Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
ASME STP-PT-074-2015 — Local Stress in Nozzles, Shells and Formed Heads from External Loads
📐 ASME STP-PT-074-2015 is a technical report sponsored by the ASME Pressure Technology Codes & Standards committee and ASME ST-LLC and prepared by the Paulin Research Group (PRG). Issued in June 2015, it computes local stresses in nozzles, shells, and formed heads under external loads across a broad parameter space — diameter ratio D/T from 7 to 2500, nozzle-to-shell diameter ratio d/D from 0 to 0.7, and wall ratio t/T from 0.1 to 10 — and provides recommended rules and supporting tables aligned with the design-by-analysis (DBA) provisions of ASME Boiler and Pressure Vessel Code Section VIII, Division 2.
💡 This report targets the extrapolation gap of influence-coefficient methods such as WRC 107 / WRC 297: those classic methods are rooted in thin-shell and moderate-thickness assumptions. Here the applicability is extended to the very thin (D/T = 2500) and thick (D/T = 7) ends, with explicit stress-classification-line (SCL) assessment procedures.
1. Background and Parameter Space 📌
Nozzles on pressure-vessel shells and formed heads carry not only internal pressure but also external loads transmitted from the attached piping: thermal-expansion thrust, self-weight, safety-valve reaction, wind, and seismic loads. These loads produce significant local membrane and bending stresses at the nozzle-to-shell junction — a critical location for fatigue and strength assessment. The classical WRC 107 / WRC 297 influence-coefficient methods have limited parameter coverage and insufficient data for formed heads and for extremely thin or thick shells.
This report establishes a unified 3D finite element analysis framework covering the following parameter space:
| Geometric Parameter | Range Covered | Meaning | Extension vs. Traditional Methods |
|---|---|---|---|
| D/T (shell diameter/wall ratio) | 7 – 2500 | Thick to extremely thin shells | Often limited to D/T≈10–100 traditionally |
| d/D (nozzle/shell diameter ratio) | 0 – 0.7 | Very small to large openings | Covers both extremes |
| t/T (nozzle/shell wall ratio) | 0.1 – 10 | Relative nozzle strength | Covers thin and thick nozzles |
| Component type | Shells, elliptical/torispherical formed heads | Structural class | Separate data for formed heads |
2. Methodology and Correlation Validation 🔧
The report uses 3D finite element analysis (both shell elements and brick/solid elements) as its backbone:
- Shell models: large-area thin-shell models solve the global stress distribution and support the parameter sweep.
- Brick comparison: Widera geometry no. 26 is re-modeled with solid elements to cross-validate the shell models, verifying the SCL locations and stress extraction.
- Correlation accuracy: FEA results are compared one by one with existing correlations (Wais, Widera, etc.), documenting deviations (e.g., up to ~35% for the Wais correlation and ~48.9% for the Widera correlation), which justifies the updated equations.
- Nonlinear behavior: the modification of local stresses by material nonlinearity and large deformation is evaluated.
- Points of inflection, maxima and minima: stress-factor trends versus t/T and D/T are systematically examined so the recommended rules remain continuous and reasonable at parameter boundaries.
⚠️ Deviations in existing correlations are not negligible: the report shows differences of roughly 35%–49% versus high-fidelity FEA for some geometric combinations. Outside the traditional validity range, blindly using old curves can be unconservative.
3. Recommended Rules and Tables 📊
Appendix 2 of the report provides the “recommended rules and supporting tables,” and Appendix 3 compiles all the finite element results. The recommended procedure assesses external-load local stress as follows:
- Select the stress-factor table by the shell/nozzle geometry (D/T, d/D, t/T) and component type (shell or formed head);
- Decompose the external loads (axial force, in-plane/out-of-plane moments, torsion, radial force, etc.) into membrane and bending components;
- Evaluate against the Section VIII Division 2 stress-classification limits (primary local membrane PL, primary bending Pb, secondary stress Q).
The report also compares the membrane/bending factors for circumferential branch moment between WRC 107 and WRC 297, explaining the sources of divergence in some parameter regions (e.g., the t/T ceiling effect in WRC 297 Figures 38 and 46), helping users identify valid ranges when citing old bulletins.
4. Engineering Practice Essentials 🛠️
- 🔍 Identify external load sources: inventory the actual pipe reactions on each nozzle (thermal displacement, dead weight, occasional loads) before selecting single or combined load cases; do not check pressure only.
- 📊 Stress-classification assessment: per Section VIII-2 Part 5, use the SCL method to separate primary and secondary stress. Bending from external loads often combines primary bending and secondary stress and must be checked against different limits.
- 🔧 Head-nozzle special case: nozzles on formed heads (elliptical, torispherical) behave differently from those on cylindrical shells; use the report’s dedicated head data rather than shell curves.
- 📋 Outside the parameter space: when the actual geometry falls outside the report’s space (e.g., t/T > 10), perform direct finite element analysis with a documented assessment report rather than extrapolating.
🚨 Safety-valve and relief-line reaction is one of the most overlooked external loads. Reaction forces often reach tens to hundreds of kilonewtons and are impulsive; they must be included in nozzle local-stress assessment or root fatigue cracking may result.
5. Frequently Asked Questions ❓
❓ Q: Why can’t WRC 107 / WRC 297 be used everywhere?
A: Those bulletins rely on thin-shell assumptions and limited test data; they under-cover the very thin (D/T>100), very thick (D/T<10), large-nozzle (d/D>0.5), and formed-head regimes, with deviations reaching 35%–49%. This report builds a new database that closes those gaps.
A: Those bulletins rely on thin-shell assumptions and limited test data; they under-cover the very thin (D/T>100), very thick (D/T<10), large-nozzle (d/D>0.5), and formed-head regimes, with deviations reaching 35%–49%. This report builds a new database that closes those gaps.
❓ Q: Why do external-load local stresses matter so much?
A: The junction is the most severe stress-concentration region. External loads superimpose on pressure stress, easily pushing local stress beyond yield and creating cyclic fatigue damage. Most nozzle-root cracking and fatigue problems trace to underestimated external loads.
A: The junction is the most severe stress-concentration region. External loads superimpose on pressure stress, easily pushing local stress beyond yield and creating cyclic fatigue damage. Most nozzle-root cracking and fatigue problems trace to underestimated external loads.
❓ Q: How does a formed-head nozzle differ from a shell nozzle?
A: The curvature, stress state, and boundary effects of formed heads differ markedly from cylindrical shells; the stress factors differ even at identical d/D and t/T. The report provides separate head datasets and curves — match the component type.
A: The curvature, stress state, and boundary effects of formed heads differ markedly from cylindrical shells; the stress factors differ even at identical d/D and t/T. The report provides separate head datasets and curves — match the component type.
🔍 Q: When should influence-coefficient methods be abandoned for direct FEA?
A: When the geometry falls outside the parameter range, loads are complex, combined, or impulsive, the head is special, or an accurate fatigue life is needed — use direct 3D FEA with a mesh-convergence study.
A: When the geometry falls outside the parameter range, loads are complex, combined, or impulsive, the head is special, or an accurate fatigue life is needed — use direct 3D FEA with a mesh-convergence study.
