🧮 Go straight to the calculator: want to skip the derivation? Open the BS 7910 fracture toughness estimator , load a standard example, and verify in one click.
1. Why estimate K_mat from CVN?
In a fitness-for-service / engineering critical assessment (FFS/ECA), fracture toughness $K_{mat}$ is the core parameter for judging whether a crack will cause brittle fracture. Yet in many real engineering situations:
- the component has been in service for years and no fresh toughness specimen can be taken;
- the code required only a Charpy impact test, so the historical data hold CVN values only;
- a full-size fracture toughness test is extremely costly and not feasible in practice.
Here, indirect conversion is the only viable route. BS 7910:2019 Annex J provides two well-validated conversion methods for exactly this purpose, for ferritic steels.
⚠️ This method applies to ferritic steels only (yield strength $\sigma_{ys} \leq 690$ MPa); it does not apply to austenitic stainless steels (which have no ductile-brittle transition).
2. Overview of the two conversion routes
| Route | Code clause | Applicability | Character |
|---|---|---|---|
| A: lower-shelf equation | BS 7910:2019, Annex J.2.1 | 3 J < $C_v$ ≤ 27 J, fracture-surface crystallinity ≥ 80% | Conservative, for the very-low-temperature brittle region |
| B: Master Curve | BS 7910:2019, Annex J.2.2 | Ductile-brittle transition region | Statistical-physical model, more accurate |
3. Route A: the lower-shelf equation (Annex J.2.1)
3.1 Conditions of use
The lower-shelf equation requires the material to be “near the lower shelf”, shown by:
- a Charpy impact energy $C_v$ between 3 J and 27 J;
- a cleavage (brittle) area fraction (crystallinity) on the fracture surface of ≥ 80%.
When $C_v > 27$ J the material has entered the transition region, where the J.2.1 equation overestimates toughness; the Master Curve method must be used instead.
3.2 The equation
$$K_{mat} = \left[(12\sqrt{C_v} - 20) \cdot \left(\frac{25}{B}\right)^{0.25}\right] + 20 \quad [\text{MPa}\sqrt{\text{m}}]$$(Source: BS 7910:2019, Annex J.2.1, Eq. J.1)
Physical meaning of each term:
- The constant 20: the theoretical minimum fracture toughness of ferritic steel, $K_{min}$ = 20 MPa√m (the asymptotic lower bound, from ASTM E1921).
- The $12\sqrt{C_v}$ term: an empirical lower-envelope coefficient fitted to a large body of ferritic-steel test data.
- $(25/B)^{0.25}$: a size correction from weakest-link theory — a thicker section contains more micro-crack initiation sites, so statistically $K_{mat}$ is lower; the exponent 0.25 = 1/β, with β = 4 the Weibull shape parameter.
3.3 Worked example 2
Input: $C_v = 20$ J, $B = 50$ mm
$$K_{mat} = [(12 \times \sqrt{20} - 20) \times (25/50)^{0.25}] + 20 = [33.66 \times 0.8409] + 20 = \mathbf{48.3 \text{ MPa}\sqrt{\text{m}}}$$4. Route B: the Master Curve method (Annex J.2.2)
The Master Curve method, proposed by Kim Wallin (1984) from a large body of ferritic-steel fracture toughness data, has one core idea: the fracture toughness of ferritic steel in the transition region follows a three-parameter Weibull distribution, and the distribution shape is material-independent (only the location parameter $T_0$ varies by material).
4.1 Step one: determine the reference transition temperature $T_0$
Way 1 (preferred): if measured ASTM E1921 Master Curve data exist, use the measured $T_0$ directly — no statistical margin is then needed.
Way 2: estimate from $T_{27J}$ (this calculator’s default route):
$$T_0 = T_{27J} - 18 \quad [°\text{C}]$$(Source: BS 7910:2019, Annex J.2.2)
$T_{27J}$ is the temperature on the CVN transition curve at $C_v = 27$ J, and the empirical offset of −18°C gives $T_0$.
4.2 Step two: the T_K confidence margin (⚠️ a key pitfall)
When $T_0$ is estimated from $T_{27J}$, BS 7910 requires an added $T_K = 25°\text{C}$ confidence margin to cover the statistical scatter of the CVN → $T_0$ conversion (90% confidence):
$$T_{eff} = T - T_0 - T_K = T - T_0 - 25°\text{C}$$This is the most common source of calculation error! Omitting $T_K$ overestimates $K_{mat}$. In standard example 1, omitting $T_K$ inflates $K_{mat}$ from 90.8 MPa√m to about 128 MPa√m.
4.3 Step three: the Master Curve calculation
Weibull scale parameter $K_0$ (the 63.2% failure fractile, for a 1T standard specimen):
$$K_0 = 31 + 77 \cdot \exp[0.019 \cdot T_{eff}] \quad [\text{MPa}\sqrt{\text{m}}]$$(Source: BS 7910:2019, Annex L.9.5.3, Eq. L.13; 31 = $K_{min}$ + 11 = 20 + 11)
$K_{mat}$ at a specified failure probability (1T basis):
$$K_{mat,1T} = 20 + \left[\ln\left(\frac{1}{1-P_f}\right)\right]^{0.25} \cdot (K_0 - 20)$$(Source: BS 7910:2019, Annex J.2.2, Eq. J.5; the ASTM E1921 three-parameter Weibull inverse)
At $P_f = 0.05$: the probability term = $[\ln(1/0.95)]^{0.25} = (0.05129)^{0.25} = 0.4759$
Thickness correction (weakest-link size effect):
$$K_{mat} = 20 + (K_{mat,1T} - 20) \cdot \left(\frac{25}{B}\right)^{0.25}$$(Source: BS 7910:2019, Annex J.2.2; ASTM E1921)
4.4 Worked example 1 (standard example)
Input: $T_{27J} = -50°\text{C}$, $T = 0°\text{C}$, $B = 60$ mm, $P_f = 0.05$
| Step | Calculation | Result |
|---|---|---|
| B1: T_0 = T_27J - 18 | = -50 - 18 | -68°C |
| T_eff (with T_K=25°C) | 0 - (-68) - 25 | 43°C |
| K_0 = 31 + 77·exp(0.019×43) | = 31 + 174.3 | 205.3 MPa√m |
| K_mat_1T = 20 + 0.4759×(205.3-20) | = 20 + 88.2 | 108.2 MPa√m |
| K_mat = 20 + 88.2×(25/60)^0.25 | = 20 + 70.9 | 90.9 MPa√m ✅ |
Code reference value: 90.8 MPa√m (BS 7910:2019, Annex J.2.5 Note 2)
5. The API 579 comparison route
API 579-1:2021, 9F.4.3.2, Eq. (9F.74) gives a finer, multi-parameter estimate of $T_0$:
$$T_0 = T_{28J} - 79 + \frac{\sigma_{ys}}{9} - \frac{C_{V-US}}{59} + \Delta T_0 \quad [°\text{C}]$$where $\Delta T_0 = 18°\text{C}$ (a fixed conservative margin) and $T_{28J} \approx T_{27J}$ (27 J ≈ 20 ft-lbs).
Compared with BS 7910’s $T_0 = T_{27J} - 18°\text{C}$, API 579 additionally accounts for:
- the yield-strength term $+\sigma_{ys}/9$: higher strength → more brittle tendency → $T_0$ shifts to higher temperature;
- the upper-shelf energy term $-C_{V-US}/59$: better ductility → $T_0$ shifts to lower temperature.
Note: API 579 treats its $T_0$ as a “directly determined value”, so $T_K = 0$ when it goes into the Master Curve (the $\Delta T_0 = 18°\text{C}$ already plays the same role).
Example 3 comparison
| Parameter | BS 7910 | API 579 |
|---|---|---|
| T_0 | -18°C | -24.1°C |
| K_mat | 44.7 MPa√m | 61.0 MPa√m |
Because API 579 additionally accounts for yield strength and upper-shelf energy, its $T_0$ is lower and its $K_{mat}$ higher (more optimistic).
6. Engineering-use advice
- Order of preference: measured $T_0$ (ASTM E1921) > API 579 multi-parameter conversion > BS 7910 single-parameter $T_{27J}$ estimate.
- Always check that $T_K = 25°C$ has been included (specific to the CVN route).
- Lower-shelf boundary: for $C_v > 27$ J, J.2.1 must not be used — switch to the Master Curve.
- Thickness effect: when the wall thickness $B > 25$ mm, the size correction lowers $K_{mat}$ markedly and cannot be ignored.
7. Quick start with the calculator
- Open the fracture mechanics calculators → choose “BS 7910 Fracture Toughness Estimation”.
- Click the “Example 1 (Master Curve, BS 7910 standard example)” button to auto-fill the parameters.
- Click “Calculate” to verify: $K_{mat} = 90.85$ MPa√m ✅.
- The right panel shows the three Master Curve probability curves ($P_f$ = 5%/50%/95%) and the position of the current assessment point.
📖 References:
- BS 7910:2019+A1:2020, Guide to methods for assessing the acceptability of flaws in metallic structures
- API 579-1/ASME FFS-1:2021, Part 9 Annex 9F — Fracture Toughness
- ASTM E1921, Standard Test Method for Determination of Reference Temperature T_0
- Wallin, K. (1984). The scatter in K_Ic results. Engineering Fracture Mechanics, 19(6), 1085-1093.