Empirical Multiphase Model for Strength Evolution of High-Chromium Steel Alloys

Duplex stainless steel 2205 (DSS2205) exhibits complex non-monotonic strength evolution during thermal ageing at elevated temperatures, limiting predictive capability for long-term component life assessment. This dissertation develops and validates the METALS (Microstructure Evolution and Thermal Ageing-Linked Strength) model. An empirical multiphase framework coupling phase-field microstructural predictions with mechanical property evolution through phenomenological energy-strength relationships. A comprehensive six-month isothermal ageing study at 500 °C established the experimental foundation, revealing characteristic strength evolution: 47.9% hardening (695 to 1028 MPa, 0-504 hours), 13.6% softening to 873 MPa (1680 hours), and a recovery to 1020 MPa (4368 hours). TEM/EELS analysis at six timepoints identified competitive Cr₂N/MoN precipitation as the governing mechanism, with late-stage Ti/V-containing phases contributing to recovery behaviour. The phase-field model incorporates a critical enhancement distinguishing compositional (deviatoric) and volumetric (dilatational) misfit strain components. Volumetric contributions, while representing a small fraction of total elastic energy during peak precipitation (~5% at 1680h), generating compressive hydrostatic stresses that significantly influence precipitate dissolution kinetics. The elastic energy grows substantially at late stages (reaching ~23% at 3700h), correlating with increased dissolution activity. GPU-accelerated spectral methods enable 4368-hour predictions in ~72 hours, demonstrating practical computational efficiency. Validation demonstrates stage-dependent accuracy: good early-stage agreement (RMSE = 6.18 MPa, 0.5% error for 0-672 hours), moderate intermediate performance (RMSE = 79.48 MPa, 7.79% error for 672-2520 hours), and reasonable late-stage predictions (RMSE = 71.59 MPa, 6.06% beyond 2520h). Parametric sensitivity analysis reveals MoN activation energy critically governs intermediate-stage behaviour, which provides a clear pathway for model refinement. The systematic late-stage underprediction identifies the need for explicit TiN/VN phase incorporation. The validated framework enables quantitative strength predictions for DSS2205 components during service exposure, supports alloy composition optimisation through computational screening, and establishes a transferable methodology applicable to other high-chromium steel systems. The research demonstrates that empirically-calibrated phenomenological coupling within thermodynamically-consistent microstructural frameworks provides actionable predictive capability even when complete mechanistic understanding remains incomplete.

Grantor dc:publisher

Stellenbosch : Stellenbosch University

Year dc:date.issued

2026

Author dc:creator

• Cupido, Llewellyn Heinrich

Advisor dc:contributor.advisor

• Mahomed, Nawaz

Language dc:language.iso

en

Repository record dc:identifier.uri

https://scholar.sun.ac.za/handle/10019.1/135708

OAI identifier oai:identifier

oai:scholar.sun.ac.za:10019.1/135708