Fatigue Screening Criteria Under ASME VIII Div.2 Explained: A Worked Example

A full fatigue analysis — elastic or elastic-plastic stress evaluation across every load combination in a vessel's service life — is one of the more expensive parts of an ASME Section VIII Division 2 design. Paragraph 5.5.2 exists to let you skip it entirely when the cyclic service is genuinely mild enough not to need it. This guide walks through Method A — the most commonly used screening path — with a fully worked reactor vessel example, using the same logic behind this site's free Fatigue Screening Calculator.

Screening Overview: Three Ways to Skip a Fatigue Analysis

Per 5.5.2.1, if any one of three screening options is satisfied for a given component, a detailed fatigue analysis is not required for that component:

• 5.5.2.2 (unnamed, often called "Method C" informally): documented successful experience with comparable equipment under similar loading — a qualitative, engineering-judgment exemption with no numeric criterion in the Code.
• 5.5.2.3, Method A: cycle counting against Table 5.9's allowable limits — but only for materials with specified minimum tensile strength ≤ 552 MPa (80,000 psi).
• 5.5.2.4, Method B: a stress-amplitude / cumulative-damage check using the design fatigue curve — applies to any material, with no tensile strength limit.

Critically, 5.5.2.1(c) caps all three options: if the specified number of cycles exceeds 10⁶, none of the screening criteria apply and a fatigue analysis is mandatory regardless of the other checks. Screening is also done per-component — one part of a vessel can be exempt while another, non-integral attachment on the same vessel is not.

Method A: Cycle Counting Against Table 5.9

Method A's entire logic is comparing a sum of classified cycle counts against a fixed allowable limit. That limit depends on both the construction type (integral vs. non-integral) and, importantly, where on the vessel the component sits:

"Non-integral" construction — pad-type reinforcements, fillet-welded attachments, threaded connections, stud-bolted attachments — carries a much stricter limit than integral (fully-penetration-welded) construction, and a formed head's knuckle region carries a much stricter limit again, in both cases, because those geometries concentrate stress more severely for the same nominal loading.

Classifying a Cycle: Full-Range, Operating, or Thermal

Every process condition in the load history gets classified into one (or more) of these categories:

• N_∆FP — Full-range pressure cycles: cycles from atmospheric pressure to design pressure and back — startup, shutdown, emergency shutdown, pressure testing, and any user-designated full-range cycle.
• N_∆PO — Operating pressure cycles: pressure variations during normal operation that exceed 20% of design pressure for integral construction, or 15% for non-integral, but don't reach the full-range threshold.
• N_∆TE — Effective thermal cycles: changes in metal temperature difference between adjacent points, weighted by Table 5.8's temperature factor (below) before being added to the total.
• N_∆Tα — Dissimilar-material thermal cycles: for welds joining materials with different thermal expansion coefficients, counted separately when a specific strain criterion is exceeded (not modeled in this site's calculator — flagged as a manual check).

Table 5.8: Effective Thermal Cycles

Not every thermal cycle counts equally toward the total — Table 5.8 weights each cycle by a factor based on the magnitude of the metal temperature difference, then that factor multiplies the number of cycles at that magnitude to get the "effective" thermal cycle count:

A 28°C swing contributes nothing to the fatigue total — it's small enough to be considered non-damaging under this screening method. A swing just over 250°C is weighted 20× — meaning even a handful of large thermal transients can dominate the total far more than dozens of small ones.

Worked Example: Reactor Vessel R-101

Consider a carbon steel (SA-516 Gr. 70, UTS = 485 MPa — well under the 552 MPa Method A limit) reactor vessel with integral (fully welded) construction, 25-year design life, design pressure 1.0 MPa.

Step 1 — Construction classification and allowable limit (Table 5.9): integral construction, "all other components" (not knuckle-region), design pressure P = 1.0 MPa:

Step 2 — Full-range pressure cycles (N_∆FP): every condition marked full-range sums directly (startup, shutdown, emergency shutdown, unplanned shutdowns, maintenance, pre-commissioning):

Step 3 — Operating pressure cycles (N_∆PO): the 80 normal-operation cycles have no significant pressure range specified (thermal-only), so they don't cross the 0.200 MPa operating threshold and contribute 0 to this category.

Step 4 — Effective thermal cycles (N_∆TE), Table 5.8: every condition has ΔT = 46°C, which falls in Table 5.8's 29–56°C band → F = 1. Each condition's cycles carry that factor:

Step 5 — Total and screening assessment:

344 ≤ 1,000 — screening passes with a 65.6% margin. No detailed fatigue analysis is required for this component under Method A. This exact five-step derivation — with every intermediate substitution — is what the Fatigue Screening Calculator now displays on-screen and includes in its printed report.

The Knuckle-Region Trap: A Marginal Case

Now suppose the exact same load history applies to a nozzle located in the knuckle region of the reactor's head — a very common real-world location for a manway or instrument nozzle. The cycle count doesn't change (still 344 total), but the allowable limit drops from 1,000 to 350 per Table 5.9's stricter knuckle-region row.

This still technically passes — but only with a 1.7% margin, compared to 65.6% for the same loading on a non-knuckle component. This is exactly the kind of result that should prompt a second look: a small increase in design life, one more planned shutdown per year, or a slightly larger thermal swing pushing a few cycles into Table 5.8's next factor band could flip this specific component to "screening failed" even though an otherwise-identical component elsewhere on the same vessel stays comfortably exempt. This is precisely why 5.5.2.1(b) requires screening to be evaluated per component, not once for the whole vessel.

Method B: Stress Amplitude and Cumulative Damage

Method B (5.5.2.4) removes Method A's 552 MPa tensile-strength restriction, at the cost of a more involved procedure: stress amplitudes for full-range pressure, operating pressure, and thermal cycles are each computed, converted to an allowable number of cycles via the material's design fatigue curve, and combined via Miner's rule:

The fatigue strength reduction factor K_fb and construction factor H_c come from Table 5.10 (and Table 5.11 for welded construction, by weld type, surface finish, and inspection quality level). This site's calculator implements the real Table 5.10/5.11 lookups, but approximates the design fatigue curve itself with a generic high-cycle S-N power law rather than the exact, material- and temperature-specific Annex 3-F curve — so treat the result as an illustrative estimate, not a substitute for the full Annex 3-F evaluation.

Method B Worked Example: The Same Reactor Vessel R-101

Using the identical load history from the Method A example above (SA-516 Gr. 70: S = 138 MPa, Sy = 260 MPa, α = 0.000012/°C, E = 200 GPa, P = 1.0 MPa, integral nonwelded construction, not knuckle-region) — this is exactly the calculator's default Method B data, so you can reproduce every number below with one click of "Load Sample Data."

Step 1 — Table 5.10/5.11 factors: integral, nonwelded construction gives K_fb directly (no weld table lookup needed), and the component isn't in a knuckle region:

Step 2 — Full-range pressure stress amplitude:

Since Se,FP = 207 MPa exceeds the 100 MPa infinite-life threshold of the approximated curve, the allowable cycles come out to N_FP ≈ 112,743 cycles. With N_∆FP = 132 cycles from the same load history as before:

Step 3 — Operating pressure stress amplitude: no operating-pressure cycles were specified in this load history (the same as Method A's N_∆PO = 0), so:

Step 4 — Thermal stress amplitude: using the maximum ΔT = 46°C from the load history:

Se,T = 55.2 MPa is below the 100 MPa threshold, so N_T = 1,000,000 (infinite-life region). With effective thermal cycles of 212 (same total as Method A's N_∆TE):

Step 5 — Cumulative damage:

Screening passes under Method B too, with an enormous margin (D = 0.0014, over 700× below the 1.0 limit) — consistent with Method A's comfortable 65.6% margin on the same load history. This kind of cross-check, where both screening paths agree, is a useful sanity check when a component sits near Method A's cycle limit but well clear of Method B's damage limit (or vice versa).

Method C: The Comparable-Equipment Exemption

The comparable-equipment exemption (5.5.2.2) is qualitative — there's no formula or numeric score. You document the comparable equipment's design, its loading history, its service years, and its inspection record, then apply engineering judgment about whether that experience genuinely supports exempting the new design, considering specific risk factors the Code calls out: non-integral construction, large threaded connections, stud-bolted attachments, partial penetration welds, major thickness changes, and knuckle-region attachments. This site's calculator reflects that by returning a documentation checklist, not a fabricated score — the exemption decision itself still requires engineering judgment.

Try It Yourself

Reproduce both worked examples above, complete with the full step-by-step MathJax derivation the calculator now displays on-screen and reproduces in its printed report:

1. Open the Fatigue Screening Calculator and select Method A.
2. Set tensile strength = 485 MPa, construction = Integral, design pressure = 1.0 MPa.
3. Click "Load Sample Data" to populate the seven process conditions above.
4. Click Calculate to see the full 5-step derivation, then switch Component Location to "Knuckle region" and recalculate to see the margin shrink from 65.6% to 1.7%.
5. Switch to Method B — it defaults to the exact same reactor vessel R-101 material properties and load history, so clicking "Load Sample Data" there reproduces the Method B worked example above.
6. Click "Print Report" on either method to get a formatted, printable version of the full step-by-step solution.