CURRENT DEVELOPMENT STATUS: The physics engine has completed preliminary verification against the current benchmark suite for all core modules. Unit consistency across all modules has been verified, with composition handling (wt% ↔ at% conversions) verified throughout the codebase. One industry-standard benchmark grade (Grade 91) has been tested against published literature ranges (preliminary verification, not validation), with all directional responses physically correct: Larson-Miller parameters matching NIMS/ECCC data (LMP 29–31), martensite start temperatures consistent with the Andrews and Kung–Rayment equations, and hardness/creep-life trends correctly ordered across grades. The optimization, uncertainty quantification, and sensitivity analysis modules are fully operational. The tool is now entering the calibration phase, where published creep-rupture and mechanical-property datasets will be used to fit the Wilshire and solid-solution-strengthening coefficients to grade-specific experimental data. Absolute property predictions will be updated accordingly before blind validation against a separate, held-out set of real-world industrial datasets.
Case study #2 (tools: Austenitic and FM) : Remaining life disposition on a Grade 91 header with an experimental FeCrAl coating in sCO2 service
Power plant operators are sitting on Grade 91 steam headers and hot reheat pipework installed in the 1990s-2000s with poorly documented heat-specific composition records. They need run/retire/repair decisions and can’t afford to take units offline for destructive testing campaigns. A single avoided unplanned outage on a large coal or gas unit is worth a lot, so a defensible remaining life report with audit trail justifies the engagement. Steel producers and OEMs developing proprietary 9-12Cr variants for 650°C+ service are running expensive iterative casting programmes with little systematic way to navigate the composition space.
An advanced ultra-supercritical operator has Grade 91 headers approaching end of original design life, now being considered for continued operation in a sCO2 Brayton cycle retrofit. The substrate composition is documented from mill certificates. A FeCrAl coating has been applied to the bore surface as a corrosion barrier. The FM steel tool runs first. It takes the certified Grade 91 composition, the documented service history — temperature, stress, operating hours — and returns a heat-specific creep risk classification and C-distance relative to the calibrated Grade 91 commercial population, with HAZ Type IV risk scoring on the weld seams. Remaining life estimation follows if heat-specific coupon data is available. This establishes whether the substrate constraint or the coating constraint is binding. If the risk classification supports continued operation, the question becomes whether the FeCrAl coating survives the remaining service window in sCO2 at the target conditions.
Austenitic then runs with the substrate identity, the sCO2 temperature and pressure, and the FeCrAl coating system. It returns parabolic oxidation rate, carburisation depth at 650°C, and coating lifetime estimate with the explicitly flagged half-order-of-magnitude uncertainty the sCO2 module carries. If the coating lifetime distribution overlaps poorly with the substrate remaining life — the P10 coating failure time is earlier than the P50 substrate exhaustion — the system flags a coating-limited outcome and the operator’s decision shifts from run/retire on the substrate to coating replacement interval.
Preliminary methodology study: “A Physics-Anchored Two-Layer Creep Risk-Ranking Framework for Grade 91 Power Plant Headers” with the following abstract: Grade 91 steel headers in ageing power plants require run, monitor, or retire decisions, but heat-to-heat compositional variability produces substantial variability in long-term rupture life that generic alloy-class equations cannot capture. This work presents a two-layer creep risk-ranking framework anchored to representative Wilshire physics parameters for Grade 91 and calibrated through a C-distribution derived from 26 certified-composition records across five independent institutional sources. To use the framework an operator supplies five inputs: multi-element certified heat composition, operating temperature, operating stress, product form, and accumulated service hours. The framework returns a risk-zone classification and a C-distance indicating the heat’s position within the calibrated Grade 91 population and identifying heats that warrant targeted coupon sampling before the next planned outage. Blind evaluation on 25 independent post-2015 records demonstrated 92% C-consistency by risk-zone classification. Application of the framework to two heats from a retired 141,000-hour service header correctly ranked the forged reducer as weaker than the seamless pipe, consistent with their observed rupture performance. The present evaluation provides initial validation for qualitative risk ranking and triage, not precision rupture-time prediction, and is positioned accordingly as a structured decision-support framework for maintenance and remaining-life disposition of ageing Grade 91 headers.
Cornacchia, G. (2026). A Physics-Anchored Two-Layer Creep Risk-Ranking Framework for Grade 91 Power Plant Headers. Zenodo. https://doi.org/10.5281/zenodo.21044733