IOT-ENABLED PREDICTIVE MAINTENANCE FOR CYBER-PHYSICAL SYSTEMS IN SAUDI ARABIA: ENHANCING RELIABILITY, EFFICIENCY AND OPERATIONAL RESILIENCE UNDER SAUDI VISION 2030

Production continuity depends directly on the reliability, efficiency, and recoverability of connected physical assets, on which Saudi Arabia’s industrial diversification and digital transformation increasingly depend. This review critically discusses Internet of Things-enabled predictive maintenance for cyber-physical systems, centering on integrating sensing, edge and cloud computing, machine learning, remaining useful life estimation, digital twins, and maintenance decision support into an operational capability aligned with Vision 2030. We used a structured integrative review approach to synthesize thirty-seven peer-reviewed journal articles published between 2020 and 2025, prioritizing studies from Q1 and Q2 journals. The synthesis shows that predictive maintenance creates value only when accurate models are embedded in a trustworthy end-to-end loop linking condition data, prognosis, operational context, maintenance planning, execution, and learning. The evidence supports temporal deep learning, hybrid and physics-informed prognostics, digital twin-assisted reasoning, and edge analytics but also points out recurring weaknesses such as scarce failure labels, model drift, interoperability, cybersecurity, false alarms, and weak integration with work management. These limitations are additionally exacerbated in the Saudi industrial context by brownfield heterogeneity, severe operating conditions, large asset portfolios, and the need to localize advanced technical capabilities. The review proposes a resilience-oriented architecture and staged deployment roadmap in which asset criticality, secure sensing, model assurance, personnel supervision, enterprise integration, and lifecycle governance are considered interdependent design requirements. It concludes that predictive maintenance aligned with Vision 2030 should be evaluated in terms of operational reliability, maintenance productivity, cyber resilience, workforce capability, and measurable business outcomes rather than predictive accuracy alone.