Compendium

Fracture Mechanics — Engineering Reference

37 min read

Fracture Mechanics — Engineering Reference

1. At a glance

Fracture mechanics quantifies the behaviour of pre-existing cracks under load. Classical strength-of-materials (see mechanics-of-materials) asks will the section yield or rupture in the bulk? — it treats the body as defect-free. Fracture mechanics asks the question that actually controls service failure of real welded, cast, forged, and additively-manufactured hardware: given a crack of size a, will it grow, and at what load?

The discipline splits into three regimes:

  • LEFM (Linear Elastic Fracture Mechanics) — small-scale yielding, K-based. Applies to brittle metals, ceramics, high-strength steels, glass, polymers below T_g, and thick sections where plane-strain constraint suppresses crack-tip plasticity.
  • EPFM (Elastic-Plastic Fracture Mechanics) — J-integral (Rice 1968) or CTOD (Wells 1961). Applies to ductile metals, weldments, ferritic steel in the upper-shelf regime, and any case where the plastic zone is not small relative to the crack length and remaining ligament.
  • Subcritical crack growth — fatigue (Paris law), stress-corrosion cracking (SCC) above K_ISCC, creep crack growth above ~0.4 T_melt, hydrogen-assisted cracking. The crack grows slowly under stresses far below the static fracture toughness, then accelerates to catastrophic failure when K reaches K_C.

Where it sits in the design stack: mechanics of materials → fatigue analysis → fracture mechanics → damage tolerance / fitness-for-service / risk-based inspection. It is the explicit failure-control discipline behind aerospace damage tolerance (FAR 25.571, MIL-A-83444), nuclear pressure-boundary integrity (ASME BPVC Sec. XI), oil-and-gas pipeline FFS (API 579-1), and offshore structure assessment (BS 7910). Modern Risk-Based Inspection (RBI per API 580 / 581) uses fracture-mechanics output — predicted crack growth from a postulated initial defect — as the dose-response curve that drives inspection intervals.

In one sentence: fracture mechanics turns a found crack into a predicted remaining life.

2. Why it matters — the case-study spine

Most of the named twentieth-century structural disasters were preventable in retrospect with fracture-mechanics analysis. The lessons codify into current standards:

  • Liberty ships (1942–1946). ~1500 hull failures, ~200 catastrophic. Welded T-1 steel went brittle below ~10 °C; transition temperature was above winter sea-water. Driver: notch-sensitivity in the unstable BCC steel + welded continuity (no riveted-plate crack arrestor). Codified into Charpy CVN minimum impact requirements (ASTM E23) and ABS/Lloyd’s plate-grade temperature limits.
  • de Havilland Comet (1954). Two Comet 1 jets disintegrated in cruise over the Mediterranean (G-ALYP, G-ALYY). Driver: fatigue crack from the corner of a square ADF antenna cutout in the pressurised fuselage. K_t ≈ 7–8 at the corner; cyclic σ_hoop drove crack to critical size in ~3000 flights. Result: round cutouts mandatory on pressurised airframes; cyclic pressurisation test (the Comet 1A water-tank test) became standard certification.
  • Aloha Airlines 243 (1988). 5 m of upper fuselage tore off a 19-year-old 737-200 at FL240 over Hawaii. Driver: multi-site damage (MSD) — many small fatigue cracks along the lap-joint rivet line, all sub-detectable individually, that linked up. Drove FAA Aging Aircraft Program and the Widespread Fatigue Damage (WFD) rules in 14 CFR 26.21.
  • Alexander L. Kielland (1980). Norwegian semi-sub flotel capsized in the North Sea, 123 dead. Driver: fatigue crack initiated at a 6-mm fillet weld on a hydrophone bracket on a bracing leg, propagated through the leg, brace separated, platform lost stability. Drove offshore-fatigue assessment per DNV-RP-C203 and NORSOK N-006.
  • F-111 (1969). Wing-pivot fitting failure during low-level training, USAF lost the aircraft. Driver: forging-defect crack in D6AC steel. First production aircraft to be redesigned with explicit damage-tolerance analysis (DTA). Drove MIL-A-83444 (USAF damage tolerance requirements) and later AFGROW.

The common thread: a detectable, sub-critical crack existed; the operator did not have a quantitative tool to predict how long it would remain sub-critical. Fracture mechanics is that tool.

3. First principles

3.1 Stress concentration vs stress intensity

A stress concentration factor K_t (dimensionless, see mechanics-of-materials §3.8) multiplies nominal stress at a hole or fillet of finite radius. For a circular hole in a wide plate, K_t = 3 (Kirsch 1898). As the notch radius shrinks toward zero, K_t → ∞ — which is physically nonsensical and tells us K_t is the wrong descriptor for a sharp crack. The right descriptor is the stress-intensity factor K, with units MPa·√m (ksi·√in), introduced by Irwin (1957):

σ_ij(r, θ) = K / √(2π·r) · f_ij(θ) + (higher-order terms)

The crack-tip stress field has a 1/√r singularity; K is the amplitude of that singularity. K depends on geometry, applied load, and crack size — not on material. The material’s resistance to crack extension is the fracture toughness K_C (plane stress) or K_IC (plane strain), a measured property with the same units.

3.2 Griffith energy balance (1921)

A.A. Griffith, working on glass for the Royal Aircraft Establishment, derived the first quantitative fracture criterion from energy:

G = dU / dA       (energy release rate per unit crack-area extension)
G = σ²·π·a / E    (central crack, infinite plate, plane stress)
G_C = 2·γ_s       (Griffith criterion, brittle solids — only surface-energy term)

Crack extends when G ≥ G_C. For ductile metals, Orowan (1948) and Irwin (1948) added a plastic-work term γ_p that dominates γ_s by 3–4 orders of magnitude:

G_C = 2·(γ_s + γ_p)

For a typical structural steel, γ_s ≈ 1 J/m², γ_p ≈ 10⁴–10⁵ J/m² — almost all of the fracture energy is plastic work in the crack-tip process zone.

3.3 Irwin K — three modes

Irwin (1957) showed G and K are equivalent in LEFM, related by:

G = K_I² / E'           E' = E (plane stress)
G = K_I² / E' + K_II² / E' + K_III² / (2G)    E' = E / (1 − ν²) (plane strain)

Three fracture modes are kinematically independent:

  • Mode I — tensile opening. The crack faces separate normal to the crack plane. The dominant failure mode in service.
  • Mode II — in-plane shear. Crack faces slide normal to the crack front but in the crack plane.
  • Mode III — out-of-plane shear (anti-plane shear). Crack faces slide parallel to the crack front (tearing).

Most engineering analysis is Mode I. Mixed-mode loading uses combined criteria (§7).

3.4 K_IC — the plane-strain toughness

The plane-strain fracture toughness K_IC is the minimum fracture toughness of a material — a true material property — measured under ASTM E399 conditions where the plastic zone is small relative to specimen thickness. Validity check:

B, a, (W − a) ≥ 2.5 · (K_IC / σ_y)²

If this is not satisfied, the measured value is an apparent toughness K_Q, not K_IC. Typical K_IC values for engineering metals:

Materialσ_y (MPa)K_IC (MPa·√m)K_IC (ksi·√in)Source
4340 steel, Q&T 200°C17005045.5ASTM E399 testing
4340 steel, Q&T 425°C14509081.9ASTM E399 testing
AISI 1018 (mild)250220 (J→K)200Upper-shelf E1820
Stainless 304L215> 200> 182Upper-shelf E1820
7075-T6 aluminum5032926.4MMPDS-19
7075-T73 aluminum4353330.0MMPDS-19
2024-T351 aluminum3253733.7MMPDS-19
Ti-6Al-4V (annealed)8807568.3MIL-HDBK-5J
Ti-6Al-4V (β-STOA)850110100.1MMPDS-19
Maraging 25017209081.9E399 testing
Soda-lime glass—0.70.64Anderson 2017
Al₂O₃ ceramic—3–52.7–4.6ASTM C1421
WC–Co cermet—10–189.1–16.4ASTM B771

Higher strength → lower toughness is the universal trade-off. The 4340 above illustrates: tempering 200 °C gives 1700 MPa / 50 MPa·√m; tempering 425 °C gives 1450 MPa / 90 MPa·√m.

3.5 Plane stress vs plane strain

  • Plane strain (thick section, B large) — through-thickness contraction is suppressed; σ_zz = ν·(σ_xx + σ_yy); higher hydrostatic constraint → smaller plastic zone → lower apparent toughness. K_IC is plane-strain.
  • Plane stress (thin section, B small) — σ_zz = 0; lower constraint → larger plastic zone → higher apparent toughness K_C. K_C falls with increasing thickness until plane-strain K_IC is reached at the transition thickness ~2.5·(K_IC/σ_y)².

Practical consequence: thinner is “tougher” in apparent K, but the thin-section failure mode is shear-lip / slant fracture rather than flat plane-strain cleavage. Aircraft skins are designed in the high-K_C plane-stress regime; pressure vessel walls in the lower-K_IC plane-strain regime.

3.6 Plastic-zone size (Irwin first-order estimate)

r_p = (1 / 2π) · (K_I / σ_y)²      plane stress
r_p = (1 / 6π) · (K_I / σ_y)²      plane strain

Small-scale yielding (SSY) — the condition for LEFM validity — requires r_p ≪ a, B, (W − a). The factor-of-2.5 in ASTM E399 §3.4 is the SSY threshold for K-dominance.

3.7 J-integral (Rice 1968)

For elastic-plastic conditions where r_p is no longer small, K is invalid but J is well-defined as a path-independent line integral around the crack tip:

J = ∮_Γ (W·dy − T_i · ∂u_i/∂x ds)

where W is strain-energy density, T_i is the traction vector, u_i is displacement, ds is arc length along contour Γ. Two key properties:

  1. Path independence for elastic (linear or non-linear) materials — Rice’s central result.
  2. Equivalence to K in LEFM: J = G = K_I²/E’ under small-scale yielding.

In the EPFM regime, J characterises the crack-tip field via the HRR singularity (Hutchinson 1968; Rice and Rosengren 1968):

σ_ij ~ (J / (α·ε_y·σ_y·I_n·r))^(1/(n+1))    Ramberg-Osgood with hardening exponent n

J_IC (ASTM E1820) is the crack-initiation toughness; J-R curves give resistance vs Δa. Conversion back to an equivalent K_J:

K_J = √(J · E')      E' = E (plane stress), E / (1 − ν²) (plane strain)

3.8 CTOD (Wells 1961)

The Crack-Tip Opening Displacement δ_t is a physically measurable opening between the deformed crack faces. Relations (small-scale yielding):

δ_t = K_I² / (m · σ_y · E')        m ≈ 1 (plane stress), 2 (plane strain)
δ_t = J / (m · σ_y)                EPFM, same m

CTOD design (BS 7910 Annex K) is the British/European complement to the American J-based approach; CTOD δ_C is measured per ASTM E1290 / ISO 12135.

3.9 T-stress and constraint quantification

The leading-order crack-tip field has a second non-singular term — the T-stress — parallel to the crack plane:

σ_xx = K_I / √(2π·r) · cos(θ/2)·(1 − sin(θ/2)·sin(3θ/2)) + T + O(√r)
σ_yy = K_I / √(2π·r) · cos(θ/2)·(1 + sin(θ/2)·sin(3θ/2)) + O(√r)

T can be negative (tensile-dominated specimens like deep-cracked SENB) — low constraint, larger plastic zone, higher apparent toughness — or positive (shallow-cracked SENT, M(T) tension specimens) — high constraint, plane-strain-like behavior. K-T characterization of the crack tip, or alternatively the Q-parameter (O’Dowd and Shih 1991) under EPFM conditions, lets engineers transfer toughness measurements between specimen geometries and structural components with different constraint. Real structural cracks (surface flaws under bending, leak-before-break through-wall cracks) typically have constraint far below the ASTM E399 standard deeply-notched specimen, so K_IC tests are conservative for those applications.

3.10 R-curve and tearing instability

Fracture resistance K_R or J_R generally rises with crack growth Δa (the “R-curve”). Instability:

dG/da  ≥  dR/da       (instability condition, stress-controlled loading)
dJ/da  ≥  dJ_R/da     (J-tearing instability, Paris-Hutchinson)

The tearing modulus T = (E/σ_y²)·dJ_R/da characterises slope of J-R curve; higher T means more stable tearing.

4. Stress-intensity-factor handbook (Mode I)

Geometry-only correction factors Y = K·√(π·a) / (σ·√(π·a)). Compiled from Tada/Paris/Irwin Stress Analysis of Cracks Handbook, 3rd ed.

GeometryK_I expressionNotes
Central through-crack in infinite plateK_I = σ·√(π·a)Y = 1.000
Central crack in finite plate, 2W widthK_I = σ·√(π·a)·sec(π·a/W)^(1/2)Feddersen secant; Y → 1 as a/W → 0
Edge crack in semi-infinite plateK_I = 1.1215·σ·√(π·a)Y = 1.1215 (free surface)
Single-edge-notched bend (SENB, 3-pt)K_I = (P·S / (B·W^1.5))·f(a/W)ASTM E399 / E1820 specimen
Compact tension (CT)K_I = (P / (B·√W))·f(a/W)ASTM E399 / E1820 specimen
Penny-shaped (embedded circular) crackK_I = (2/π)·σ·√(π·a)Y = 2/π ≈ 0.637
Semi-elliptical surface flawNewman–Raju (1981) closed-form Y(a/c, a/t, φ)Pipelines, pressure-vessel walls
Crack from a circular hole (Bowie 1964)K_I = σ·F_B(L/r)·√(π·L)L = crack length from hole edge
Through-wall axial crack in pressurised cylinderK_I = σ_hoop·M_T(a, R, t)·√(π·a)Folias bulging factor M_T
Through-wall circumferential crack in pipeK_I = σ_axial·M_C(a, R, t)·√(π·a)Used in B31G, API 1163
Point load P on crack faces, infinite plateK_I = P / √(π·a)Green’s function building block
Wedge-opening load on edge crackK_I = (2·P / √(π·a))·g(c/a)DCB, wedge-opening-load (WOL)

Y(a/W) polynomials for the standard specimens (ASTM E399 §A4 verbatim):

  • CT: f(a/W) = ((2 + a/W) / (1 − a/W)^1.5) · [0.886 + 4.64(a/W) − 13.32(a/W)² + 14.72(a/W)³ − 5.6(a/W)⁴]
  • SENB: f(a/W) = (3·(a/W)^0.5 · [1.99 − (a/W)(1−a/W)(2.15 − 3.93(a/W) + 2.7(a/W)²)]) / (2·(1 + 2·a/W)·(1 − a/W)^1.5)

For arbitrary geometry use FE methods (§8) — the J-integral domain extraction in Abaqus, the displacement-correlation technique in ANSYS, or the M-integral in FRANC3D.

5. Fatigue crack growth

5.1 Paris law (Paris & Erdogan 1963)

da/dN = C · (ΔK)^m         ΔK = K_max − K_min = ΔS · √(π·a) · Y

C and m are material constants for a given R-ratio R = K_min/K_max, environment, and frequency. da/dN in m/cycle, ΔK in MPa·√m. The exponent m sits in 2–4 for most metals; m = 3 is a useful default.

MaterialC (m/cycle, MPa·√m units)mR-ratioSource
Ferritic-pearlitic steel6.9 × 10⁻¹²3.0R ≈ 0BS 7910 §8
Martensitic steel1.4 × 10⁻¹¹2.25R ≈ 0BS 7910 §8
Austenitic stainless 3045.6 × 10⁻¹²3.25R ≈ 0BS 7910 §8
2024-T3 aluminum1.6 × 10⁻¹¹3.59R = 0.1MMPDS-19
7075-T6 aluminum1.5 × 10⁻¹¹3.0R = 0.1NASGRO database
Ti-6Al-4V1.0 × 10⁻¹¹3.2R = 0.1NASGRO database

Caveat — these constants are units-bracketed. Many handbooks tabulate C for ΔK in ksi·√in or MN·m^(−3/2); always check.

5.2 The three regimes of da/dN vs ΔK (log-log)

  • Region I — threshold. Below ΔK_th cracks do not grow under continued cycling. ΔK_th ≈ 5–10 MPa·√m for ferritic steels, 2–4 MPa·√m for Al alloys, ~3 MPa·√m for Ti alloys, all at R = 0. Falls with increasing R.
  • Region II — Paris regime. Linear in log-log; Paris equation applies. 10⁻⁹ < da/dN < 10⁻⁶ m/cycle.
  • Region III — fast fracture approach. da/dN rises sharply as K_max → K_IC (or J_IC). Captured by Forman:
da/dN = C · (ΔK)^m / ((1 − R) · K_C − ΔK)        Forman 1967

5.3 Closure, R-ratio effects, NASGRO equation

Crack closure (Elber 1970) — fatigue crack faces contact each other before zero applied load is reached, reducing the effective ΔK:

ΔK_eff = K_max − K_op             where K_op > K_min

Plasticity-induced closure (Elber), roughness-induced closure (Suresh & Ritchie 1982), and oxide-induced closure (Suresh, Zamiski, Ritchie 1981) explain why short cracks grow faster than long cracks at nominally identical ΔK — short cracks have not yet built up a closure wake.

NASGRO equation (Forman & Newman, NASGRO Reference Manual) — the industry-standard fit for aerospace damage tolerance, embedded in NASA NASGRO and USAF AFGROW:

da/dN = C · [((1 − f) / (1 − R)) · ΔK]^n · (1 − ΔK_th/ΔK)^p / (1 − K_max/K_C)^q

with f the closure function and p, q tailored fit-exponents. Reduces to Paris in Region II at high ΔK with no threshold or fast-fracture proximity.

5.4 Subcritical environmentally-assisted cracking

Beyond pure fatigue, three environment-driven mechanisms grow cracks at constant or quasi-static load below K_IC:

  • Stress-corrosion cracking (SCC). Threshold K_ISCC below which growth stops; above threshold, da/dt vs K curves typically show three regions (I rising, II plateau, III rising near K_IC). Common pairs: high-strength steel + chlorides / H₂S (sour service); austenitic stainless + hot chlorides; α-β titanium + methanol / N₂O₄; Al 7000-series + chloride. Plateau da/dt for high-strength steel in seawater: 10⁻⁹ to 10⁻⁷ m/s — orders of magnitude faster than fatigue at the same K. Controlled by NACE MR0175 / ISO 15156 (sour service: cap material hardness at 22 HRC for carbon steel).

  • Hydrogen-induced cracking (HIC, HE). Atomic hydrogen diffuses to crack-tip triaxiality maxima, lowers cohesive strength (HEDE) or enhances local plasticity (HELP); K_IH thresholds 10–30 % of K_IC for σ_u > 1000 MPa steel. Sources: cathodic protection, galvanising, electroplating, weld-hydrogen, sour service. Mitigated by low-hydrogen consumables (E7018-H4), post-weld bake (200–250 °C), and capping σ_u.

  • Creep crack growth. Above ~0.4 T_melt, the relevant driving force becomes the C* integral (steady-state creep analogue of J) or C_t (transient). da/dt = D · (C*)^φ with φ ≈ 0.8–0.9. Used for high-temperature plant assessment via R5 (UK CEGB / EDF) and API 579 Part 10.

5.5 Short-crack effect

For cracks shorter than ~10 grain diameters (microstructurally short) or shorter than the cyclic plastic-zone size (mechanically short), the long-crack ΔK_th over-predicts threshold — short cracks propagate at apparent ΔK below the long-crack threshold. Kitagawa-Takahashi diagram (1976) plots fatigue strength vs crack length with a transition between smooth-bar fatigue limit (Δσ_e) and long-crack threshold (ΔK_th):

ΔK_th = ΔK_th,∞ · √(a / (a + a₀))         El Haddad 1979 correction
a₀ = (1/π) · (ΔK_th / Δσ_e)²

For ferritic steels a₀ is typically 50–500 μm; for high-strength Al ~10–50 μm.

6. Worked examples

6.1 Example A — brittle plate sizing under static load (LEFM)

Problem. A wide plate of 4340 steel Q&T at 200 °C (σ_y = 1700 MPa, K_IC = 50 MPa·√m) has an edge crack a = 5 mm detected by ultrasonic inspection. Determine the maximum allowable nominal tensile stress with a factor of safety of 2 against fracture. Confirm plane-strain validity for a 25-mm-thick plate.

Step 1 — Stress-intensity for edge crack.

K_I = 1.1215 · σ · √(π·a)
    = 1.1215 · σ · √(π · 0.005 m)
    = 1.1215 · σ · 0.1253
    = 0.1405 · σ

Step 2 — Critical stress at fracture (K_I = K_IC).

σ_critical = K_IC / (1.1215 · √(π·a))
           = 50 MPa·√m / 0.1405 √m
           = 356.0 MPa

Step 3 — Allowable stress (FoS = 2).

σ_allow = σ_critical / 2 = 178 MPa

Step 4 — Plane-strain validity.

B_min = 2.5 · (K_IC / σ_y)² = 2.5 · (50 / 1700)² · 1 m = 2.5 · 8.65×10⁻⁴ m = 2.16 mm

The 25-mm plate exceeds B_min by an order of magnitude — plane-strain assumption is solid. Result: the plate may be loaded to 178 MPa nominal, which is well below σ_y (~1700 MPa). The detected crack — not yield — governs the design. Note also: σ_allow / σ_y = 178 / 1700 = 0.105, so a defect-tolerant design has thrown away ~90 % of the material’s “strength” — this is the cost of high-strength, low-toughness steel.

6.2 Example B — fatigue life from initial detected flaw

Problem. 7075-T6 aluminum panel, edge crack a_i = 0.5 mm, cyclic remote stress ΔS = 100 MPa (R = 0), K_C = 30 MPa·√m, Paris constants C = 1.5×10⁻¹¹, m = 3.0 (m/cycle, MPa·√m units), Y = 1.1215 (edge crack), σ_max = 100 MPa. Estimate cycles to failure N_f.

Step 1 — Critical crack size. At failure, K_max = K_C:

K_C = σ_max · Y · √(π · a_c)
30 = 100 · 1.1215 · √(π · a_c)
√(π · a_c) = 30 / 112.15 = 0.2675 √m
π · a_c = 0.0716 m
a_c = 22.8 mm

Step 2 — Set up Paris ODE.

da / dN = C · (ΔK)^m = C · (ΔS · Y · √(π·a))^m
       = 1.5×10⁻¹¹ · (100 · 1.1215 · √(π · a))^3
       = 1.5×10⁻¹¹ · (112.15)^3 · (π · a)^1.5
       = 1.5×10⁻¹¹ · 1.411×10⁶ · π^1.5 · a^1.5
       = 1.5×10⁻¹¹ · 1.411×10⁶ · 5.568 · a^1.5
       = 1.178×10⁻⁴ · a^1.5   [m/cycle, a in m]

Step 3 — Separate and integrate.

∫_{a_i}^{a_c} da / a^1.5 = ∫_0^{N_f} 1.178×10⁻⁴ dN
[ −2·a^(−0.5) ]_{0.0005}^{0.0228} = 1.178×10⁻⁴ · N_f
−2·(1/√0.0228) + 2·(1/√0.0005) = 1.178×10⁻⁴ · N_f
−13.24 + 89.44 = 1.178×10⁻⁴ · N_f
76.20 = 1.178×10⁻⁴ · N_f
N_f = 6.47×10⁵ cycles ≈ 650 000 cycles

Comment. The integral is heavily weighted toward the initial crack length (a^1.5 in denominator). Roughly 80 % of the life is consumed growing from 0.5 mm to ~5 mm. Doubling a_i (e.g. due to NDT under-sizing) cuts N_f by a factor of ~2.8. This is why damage tolerance assumes the initial flaw is the largest the inspection method could miss — typically 1–2 mm for shop-floor UT, larger for in-service inspection.

6.3 Example C — J-integral / CTOD with ductile blunting (EPFM)

Problem. AISI 4140 quenched-and-tempered to σ_y = 700 MPa, σ_u = 900 MPa, E = 210 GPa, ν = 0.30. J-IC (ASTM E1820, SENB specimen) measured at 200 kJ/m². Find the equivalent plane-strain K_J and the CTOD δ_t at crack initiation.

Step 1 — Plane-strain modulus.

E' = E / (1 − ν²) = 210 000 / (1 − 0.09) = 210 000 / 0.91 = 230 769 MPa = 230.8 GPa

Step 2 — K_J from J.

K_J = √(J · E') = √(200 000 J/m² · 230.8×10⁹ Pa)
    = √(4.615×10¹⁶ Pa²·m)
    = 2.148×10⁸ Pa·√m
    = 214.8 MPa·√m

Step 3 — CTOD via J / (m·σ_y). Plane strain m = 2; some BS 7910 formulations use the flow stress σ_f = (σ_y + σ_u)/2 = 800 MPa:

δ_t = J / (m · σ_y) = 200 000 / (2 · 700×10⁶) = 1.43×10⁻⁴ m = 0.143 mm
δ_t (flow-stress) = 200 000 / (2 · 800×10⁶) = 0.125 mm

Comment. This 4140 has much higher resistance to fracture than the 4340 of Example A despite similar nominal strength — at K_J ≈ 215 MPa·√m it tolerates 18× larger cracks at the same stress. This is the practical case for spec’ing intermediate-temper Q&T steel (4140 425 °C temper) over very-low-temper 4340 for structural applications: the toughness gain outweighs the modest yield-strength loss.

6.4 Example D — leak-before-break (LBB) check on a pressurised cylinder

Problem. ASME-stamped austenitic stainless 304L pressure vessel, mean radius R = 500 mm, wall thickness t = 25 mm, internal pressure p = 8 MPa, σ_y = 215 MPa, K_JC = 200 MPa·√m (upper-shelf). Confirm LBB: a through-thickness axial crack must remain stable at a length 2c sufficient for visible leak detection (typically 2c_leak ≈ 2·t = 50 mm minimum).

Step 1 — Hoop stress.

σ_hoop = p · R / t = 8 · 500 / 25 = 160 MPa

Step 2 — Folias bulging factor for axial through-wall crack (length 2c = 50 mm, half-length c = 25 mm):

λ = c / √(R · t) = 0.025 / √(0.500 · 0.025) = 0.025 / 0.1118 = 0.2236
M_T = √(1 + 1.255·λ² − 0.0135·λ⁴)
    = √(1 + 1.255·0.0500 − negligible)
    = √(1.0628) = 1.031

Step 3 — K_I at the postulated leak length.

K_I = σ_hoop · M_T · √(π · c)
    = 160 · 1.031 · √(π · 0.025)
    = 165.0 · 0.2802
    = 46.2 MPa·√m

Step 4 — Critical crack half-length (LBB margin). At K_I = K_JC:

c_crit = (K_JC / (σ_hoop · M_T_crit))² / π

Iterating with M_T at c_crit (M_T grows with c, so this requires iteration; first pass with M_T = 1.031):

c_crit,0 = (200 / (160 · 1.031))² / π = (1.213)² / π = 0.469 m

At that size M_T is much larger — repeating with M_T ≈ 5 gives c_crit ≈ 0.094 m. Converging at ~85 mm.

Step 5 — LBB margin. 2c_leak = 50 mm; 2c_crit ≈ 170 mm. Ratio c_crit / c_leak ≈ 3.4 — comfortable LBB margin. Result: the vessel will leak through a 50-mm through-wall crack at K_I = 46.2 MPa·√m (well below K_JC = 200), giving operators time to detect pressure loss and depressurise before reaching the critical length where the wall would unzip in a longitudinal rupture.

Comment. LBB is the basis of nuclear primary-circuit piping integrity (ASME III-NB, US NRC GDC-4). The assessment must also verify (a) crack opening sufficient for leakage flow rate above the detection limit; (b) no environmental degradation pathway (no SCC, no IGSCC); (c) sufficient toughness in the worst-case heat-affected zone of welds, not just in the parent metal.

7. Failure assessment diagram (FAD) — the unified LEFM/EPFM screen

The Failure Assessment Diagram (FAD) — Dowling and Townley (1975), R6 procedure (CEGB then BSI), now embedded in BS 7910 Annex 7 and API 579-1 Part 9 — is the practical engineer’s tool for FFS assessment. It collapses LEFM and EPFM behavior onto a single 2-D map:

K_r = K_I / K_mat                  vertical axis (brittle fracture proximity)
L_r = σ_ref / σ_y                  horizontal axis (plastic collapse proximity)

Assessment point (L_r, K_r) is computed for the postulated flaw; it must lie inside the FAD curve f(L_r). Three FAD options progressively more refined:

  • Option 1 (no stress-strain data): f(L_r) = (1 − 0.14·L_r²) · [0.3 + 0.7·exp(−0.65·L_r⁶)] for L_r ≤ L_r,max
  • Option 2 (stress-strain curve known): f(L_r) = [E·ε_ref / (L_r·σ_y) + L_r³·σ_y / (2·E·ε_ref)]^(−0.5)
  • Option 3 (full FE analysis): J / J_e curve from elastic-plastic FE.

L_r,max = σ_f / σ_y where flow stress σ_f = (σ_y + σ_u)/2 caps the cut-off. Three failure modes are recovered as limits:

  • K_r = 1, L_r ≈ 0 → pure brittle fracture (LEFM, K_IC governs)
  • K_r ≈ 0, L_r = L_r,max → pure plastic collapse (limit load governs)
  • interior of curve → elastic-plastic interaction; J/CTOD analysis required

The FAD is the single most-used assessment tool in pipeline integrity and pressure-vessel FFS practice; CRACKWISE and Signal-FFS automate it.

8. Engineering judgement — damage tolerance philosophy

Three philosophies are simultaneously in use across industry; choosing among them is the single highest-leverage design decision:

  1. Safe-life. Component is designed and tested to last a finite service life with no cracks initiating to a detectable size. Demonstrated by full-scale fatigue test to 2–4× design life. No in-service inspection between overhauls. Used on: helicopter rotor heads and dynamic components, landing-gear (selected items), turbine-engine rotors (LCF-limited). Backstop: scatter factor on test life (Mil-HDBK-516 prescribes ×4 typical for rotorcraft dynamic components).

  2. Fail-safe. Multiple load paths; failure of one member is detected and corrected before the redundant members fail. Pioneered by Pugsley and Frost at RAE in the 1950s. Used on: pressurized fuselage skin (tear-strap bays), redundant lugs, multi-spar wings. Demonstrated by residual-strength test with the primary path severed. Cost: weight from redundancy.

  3. Damage tolerance. Assumes a manufacturing or service-induced flaw exists at the largest size that NDE could miss; demonstrates the structure can carry limit load with that flaw and that fatigue crack growth from that flaw to the critical size will be detected by scheduled in-service inspection. Codified for USAF since MIL-A-83444 (1974), for FAA transports since FAR 25.571 amendment 45 (1978). The bulk of modern airliner structure (737, 777, A320, A350) is damage-tolerant.

Inspection-interval rule (USAF MIL-STD-1530D):

T_insp = (N_f(a_NDE → a_c)) / N_factor       N_factor typically 2

The crack must grow from the smallest reliably-detected flaw a_NDE to the critical size a_c in at least two inspection intervals, so that any single missed inspection still allows the next one to catch the crack before failure. Leak-before-break (LBB) is the pressure-vessel analogue: a through-thickness axial crack must reach a length stable under operating pressure (so the vessel leaks visibly) before reaching the critical length at which it would unzip catastrophically — used in nuclear primary-circuit piping per ASME BPVC III-NB and 10 CFR 50 GDC-4.

IndustryDominant philosophyCode reference
Commercial transport airframesDamage tolerance + WFDFAR 25.571 amdt 96; AC 25.571
Military airframes (USAF)Damage toleranceMIL-STD-1530D; ASIP
Rotorcraft dynamic componentsSafe-lifeMil-HDBK-516C §7.7
Turbine-engine rotorsSafe-life + RFCFAR 33.70; ENSIP
Pressure vesselsLimit load + FFSASME BPVC VIII; API 579
Nuclear primary pipingLBBASME III-NB; 10 CFR 50
Oil & gas pipelinesFFS + RBIAPI 579, API 580, B31.G
Bridges (steel)Fatigue category + FCMAASHTO LRFD §6.6

9. Charpy-to-K_IC correlations and the Master Curve

K_IC testing per ASTM E399 is expensive (large specimens, instrumented machine, valid-test ratio ~40 %). Industrial practice uses cheap Charpy V-notch (CVN, ASTM E23) impact data — a small specimen broken by a swinging pendulum — and correlates back to K_IC via two long-standing empirical relations:

  • Rolfe-Novak-Barsom (upper-shelf, ferritic steels):

    (K_IC / σ_y)² = 5 · (CVN / σ_y − 0.05)        [SI: K MPa·√m, σ_y MPa, CVN J]

  • Barsom-Rolfe transition (intermediate temperatures, ferritic steels):

    K_IC = 14.6 · √CVN                              [SI: K MPa·√m, CVN J]

These are screening tools — actual fracture testing is needed for damage tolerance work. The modern alternative is the Master Curve (Wallin 1984; ASTM E1921), which fits a universal three-parameter Weibull distribution to fracture-toughness data of ferritic steels in the transition region:

K_JC,median = 30 + 70 · exp(0.019·(T − T_0))    [SI: K_JC MPa·√m, T °C]

T_0 is the reference temperature where median K_JC = 100 MPa·√m for a 1-T specimen, found from a small ASTM E1921 dataset (6–12 specimens). The framework gives statistically rigorous lower-bound toughness curves for nuclear pressure-vessel integrity assessment, weld qualifications, and aging-management surveillance.

10. Welded-joint fracture assessment — practical workflow

Welds dominate real-world fracture and fitness-for-service work because they concentrate three risk factors simultaneously: geometric stress concentrators (toe radius), residual tensile stress (≈ σ_y), and heterogeneous metallurgy (parent / HAZ / weld-metal toughness gradients). Standard BS 7910 workflow:

  1. Categorise the flaw — planar (crack-like: LOF, cracks, undercut) or volumetric (porosity, slag). Only planar flaws get fracture-mechanics treatment; volumetric flaws are screened by workmanship limits in the relevant fab code (ASME IX, AWS D1.1).
  2. Size the flaw by NDE (UT, PAUT, ToFD, RT) and bound the dimensions with NDE uncertainty.
  3. Recharacterise the actual shape to an idealised semi-elliptical surface crack or through-wall crack per BS 7910 §6.4.
  4. Assemble the loads — primary (pressure, weight, thermal), secondary (residual stress, thermal), and seismic / accidental as applicable.
  5. Compute K_I using BS 7910 Annex M solutions or FE.
  6. Compute σ_ref for the L_r axis using limit-load solutions in BS 7910 Annex P (pressure-vessel, plate, pipe-girth-weld closed forms).
  7. Place the (L_r, K_r) point on the FAD (§7); inside curve = acceptable.
  8. Sensitivity / partial safety factors — BS 7910 §5 prescribes PSFs depending on consequence-of-failure category and reserve factor (RF) for stress and toughness inputs. RF target typically 1.5–2.0.

The same workflow underpins API 579 Part 9 Level 2 assessments. CRACKWISE (TWI) and Signal-FFS (DNV) automate it.

11. Probabilistic fracture mechanics

Deterministic FFS uses lower-bound inputs and computes a yes/no answer. Probabilistic fracture mechanics (PFM) treats K_IC, a_initial, da/dN constants, and applied loads as distributions and computes a probability of failure (PoF). Standard frameworks:

  • PRAISE (Lawrence Livermore, 1980s onward) — piping reliability via crack growth simulation.
  • PRO-LOCA / xLPR (US NRC + EPRI) — leak-before-break uncertainty quantification.
  • STAR6 and VOCALIST (EU, EDF) — nuclear pressure-vessel integrity.
  • PROMETHEUS (NRG, Netherlands) — aging-management PoF.

Inputs:

  • K_IC ~ 3-parameter Weibull (ASTM E1921 Master Curve).
  • a_initial ~ POD-curve-derived (MIL-HDBK-1823); log-normal typical.
  • da/dN ~ log-normal on C, normal on m (Virkler dataset is the canonical fit calibration).
  • Load ~ histogram of measured spectra (rainflow counted, see fatigue-analysis).

Output: PoF vs time. ALARP (As Low As Reasonably Practicable) target for nuclear ~10⁻⁶/yr per component; for oil-and-gas pipelines per API 581 ~10⁻⁴/yr is common. Monte Carlo (10⁶–10⁸ trials) and importance sampling are the workhorses; first-order reliability methods (FORM/SORM) accelerate the rare-event tail.

12. Composite and additively-manufactured materials

The classical LEFM/EPFM toolkit was built around metals. Two material classes need modification:

Fibre composites. Self-similar crack growth is the exception, not the rule — delamination, fibre pull-out, fibre bridging, transverse matrix cracking, and ply splitting compete. Methods:

  • VCCT (virtual crack closure technique) — Mode I/II/III G computed from FE nodal forces and displacements (Rybicki & Kanninen 1977). Standard for delamination in laminated composites; implemented in Abaqus, ANSYS, NASTRAN.
  • Cohesive-zone modelling (CZM) — interface elements with bilinear or exponential traction-separation laws (Dugdale 1960; Barenblatt 1962; Camanho & Dávila 2003). Governs initiation and propagation without requiring an initial flaw.
  • Standards — ASTM D5528 (Mode I, DCB), ASTM D7905 (Mode II, ENF), ASTM D6671 (mixed-mode bending). Material parameter G_IC typical 100–500 J/m² for thermoset CFRP, 1000–2500 J/m² for thermoplastic (PEEK, PEKK) CFRP.

Additively-manufactured metals. Process-induced defects (lack-of-fusion, gas porosity, keyholing) act as ready-made fatigue crack initiators. The Murakami-area approach treats defects as equivalent surface cracks:

ΔK_th = ΔK_th,smooth · ((area)^(1/2) / a₀)^(1/6)        Murakami 1989

with √area the defect-projected area on the plane normal to σ_max. AM-relevant standards: ASTM F3122 (DTD methodology for AM), ASTM E3122 (NDE for AM), NASA-STD-6030/6033 (spaceflight AM), ASME BPVC-VIII Code Case 2901 (AM in pressure equipment). Anisotropic toughness — build direction has lower K_IC than transverse — is universal in LPBF Ti-6Al-4V, Inconel 718, AlSi10Mg, 17-4 PH.

13. Edge cases & gotchas

  1. Constraint loss in thin sections. A K_IC measured on a thick CT specimen does not apply to a 2-mm-thick skin. Apparent K_C is higher; correlate via K_C(B) curves (Anderson 2017 §2.10) or do an EPFM analysis directly.

  2. Mixed-mode loading. When Mode II or III are present, use a combined criterion. Maximum tangential stress (Erdogan & Sih 1963):

    K_I·sin θ + K_II·(3 cos θ − 1) = 0       (gives kink angle θ)
K_eq = K_I·cos³(θ/2) − 3·K_II·cos²(θ/2)·sin(θ/2)

    Alternatives: maximum energy-release-rate (Hussain et al. 1974), minimum strain-energy-density (Sih 1973).

  3. Subcritical growth far below K_IC. Stress-corrosion cracking, hydrogen embrittlement, and creep crack growth move cracks at K ≪ K_IC. K_ISCC — the stress-corrosion threshold — can be 10–30 % of K_IC for high-strength steels in chloride or H₂S environments. NACE MR0175 / ISO 15156 limits steel hardness in sour service explicitly to keep K_ISCC up.

  4. Residual stresses superpose linearly. Welding residual stress σ_res ≈ σ_y locally near the weld toe. Effective K = K_applied + K_residual. Shot-peening, autofrettage, and laser shock peening exploit this in reverse — introduce compressive σ_res to reduce K and extend life. BS 7910 Annex Q has parametric residual-stress profiles for welds.

  5. Multi-site damage (MSD). Many small cracks at near-uniform-stressed sites (lap-joint rivet rows, weld toes, turbine-disc rim) interact when their plastic zones overlap or when they link up. Single-crack damage-tolerance methods become non-conservative. Capture with FRANC3D explicit multi-crack growth or BS 7910 Annex G.

  6. Low-temperature transition in ferritic steel. BCC iron and ferritic steels exhibit a ductile-to-brittle transition. CVN (Charpy V-notch, ASTM E23) drops sharply from upper-shelf (~150–300 J) to lower-shelf (~5–20 J) over a 30 °C window. Service temperature must be above NDT (Nil-Ductility Transition); ABS / DNV plate grades (A, B, D, E, F) progressively guarantee lower CVN-test temperatures.

  7. Embrittlement modes to flag in failure analysis:

    • Temper embrittlement (Cr-Mo-Ni steels held at 350–550 °C): P, Sb, Sn segregation to prior austenite grain boundaries; intergranular fracture.
    • Hydrogen embrittlement (high-strength steels above ~1000 MPa σ_u): cathodic protection, electroplating, sour service, weld-metal hydrogen. Drops K_IC by 50–80 %.
    • Radiation embrittlement (nuclear pressure vessels): neutron flux raises DBTT by 50–150 °C over plant life; monitored with surveillance capsules per ASTM E185.
    • Sigma-phase embrittlement (stainless 300-series held 600–900 °C): brittle FeCr σ precipitates.

  8. Surface-flaw vs through-thickness. Newman-Raju (1981) gives Y(a/c, a/t, φ) for semi-elliptical surface cracks; depth a and length 2c grow at different rates because Y varies along the crack front. Aspect ratio drifts toward a/c ≈ 0.7–0.8 in steady-state fatigue.

  9. Crack path is not always normal to σ_max. Anisotropy, residual stress, and mixed-mode loading cause turning. Predict with FE re-meshing (FRANC3D, ABAQUS XFEM with arbitrary crack growth, ANSYS SMART crack growth).

  10. High R-ratio effects. At R > 0.5, closure does not develop fully and ΔK_eff → ΔK; threshold drops. Forman R-correction or NASGRO closure function required for life predictions on rotating components with R ≈ 0.7–0.9 (turbine discs, helicopter rotor heads).

14. Creep crack growth — practical engineering form

For service above ~0.4 T_melt (carbon steel above ~400 °C, austenitic stainless above ~500 °C, Cr-Mo-V rotors above ~480 °C), time-dependent crack growth under steady load becomes the dominant subcritical mechanism. The relevant driving force is the C* integral (Landes & Begley 1976), the steady-state-creep analogue of J:

C* = ∮_Γ (W*·dy − T_i·(∂u̇_i/∂x) ds)        W* = strain-rate density

Power-law fit:

da/dt = D · (C*)^φ                            φ ≈ 0.8–0.9 typical

In the small-scale-creep regime where elastic strain dominates and creep is confined to the crack tip, the C_t parameter (Saxena 1986) replaces C* with a transient correction. Reference-stress methods (Webster & Ainsworth 1994) give engineering estimates without solving the full visco-plastic field:

C* = σ_ref · ε̇_ref · K² / σ_ref²              from Norton creep law ε̇ = A·σ^n

Standards: ASTM E1457-19 (creep crack growth testing using compact-type specimens), API 579-1 Part 10 (creep FFS), R5 (UK CEGB / EDF Energy procedure for high-temperature plant). Typical service applications: superheater outlet headers, main steam piping, gas-turbine combustor liners, nuclear reactor pressure vessel above ratcheting limits.

15. Dynamic fracture and crack arrest

For impact, blast, projectile, and pressure-burst loads, fracture-mechanics parameters become rate-dependent:

  • K_Id (dynamic initiation toughness) — measured under instrumented Charpy or pre-cracked drop-weight tests; typically K_Id < K_IC for ferritic steel (rate-sensitivity of σ_y inverts the toughness ranking in the transition region).
  • K_Ia (crack-arrest toughness) — the K below which a running crack arrests. Drop-weight tear test (NDT, ASTM E208) and crack-arrest test (ASTM E1221) give K_Ia. Reactor pressure vessel codes (ASME III) require K_Ia margin above operating K_I for all postulated transients.
  • Dynamic crack-tip stress field depends on crack velocity v_c relative to Rayleigh wave speed C_R; K → 0 as v_c → C_R. Practical crack velocities in steel: 500–1500 m/s.
  • Stress-wave loading — a square pressure pulse arriving at a crack tip drives K(t) that rings up at speed C_L; the design parameter is the peak dynamic K, not the static-equivalent K.

16. Tools, standards, software

16.1 Standards (current revisions)

StandardTitle / scope
ASTM E399-22K_IC plane-strain fracture toughness (CT, SENB, arc-bend, disc)
ASTM E1820-22J-integral, J_IC, J-R curve, CTOD (unified standard)
ASTM E647-23Fatigue crack growth rate da/dN vs ΔK
ASTM E1290-08(2014)CTOD δ measurement (largely superseded by E1820)
ASTM E1921-22Master Curve for ferritic-steel reference temperature T_0
ASTM E23-23aCharpy V-notch impact test
ISO 12135:2021Unified fracture-toughness test (international counterpart to E1820)
ISO 12108:2018Metallic materials — Fatigue crack growth rate
BS 7910:2019Fitness-for-service of metallic structures (UK, widely adopted)
API 579-1 / ASME FFS-1 (2021)Fitness-for-service for pressure equipment (refining, petrochem)
ASME BPVC Section XI (2023)In-service inspection of nuclear power-plant components
FAR 25.571 / EASA CS-25.571Damage tolerance and fatigue evaluation of structure (large aircraft)
MIL-STD-1530DUSAF Aircraft Structural Integrity Program (ASIP)
MIL-A-83444 (historical)Airplane damage-tolerance requirements (now in MIL-STD-1530)
MMPDS-19 (2024)Metallic Materials Properties Development & Standardization
API 580 / 581Risk-Based Inspection (RBI) — uses fracture-mechanics input

16.2 Software

  • NASGRO (NASA + Southwest Research Institute) — fatigue crack growth, damage tolerance, SIF library; NASGRO equation reference implementation. Aerospace standard.
  • AFGROW (USAF / LexTech) — fatigue crack growth, multi-load-spectrum, SIF library. USAF damage-tolerance standard.
  • Abaqus — contour-integral J extraction, XFEM (extended FEM) for arbitrary crack growth without remeshing, cohesive-zone modeling, virtual crack closure technique (VCCT) for composites.
  • ANSYS Mechanical — SMART (Separating Morphing and Adaptive Remeshing Technology) crack growth, contour-integral J, fracture-toolbox SIF extraction.
  • FRANC3D (Cornell Fracture Group / FAC) — purpose-built 3-D crack growth with adaptive remeshing; widely used in turbine-engine community.
  • BEASY — boundary-element fracture analysis (commercial, BEASY Ltd).
  • CRACKWISE (TWI) — BS 7910 automation; FFS assessments.
  • Signal-Fitness for Service (DNV) — API 579 / BS 7910 automation.
  • CalculiX — open-source FEA with J-integral capability via Abaqus-compatible input.

16.3 Non-destructive evaluation (the input)

Fracture-mechanics calculations are only as good as the assumed initial flaw size. NDE methods and their typical resolution (covered in joining-welding §NDE):

  • UT phased-array (PAUT) — 1–3 mm flaw resolution, depth-encoded; standard for pressure vessels, pipelines, thick welds.
  • Time-of-flight diffraction (ToFD) — 0.5–1 mm; sizing accuracy ±1 mm typical.
  • Eddy current (ET) — surface and near-surface, 0.3–1 mm in conductive materials.
  • Magnetic particle (MT) — surface-breaking only, ferromagnetic substrate.
  • Liquid penetrant (PT) — surface-breaking, any non-porous material.
  • Radiography (RT) — film or digital; volumetric, ~2 % of thickness typical sensitivity.

Probability-of-detection (POD) curves (per MIL-HDBK-1823) feed the assumed initial flaw a_NDE used in damage-tolerance calculations: a_NDE ≈ a_{90/95}, the flaw size detected 90 % of the time at 95 % confidence.

17. Cross-references

  • mechanics-of-materials — prerequisite; stress, strain, principal stresses, plane-stress vs plane-strain definitions.
  • beam-theory — bending stresses feeding K through Mode I solutions; SIF expressions for cracked beams.
  • materials-steel — K_IC, CVN, DBTT for the steel families discussed here; heat-treatment effects on toughness.
  • materials-aluminum — K_IC and Paris constants for 2024, 7075, 6061 series.
  • materials-selection — toughness-vs-strength Ashby map; K_IC / σ_y as the leak-before-break index.
  • joining-welding — weld defects, HAZ toughness, residual-stress fields, NDE methods.
  • vibration-dynamics — cyclic-load spectra feeding fatigue-life calculations.
  • fatigue-analysis (planned, companion in this batch) — high-cycle / low-cycle fatigue, S-N curves, strain-life, cumulative damage; consumes the Paris-law output.
  • fem-fea (planned, companion) — J-integral extraction, XFEM, SMART, VCCT implementations.
  • structural-dynamics (planned) — dynamic-K, stress-wave-loaded cracks, impact fracture.
  • manipulator-design — robot joints and link fatigue; the Paris-law output sets inspection intervals on cyclic-load-bearing welded structure.
  • scientific — .inp (Abaqus), .cdb (ANSYS), .bdf (NASTRAN) for fracture models.

18. Citations

  1. Anderson, T. L. Fracture Mechanics: Fundamentals and Applications, 4th ed. CRC Press, 2017. ISBN 978-1498728133. The canonical English-language graduate text.
  2. Hertzberg, R. W.; Vinci, R. P.; Hertzberg, J. L. Deformation and Fracture Mechanics of Engineering Materials, 5th ed. Wiley, 2012. ISBN 978-0470527801. Strong on materials-science side, fatigue mechanisms.
  3. Broek, D. Elementary Engineering Fracture Mechanics, 4th ed. Kluwer, 1986. ISBN 978-9024726561. Concise practical introduction, still in use as design-engineer’s primer.
  4. Tada, H.; Paris, P. C.; Irwin, G. R. The Stress Analysis of Cracks Handbook, 3rd ed. ASME Press, 2000. ISBN 978-0791801536. The SIF reference; ~600 geometries.
  5. Suresh, S. Fatigue of Materials, 2nd ed. Cambridge, 1998. ISBN 978-0521578479. Canonical fatigue text — closure, short cracks, mechanisms.
  6. Murakami, Y. Stress Intensity Factors Handbook, 5 vols. Pergamon / Elsevier, 1987–2003. Complementary SIF compendium to Tada/Paris/Irwin.
  7. Griffith, A. A. “The phenomena of rupture and flow in solids.” Phil. Trans. Roy. Soc. London A 221 (1921) 163–198. The origin of fracture mechanics.
  8. Irwin, G. R. “Analysis of stresses and strains near the end of a crack traversing a plate.” J. Appl. Mech. 24 (1957) 361–364. K-factor.
  9. Rice, J. R. “A path-independent integral and the approximate analysis of strain concentration by notches and cracks.” J. Appl. Mech. 35 (1968) 379–386. The J-integral.
  10. Wells, A. A. “Unstable crack propagation in metals: cleavage and fast fracture.” Proc. Crack Propagation Symposium, Cranfield (1961). CTOD.
  11. Paris, P. C.; Erdogan, F. “A critical analysis of crack propagation laws.” J. Basic Eng. 85 (1963) 528–534. Paris law.
  12. Elber, W. “Fatigue crack closure under cyclic tension.” Eng. Fract. Mech. 2 (1970) 37–45. Plasticity-induced closure.
  13. Newman, J. C.; Raju, I. S. “An empirical stress-intensity-factor equation for the surface crack.” Eng. Fract. Mech. 15 (1981) 185–192. Semi-elliptical surface flaw solution.
  14. Erdogan, F.; Sih, G. C. “On the crack extension in plates under plane loading and transverse shear.” J. Basic Eng. 85 (1963) 519–525. Mixed-mode criterion.
  15. Hutchinson, J. W.; Rice, J. R.; Rosengren, G. F. (separate 1968 papers, J. Mech. Phys. Solids 16). HRR singular field.
  16. Bowie, O. L. “Analysis of an infinite plate containing radial cracks originating at the boundary of an internal circular hole.” J. Math. Phys. 35 (1956) 60–71.
  17. Kitagawa, H.; Takahashi, S. “Applicability of fracture mechanics to very small cracks.” 2nd ICM Conf. (1976) 627–631. Short-crack threshold.
  18. ASTM E399-22 — Standard Test Method for Linear-Elastic Plane-Strain Fracture Toughness K_IC of Metallic Materials. ASTM International, 2022.
  19. ASTM E1820-22 — Standard Test Method for Measurement of Fracture Toughness. ASTM International, 2022.
  20. ASTM E647-23 — Standard Test Method for Measurement of Fatigue Crack Growth Rates. ASTM International, 2023.
  21. ASTM E1290-08(2014) — Standard Test Method for Crack-Tip Opening Displacement (CTOD) Fracture Toughness Measurement. ASTM International, 2014.
  22. ASTM E1921-22 — Standard Test Method for Determination of Reference Temperature, T_0, for Ferritic Steels in the Transition Range. ASTM International, 2022.
  23. BS 7910:2019 — Guide to methods for assessing the acceptability of flaws in metallic structures. BSI, 2019.
  24. API 579-1 / ASME FFS-1 (2021) — Fitness-For-Service. API / ASME, 2021. The pressure-equipment FFS code.
  25. MMPDS-19 (2024) — Metallic Materials Properties Development and Standardization. Battelle, 2024. Successor to MIL-HDBK-5J; aerospace allowables.
  26. NASGRO Reference Manual v10. NASA Johnson Space Center / Southwest Research Institute, 2023. Algorithm and equation documentation.
  27. MIL-HDBK-1823A — Nondestructive Evaluation System Reliability Assessment. DoD, 2009. POD methodology underpinning damage tolerance.
  28. Wallin, K. “The scatter in K_IC results.” Eng. Fract. Mech. 19 (1984) 1085–1093. Master-curve foundation.
  29. O’Dowd, N. P.; Shih, C. F. “Family of crack-tip fields characterized by a triaxiality parameter — I. Structure of fields.” J. Mech. Phys. Solids 39 (1991) 989–1015. Q-parameter constraint quantification.
  30. Forman, R. G.; Mettu, S. R. “Behavior of surface and corner cracks subjected to tensile and bending loads in Ti-6Al-4V alloy.” Fracture Mechanics: 22nd Symposium, ASTM STP 1131 (1992). NASGRO baseline.
  31. Murakami, Y.; Endo, M. “Quantitative evaluation of fatigue strength of metals containing various small defects or cracks.” Eng. Fract. Mech. 17 (1989) 1–15. √area approach for defects in AM and cast metals.
  32. Rybicki, E. F.; Kanninen, M. F. “A finite element calculation of stress intensity factors by a modified crack closure integral.” Eng. Fract. Mech. 9 (1977) 931–938. VCCT method.
  33. Camanho, P. P.; Dávila, C. G. “Mixed-mode decohesion finite elements for the simulation of delamination in composite materials.” NASA/TM-2002-211737, 2002. CZM implementation.
  34. Pugsley, A. G. The Safety of Structures. Edward Arnold, 1966. Fail-safe philosophy.
  35. ASTM E1221-12a(2018) — Standard Test Method for Determining Plane-Strain Crack-Arrest Fracture Toughness K_Ia. ASTM, 2018.
  36. ASTM E208-22 — Standard Test Method for Conducting Drop-Weight Test to Determine NDT Temperature of Ferritic Steels. ASTM, 2022.
  37. ASTM D5528-22 / D7905-19 / D6671-22 — Mode I, Mode II, and mixed-mode interlaminar fracture toughness of laminated composites. ASTM, current revisions.
  38. NACE MR0175 / ISO 15156:2020 — Materials for use in H₂S-containing environments in oil and gas production. NACE / ISO, 2020. Sets K_ISCC-protective hardness caps for sour service.
  39. ASME BPVC Section XI (2023) — Rules for In-service Inspection of Nuclear Power Plant Components. ASME, 2023. Codifies LBB and probabilistic fracture-mechanics-informed inspection.
  40. EN 1993-1-9:2005 — Eurocode 3: Design of steel structures — Fatigue. CEN, 2005. European fatigue and fracture detail categories for steel structures.

Graph View