ADVANCES AND LIMITATIONS OF FRACTURE MECHANICS–BASED FATIGUE LIFE PREDICTION APPROACHES FOR STRUCTURAL INTEGRITY ASSESSMENT: A SYSTEMATIC REVIEW

Abstract

This mixed-evidence study synthesizes the state of fracture mechanics–based fatigue life prediction and tests what actually improves accuracy and adoption in service-realistic settings. We systematically reviewed 96 peer-reviewed studies, mapped capabilities across ΔK–Paris/Forman/NASGRO baselines, closure-aware and elastic–plastic crack-tip response models, environment-enriched formulations, short-crack treatments, and probabilistic uncertainty methods, and then triangulated the literature with a quantitative cross-sectional survey and explanatory case studies. Design: quantitative cross-sectional, case-based. Sample: three enterprise-relevant structural integrity cases spanning aerospace variable-amplitude spectra in air, offshore steel in saline exposure, and welded details with residual stress. Key variables included model family, treatment of variable-amplitude sequence effects, short-crack and near-threshold behavior, environment conditioning, uncertainty practice, data quality, organizational support, usability burden, adoption, and perceived predictive accuracy. Analysis plan: descriptive synthesis and evidence mapping, Likert-scale reliability checks, correlation and multiple regression with robust errors, and case-level benchmarking using RMSE, bias, and predictive-interval coverage. Headline findings: sequence-aware closure or elastic–plastic models reduced life-prediction error by roughly 18–27 percent under variable-amplitude loading; environment-enriched formulations improved accuracy in chloride media and lifted 90 percent interval coverage from about 0.71 to approximately 0.87 relative to air-calibrated surrogates; short-crack modifications cut near-threshold bias by roughly 30–35 percent. Organizational support, training, and data quality showed strong positive associations with adoption and perceived accuracy, while high usability burden suppressed adoption. Implications: for decision-ready integrity assessments, organizations should standardize a tiered pipeline that matches model physics to service conditions, propagate uncertainty into inspection planning, and invest in spectra governance and workflow automation to translate technical capability into routine, auditable practice.