Digital Twin–Based Corrosion-Fatigue Crack Growth Prediction in 316L Stainless Steel and Alloy 690 for Nuclear Primary-Circuit Integrity Assessment

Abstract

This study develops a digital-twin framework for corrosion-fatigue crack-growth prediction and integrity assessment of Type 316L austenitic stainless steel and nickel-base Alloy 690 in pressurized-water-reactor primary-circuit environments. The proposed framework couples linear-elastic fracture mechanics, a Paris–Walker-type crack-growth core, material- and environment-dependent acceleration terms, sequential Bayesian updating, uncertainty quantification, and risk-informed inspection logic. The research is structured as a physics-informed simulation study supported by verified literature and nuclear-industry standards rather than as a report of new plant or laboratory measurements. A synthetic population of 800 crack-growth trajectories is used to stress-test model behavior across material class, loading frequency, stress ratio, temperature, dissolved-hydrogen surrogate, cold-work state, initial flaw size, and critical crack size. Three prognostic strategies are compared: a static Paris-law baseline, an environment-corrected static model, and the proposed continuously updated digital twin. In the synthetic evaluation, sequential updating materially reduces remaining-useful-life error, improves crack-size tracking, and produces substantially better calibrated uncertainty intervals than static approaches, particularly under heterogeneous material and environmental states. The study also demonstrates how the twin can translate posterior crack-growth states into probability-of-threshold-exceedance estimates and inspection recommendations without replacing code-required flaw evaluation or engineering authorization. The principal contribution is an integrated integrity-assessment architecture that connects experimentally established corrosion-fatigue and primary-water stress-corrosion-cracking mechanisms with state-synchronized digital-twin prognostics for safety-critical nuclear primary-circuit components. The framework is intended as a research methodology that can later be calibrated and validated using qualified fatigue-crack-growth, environmentally assisted cracking, nondestructive examination, plant transient, and water-chemistry datasets.