Multi-Perspective Trust Framework for Digital Twin Systems: Architectural Design, Massive Twinning, and Stakeholder Assurance

Digital twins (DTs) are increasingly vital in Industry 4.0 and smart infrastructure, yet their widespread adoption hinges on the assurance of trust. While prior studies emphasize architectural capabilities and security concerns, they often overlook trust as a multidimensional and systemic property. This paper introduces a multi-perspective trust framework for DT systems that integrates architectural design, large-scale twinning dynamics, and stakeholder-centric assurance. We propose a five-layer reference architecture that embeds trust as a cross-cutting concern and supports measurable trust constructs through a taxonomy of behavioral and non-behavioral attributes. To address scalability in federated twinning environments, we develop a cluster-based trust management scheme for massive twinning scenarios enabling autonomous and resilient trust assessment. We also synthesize strategies for aligning technical trust metrics with stakeholder expectations, balancing transparency, interpretability. An operational catalogue of trust metrics enables empirical benchmarking and lifecycle trust evaluation. We evaluate the framework in a predictive-maintenance case study of a smart manufacturing line with 500–2000 twins. The cluster-based scheme improves trust-propagation communication by about 40% compared to fully distributed baselines. This is accomplished while maintaining similar prediction accuracy, anomaly detection delay, and trust stability, indicating that it can scale to large DT ecosystems without degrading trust assessments.