The transition to Smart Manufacturing requires moving beyond static, one-time design optimisation toward dynamic, lifecycle-aware systems. While structural optimisation methods generate high-performance designs and Digital Twins provide virtual representations, a disconnect persists between offline optimal designs and real-world, variable operating conditions. This paper bridges that gap by proposing a novel closed-loop framework that integrates an AI-driven Digital Twin with adaptive structural re-optimisation. The core contribution is a MATLAB-centric correlation and learning engine that continuously aligns the digital model with its physical counterpart using real-time sensor data. Using machine learning to create surrogate models, the system quantifies simulation-to-reality discrepancies and triggers automatic CAD model updates. Case studies from the aerospace and automotive sectors demonstrate the framework's capability to enable usage-driven adaptation, predictive maintenance, and continuous lifecycle optimisation, thereby advancing the state of the art in adaptive, data-driven design for manufacturing.
Ahmed Yousif (Wed,) studied this question.