Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
ASME STP-NU-013-2008 — Improvement of ASME NH for Grade 91 Negligible Creep and Creep-Fatigue
📐 ASME STP-NU-013-2008 is a nuclear-industry technical report issued by ASME Standards Technology, LLC (ASME ST-LLC) with funding from the U.S. Department of Energy (DOE). Prepared by Bernard Riou of Areva NP Inc. and dated September 3, 2008, the 147-page report addresses two specific improvements to the rules of Subsection NH of ASME Section III (Class 1 components in elevated-temperature service) as applied to Grade 91 (modified 9Cr-1Mo) steel. The first concerns the criteria by which creep may be considered negligible; the second concerns the creep-fatigue evaluation procedure. The document is organized in three parts: Part 1 covers negligible creep conditions for Grade 91, Part 2 covers creep-fatigue, and Part 3 proposes a test program to validate both revisions.
💡 Subsection NH is the world’s only construction-code framework that systematically covers elevated-temperature nuclear components in the range of roughly 370 to 815 °C through time-dependent allowable stresses and time-fraction damage accumulation. Because Grade 91 is the material of choice for Generation IV steam generators and main steam piping, the adequacy of its creep and creep-fatigue rules directly controls wall thickness, service life, and safety margins.
1. Background: Why Grade 91 Demands Dedicated NH Revisions 🏭
Grade 91 (SA-213 T91 seamless tube, SA-335 P91 pipe, SA-387 Gr.91 plate) is a modified martensitic steel strengthened by MX carbonitrides and M23C6 precipitates. It offers far higher creep strength and thermal-fatigue resistance than Grade 22 (2.25Cr-1Mo) in the 550–620 °C range, which is why it was selected for the steam generators, intermediate heat exchangers, and main steam lines of sodium-cooled fast reactors, supercritical water reactors, and ultra-supercritical fossil plants. Subsection NH permits Grade 91 service up to 649 °C (1200 °F), and its design assessment rests on three pillars: time-dependent allowable stresses (St, Smt), strain and deformation limits, and fatigue/creep-fatigue damage accumulation rules.
The report argues that these rules were originally calibrated on 1970s–1980s test data for 304/316 stainless steels and 2.25Cr-1Mo steel, and that transplanting them to Grade 91 is risky for two reasons. First, if the negligible creep boundary is set too wide, designers will wrongly waive creep and creep-fatigue evaluations for components that actually operate deep in the creep regime; if set too narrow, many low-temperature locations that genuinely merit simplification are forced into complex inelastic analysis. Second, the creep-fatigue procedure — strain range determination, initial hold-time stress, stress relaxation description, and creep damage accumulation — may not be conservative for a material that softens cyclically and oxidizes readily. The report quantifies both concerns.
2. Negligible Creep Criteria: ASME, RCC-MR, and Japanese Rules Compared 📊
The report systematically compares how the three major technical systems define “negligible creep.” The ASME system relies primarily on temperature limits plus the time-dependent allowable-stress framework, letting designers waive creep evaluations outright when conditions are met. The French RCC-MR rules offer a creep-strain criterion with an alternative stress-relaxation criterion, both evaluated with the RCC-MR creep constitutive law. The Japanese fast-reactor design standard adopts a 0.2% creep-strain criterion with its own creep strain law. The report constructs “negligible creep curves” for Grade 91 using the RCC-MR, ORNL, and Japanese creep laws, and then superimposes the curves obtained from the different criteria on a stress-versus-temperature plot to compare conservatism.
| Criterion Source | Basis | Typical Form | Engineering Implication |
|---|---|---|---|
| ASME NH | Temperature limit + time-dependent allowable stresses | Creep evaluation waived if design temperature is below a material-specific limit and stress is below the time-independent allowable | Simple to apply, but not directly tied to creep strain; conservatism near the boundary is hard to quantify |
| RCC-MR (France) | Creep-strain / stress-relaxation criterion | Cumulative equivalent creep strain over design life below a stated limit (~1%), or relaxed stress below a threshold | Requires creep constitutive data and full stress history; physically meaningful but computation-heavy |
| Japanese standard | 0.2% creep-strain criterion | Stress–temperature boundary corresponding to 0.2% inelastic strain | Strict; yields the narrowest negligible-creep region for Grade 91 |
| Report proposal | Reference-stress method based on stress-to-rupture | Unified curve built on the ratio of reference stress to rupture strength | Connects naturally to the St framework; suitable basis for an NH revision |
The comparison shows that the curve built on the ASME 0.2% creep-strain criterion is the most conservative. The criteria diverge rapidly in the mid-temperature range (roughly above 450 °C), where the equivalent-temperature differences can reach tens of degrees Celsius, while they converge at the low-temperature end. The report therefore concludes that the negligible-creep boundary for Grade 91 must not be copied from austenitic stainless steel values but must be re-derived from Grade 91’s own mid-temperature creep data.
⚠️ The negligible-creep criterion is an admission ticket to simplified design, not a proof of safety. Once the design temperature or stress crosses the boundary, the creep strain, creep ratcheting, and creep-fatigue evaluations must all be executed; extrapolating low-temperature experience into the elevated-temperature range is the most common compliance error in high-temperature component design.
3. Re-Fitting Mid-Temperature Creep Data and Rupture Strength 🔬
Chapter 4 of the report deals with Grade 91 creep behavior in the mid-temperature range (about 400–500 °C). The authors first correlate minimum creep rate with time to rupture, then with stress through a power law, and finally re-fit the RCC-MR creep strain law to a merged multi-source database (ORNL, Japanese, and European data), producing new creep strain equations that are compared one by one with existing laws. The rupture strength is then estimated from three independent data sets (RCC-MR, ORNL, Japanese) plus the “minimum commitment method,” with special attention to the reliability of mid-temperature extrapolation.
The engineering conclusion is valuable: Grade 91’s actual mid-temperature rupture strength exceeds the values extrapolated from high-temperature data, so the current NH allowables in this band are conservative and there is legitimate room to raise them. At the same time, scatter among data sources is amplified in the mid-temperature band, so fitting a single database could produce non-conservative numbers. The report therefore insists that any revision of the negligible-creep curve must be based on merged multi-source data and on minimum (not average) rupture strength.
4. Creep-Fatigue: Structural Differences Among Three Procedures ⚙️
Part 2 dissects the ASME NH, RCC-MR, and Japanese DDS creep-fatigue procedures element by element. The ASME procedure starts from elastic analysis, amplifies the nominal elastic strain range with a plasticity correction factor, sets the initial hold-time stress from the elastic calculation with relaxation correction, accumulates creep damage by the time-fraction rule (Dc = Σ t/Td), and adds it linearly to fatigue damage (Df = Σ n/Nd) against a damage limit. The RCC-MR procedure is built on ductility exhaustion, expressing creep damage as the integral of inelastic strain rate over creep ductility, which is closer to the physical damage process. The Japanese DDS procedure adopts yet different simplifications for strain range and relaxation.
| Evaluation Step | ASME NH | RCC-MR | Japanese DDS |
|---|---|---|---|
| Strain range determination | Elastic analysis + plasticity correction factor | Inelastic analysis or elastoplastic correction | Primarily inelastic analysis |
| Initial hold-time stress | Elastic value with relaxation correction | From inelastic relaxation curve | From measured relaxation data |
| Stress relaxation description | Simplified, conservative relation | Integral relaxation with creep law | Based on measured relaxation curves |
| Creep damage accumulation | Time-fraction rule | Ductility exhaustion | Combination of time-fraction and ductility exhaustion |
| Typical tendency | Conservative for long holds and high stress | May be non-conservative for creep-brittle materials | Depends on the national material database |
Using existing Grade 91 creep-fatigue data (tensile holds, compressive holds, varying hold times, temperatures 500–600 °C), the report validates all three procedures and highlights several warning signals: mean stress has a strong effect on life; compressive holds are not always benign; oxidation in air accelerates crack initiation compared with vacuum tests; and cyclic softening of the modified 9Cr-1Mo steel accelerates creep damage accumulation, creating a “softening–damage” positive feedback. None of these mechanisms is explicitly captured in the simplified NH procedure, and they are the principal sources of prediction error.
🚨 The interaction of cyclic softening and oxidation can significantly shorten Grade 91 creep-fatigue life. If elasticity-correction factors and damage limits calibrated for austenitic stainless steels are reused in long-hold, high-temperature, air-environment conditions, non-conservative life predictions may result. Key components should rely on measured material data or more refined inelastic analysis.
5. The Proposed Validation Test Program 🧪
Part 3 lays out a complete test matrix for verifying the two revisions. Materials should cover multiple heats of Grade 91 tube and plate, with chemistry, heat treatment, grain size, and room/high-temperature tensile properties recorded. Creep tests should span 500–650 °C at stress levels giving rupture lives from 103 to 105 h. Creep-fatigue tests should include pure fatigue, tensile holds, compressive holds, and symmetric holds with graded hold times, plus air-versus-vacuum (or inert atmosphere) pairs. Additional tests after prior aging are scheduled to separate microstructural degradation from loading-induced damage. Only such a matrix, the report insists, can support both the negligible-creep curve and the creep-fatigue procedure revisions simultaneously.
6. Engineering Practice Essentials 🛠️
- 🔍 Temperature-zoned design strategy: assess Grade 91 components by location. Use the negligible-creep criterion to simplify low-temperature regions, but complete creep strain, creep ratcheting, and creep-fatigue evaluations for high-temperature regions; never apply a single conclusion to the whole component.
- 🔧 Weldments are the weak link: the fine-grained heat-affected zone (Type IV region) of Grade 91 welds has markedly lower creep strength than the base metal, and creep failures occur there. A weld strength reduction factor must be applied; see the STP-NU-019 series for supporting data.
- 📊 Data quality governs conclusions: allowable creep stresses rely on extrapolation, and the number and scatter of data sources directly determine conservatism. Require multi-heat long-term creep data in procurement specifications.
- 📋 Manufacturing and heat-treatment control: Grade 91 is extremely sensitive to normalize-and-temper practice, hardness limits, and δ-ferrite content; the postweld heat-treatment window is narrow, and any overtemperature irreversibly damages creep strength.
- 🧭 In-service surveillance: creep strain monitoring and replication metallography on hot components can detect Type IV cracks and microstructural degradation at an early stage.
✅ The report’s core contribution is converting the “whether to waive creep evaluation” decision from an empirical judgment into a quantifiable, comparable criterion system, and providing both data and methodology support for the subsequent evolution of the NH allowables for Grade 91. It is essential reading for anyone following Subsection NH development.
7. Frequently Asked Questions ❓
❓ Q: What is the “negligible creep” condition and why does it matter so much in design?
A: Negligible creep means the accumulated creep strain over the design life is small enough to ignore, so the whole family of creep strain limits, creep ratcheting, and creep-fatigue evaluations can be waived. It is the “simplification gate” of high-temperature design and directly controls analysis effort and wall thickness, so its width must be rigorously justified.
A: Negligible creep means the accumulated creep strain over the design life is small enough to ignore, so the whole family of creep strain limits, creep ratcheting, and creep-fatigue evaluations can be waived. It is the “simplification gate” of high-temperature design and directly controls analysis effort and wall thickness, so its width must be rigorously justified.
❓ Q: Which is more conservative for creep-fatigue — ASME NH or RCC-MR?
A: There is no absolute answer. The ASME time-fraction rule tends to be conservative for long holds and high stress; RCC-MR’s ductility-exhaustion method is physically more reasonable but depends heavily on creep ductility data and can be non-conservative for creep-brittle materials. The report recommends evaluating per material and duty rather than choosing “the stricter” side blindly.
A: There is no absolute answer. The ASME time-fraction rule tends to be conservative for long holds and high stress; RCC-MR’s ductility-exhaustion method is physically more reasonable but depends heavily on creep ductility data and can be non-conservative for creep-brittle materials. The report recommends evaluating per material and duty rather than choosing “the stricter” side blindly.
❓ Q: Have the proposed revisions already entered the ASME Code?
A: No. This report is a research output of ASME ST-LLC that provides technical justification for Code development; it has no Code status by itself. The recommendations take effect only after review, ballot, and publication by the relevant ASME BPVC Section III committees, either in a later edition or in a Code Case.
A: No. This report is a research output of ASME ST-LLC that provides technical justification for Code development; it has no Code status by itself. The recommendations take effect only after review, ballot, and publication by the relevant ASME BPVC Section III committees, either in a later edition or in a Code Case.
❓ Q: Why is there room to raise Grade 91 allowables in the mid-temperature range?
A: Because current allowables are extrapolated from high-temperature long-term data, while Grade 91’s actual mid-temperature rupture strength is higher. The report cautions, however, that mid-temperature scatter across data sources is larger, so any increase must be based on merged multi-source data and minimum rupture strength — otherwise safety margins are sacrificed.
A: Because current allowables are extrapolated from high-temperature long-term data, while Grade 91’s actual mid-temperature rupture strength is higher. The report cautions, however, that mid-temperature scatter across data sources is larger, so any increase must be based on merged multi-source data and minimum rupture strength — otherwise safety margins are sacrificed.
