and surface length as they grow. Assuming a fixed aspect ratio can bias the predicted stress intensity and remaining ligament. Where shape evolution materially affects life, propagate multiple crack-front points or use a recognised shape-update rule. For embedded flaws, interaction with a free surface may eventually transform the geometry into a surface crack.
Defining the Terminal Condition
The terminal flaw size should be established using the governing failure criterion rather than an arbitrary depth fraction. It may be the point where the FAD reaches its failure boundary, where J or CTOD reaches an allowable resistance, where net-section collapse occurs, where leakage becomes unacceptable, or where a code-defined limit is reached. The terminal condition can itself change with load or material degradation.
Initial Flaw Size and Inspection Uncertainty
Remaining life is highly sensitive to the starting crack size. A measured indication should therefore be increased or otherwise treated to account for sizing uncertainty in accordance with the assessment method. If the calculation begins at a detectable flaw size for inspection planning, that size must be tied to the actual inspection technique and probability of detection, not an idealised zero-size crack.
Environmental and Dwell Effects
Corrosive environment, hydrogen, high temperature and sustained dwell can introduce time-dependent crack growth in addition to cycle-dependent fatigue. A pure fatigue law may then be non-conservative. The assessment should identify whether corrosion fatigue, stress-corrosion cracking, creep crack growth or other time-dependent mechanisms need to be combined with cyclic growth.
Load Reconstruction and Future Duty
Past operating data can help calibrate the severity of the load spectrum, but remaining-life prediction is governed by the future duty being authorised. A life-extension assessment should therefore state the future cycles, transients and environment explicitly. If future duty is uncertain, analyse bounding duty cases or tie the acceptance to monitored operating limits.
Verification and Sensitivity
- Stress-intensity solution is verified — Check geometry factors or FEA against a reference case.
- Growth integration is step-size independent — Crack increment should not materially change predicted life.
- Critical flaw is recalculated as geometry changes — Do not assume one fixed terminal size if load or ligament changes.
- Dominant uncertainties are varied — Initial flaw size, C/m, spectrum severity and toughness usually deserve sensitivity checks.
- Units and cycle definitions are controlled — Small unit errors in crack-growth constants can produce orders-of-magnitude life errors.
Life Is a Distribution of Evidence, Not a Stopwatch
Crack-growth predictions can be highly sensitive to uncertain inputs, so reporting an exact remaining-cycle count can imply false precision. A defensible conclusion presents the central or conservative prediction, sensitivity range, inspection assumptions and terminal criterion. The result should support an actionable inspection or operating decision rather than merely produce a life number.
The most important crack-growth input is often the starting flaw size, not the numerical integration method.
Engineering judgement — governing sensitivities
For Fatigue Crack Growth & Critical Flaw Evolution, the most useful review question is not simply whether the solver has produced a plausible contour or scalar result, but whether the model preserves the evolution from the detected or assumed flaw to the limiting state under the actual variable-amplitude service history. Closure, retardation, threshold and environment can matter when the growth interval is long. This is where apparently small modelling choices can change the engineering conclusion. The analyst should identify the variables that can move the governing response, separate physical uncertainty from deliberate conservatism, and show that the selected modelling fidelity is proportionate to the decision being supported. Where the response is close to an acceptance boundary, sensitivity cases should bracket credible changes rather than apply arbitrary percentage perturbations.