Artificial Intelligence for Predictive Maintenance of Mission-Critical Engineered Systems: A Systematic Review of Methods, Frameworks, and Future Directions
Oct 2026· Iconic research and engineering journals· 0 citations· 73 references
TL;DR
A PRISMA-compliant systematic review of the predictive maintenance literature published between January 2018 and June 2026 is reported, culminating in a conceptual unifying framework, referred to here as NEXUS-AI, that combines physical priors, neuro-symbolic reasoning, and human oversight to support verifiable reliability in high-consequence settings.
Abstract
Asset management in safety-critical industries is moving away from calendar-driven preventive schedules and toward autonomous, data-driven decision-making enabled by artificial intelligence (AI) and machine learning (ML). This work reports a PRISMA-compliant systematic review of the predictive maintenance (PdM) literature published between January 2018 and June 2026, covering domains such as aerospace, energy, and advanced manufacturing where component degradation must be tracked with high reliability. A structured search across four bibliographic databases (IEEE Xplore, Scopus, Web of Science, and the ACM Digital Library) produced an initial pool of 2,847 candidate records; after multi-stage screening, 127 peer-reviewed studies were retained for detailed synthesis. The retained literature is organized into six methodological threads: purely data-driven learning, physics-informed machine learning, digital-twin architectures, causal and explainable AI, uncertainty quantification, and human-AI collaboration. Although deep architectures achieve strong benchmark accuracy, the synthesis surfaces recurring shortcomings around out-of-distribution robustness, epistemic uncertainty handling, and model opacity. Eight recurring research gaps are catalogued, and a forward-looking agenda is proposed, culminating in a conceptual unifying framework, referred to here as NEXUS-AI, that combines physical priors, neuro-symbolic reasoning, and human oversight to support verifiable reliability in high-consequence settings.
This study introduces a next-generation AI-driven framework that transforms Reliability-Centered Maintenance (RCM) from static, time-based scheduling into a dynamic, risk-optimized decision system. In industrial environments, limited failure history often forces conservative preventive maintenance (PM), resulting in...
T. Sitthiwasuwat, K. Sasanas, L. Unjantee et al.· ADIPEC· 0 citations
This paper proposes an end-to-end explainable predictive maintenance framework consisting of a contextual probabilistic prognosticator based on convolutional and self-attention networks and a SHAP-based attribution and a risk-sensitive decision layer that is formulated as either a mixed-integer linear programming or di...
K. Selvi· 2026 International Conferenc...· 0 citations
With the rapid advancement of industrial intelligence, equipment predictive maintenance (PdM) has emerged as a critical enabler for intelligent operation and maintenance (O&M) systems. However, conventional machine learning-based PdM methods are inherently constrained by scarce labeled fault samples, weak cross-conditi...
AI should be viewed not as a substitute for engineering judgment, but as a powerful tool whose value depends on disciplined integration with physical reasoning, scientific rigor, and professional responsibility, according to the central message.
Americo Cunha, C. Nataraj, L. Martínez-Piñeiro Lorenzo et al.· Journal of the Brazilian Soc...· 0 citations
Achieving cost-efficient and high-reliability operations is increasingly critical as the energy industry advances toward greater digitalization and sustainability. Drilling services fleets alone account for substantial annual maintenance expenditures, highlighting the need for data-driven approaches that improve asse...
Junko Hutahaean, S. Postovalov, M. Tergeist· SPE Asia Pacific Oil & G...· 0 citations
High predictive accuracy on aircraft maintenance benchmarks does not, by itself, demonstrate validity in operational settings or readiness for airworthiness-compatible decisions. This critical evidence synthesis examines prognostics, fault detection, inspection, digital twins, maintenance optimization, and governance....
I. Alomar· Applied Informatics· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.