From Monitoring to Supervised Control: A Capability- Maturity Perspective on Digital Twins in Smart Manufacturing

Abstract

Manufacturing digital twins are increasingly promoted as a route from shop-floor visibility to data-driven operational improvement. In practice, however, many implementations remain descriptive dashboards and do not demonstrate measurable gains in quality, throughput, maintenance, or energy performance. Existing reviews also tend to classify studies by algorithm family rather than by the manufacturing capability delivered. This paper develops a capability-maturity lens for AI-enabled manufacturing digital t wins by aligning the NIST progression from descriptive to intelligent capabilities with the ISO 23247 reference architecture and the digital thread. Based on PRISMA-guided screening of 34 studies published between 2018 and 2026, the review shows that curre nt literature is concentrated in descriptive monitoring and predictive analytics, while prescriptive decision support and supervised intelligent control remain comparatively limited. More importantly, many studies do not report standard-compliant integration pathways or KPI-validated shop-floor outcomes. The paper therefore synthesises scenario-level evidence, identifies transition barriers between maturity bands, and proposes an engineering agenda for trustworthy progression toward supervised, auditable control.